entry_point
stringlengths 1
65
| original_triton_code
stringlengths 4.5k
619k
| python_code
stringlengths 208
60.9k
| triton_code
stringlengths 1.15k
275k
| repo_name
stringlengths 7
115
| module_name
stringlengths 1
65
| synthetic
bool 1
class | uuid
int64 0
18.5k
| licenses
listlengths 1
6
| stars
int64 0
19.8k
| sha
stringlengths 40
40
| repo_link
stringlengths 72
180
|
|---|---|---|---|---|---|---|---|---|---|---|---|
MarginCosineProduct
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ub/cubndnqsd3c7k3peegeanrq2reqgcl7uzhnt4z7syuddwnan5w36.py
# Topologically Sorted Source Nodes: [w2, ger, clamp, cosine, scatter_, mul, sub, output], Original ATen: [aten.linalg_vector_norm, aten.mul, aten.clamp, aten.div, aten.scatter, aten.sub]
# Source node to ATen node mapping:
# clamp => clamp_min
# cosine => div
# ger => mul
# mul => mul_1
# output => mul_2
# scatter_ => scatter_upon_const_tensor
# sub => sub
# w2 => pow_3, pow_4, sum_2
# Graph fragment:
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_1, 2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_3, [1]), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_2, 0.5), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %pow_4), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%mul, 1e-08), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mm, %clamp_min), kwargs = {})
# %scatter_upon_const_tensor : [num_users=1] = call_function[target=torch._inductor.fx_passes.post_grad.scatter_upon_const_tensor](args = (), kwargs = {shape: [4, 4], background_val: 0, dtype: torch.float32, dim: 1, selector: %view_1, val: 1.0})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%scatter_upon_const_tensor, 0.4), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div, %mul_1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, 30.0), kwargs = {})
triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0 = async_compile.triton('triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i64', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr2 + (4*x0), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr2 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr2 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp21 = tl.load(in_ptr2 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp14 = tmp13 * tmp13
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = tmp12 * tmp24
tmp26 = 1e-08
tmp27 = triton_helpers.maximum(tmp25, tmp26)
tmp28 = tmp0 / tmp27
tmp30 = x0
tmp31 = tmp29 == tmp30
tmp32 = 1.0
tmp33 = 0.0
tmp34 = tl.where(tmp31, tmp32, tmp33)
tmp35 = 0.4
tmp36 = tmp34 * tmp35
tmp37 = tmp28 - tmp36
tmp38 = 30.0
tmp39 = tmp37 * tmp38
tl.store(in_out_ptr0 + (x2), tmp39, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [ip], Original ATen: [aten.mm]
extern_kernels.mm(primals_2, reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [w2, ger, clamp, cosine, scatter_, mul, sub, output], Original ATen: [aten.linalg_vector_norm, aten.mul, aten.clamp, aten.div, aten.scatter, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0.run(buf2, buf0, primals_2, primals_1, primals_3, 16, grid=grid(16), stream=stream0)
del primals_3
return (buf2, primals_1, primals_2, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.int64)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import Parameter
import torch.utils.data
import torch.optim
def cosine_sim(x1, x2, dim=1, eps=1e-08):
ip = torch.mm(x1, x2.t())
w1 = torch.norm(x1, 2, dim)
w2 = torch.norm(x2, 2, dim)
return ip / torch.ger(w1, w2).clamp(min=eps)
class MarginCosineProduct(nn.Module):
"""Implement of large margin cosine distance: :
Args:
in_features: size of each input sample
out_features: size of each output sample
s: norm of input feature
m: margin
"""
def __init__(self, in_features, out_features, s=30.0, m=0.4):
super(MarginCosineProduct, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.s = s
self.m = m
self.weight = Parameter(torch.Tensor(out_features, in_features))
nn.init.xavier_uniform_(self.weight)
def forward(self, input, label):
cosine = cosine_sim(input, self.weight)
one_hot = torch.zeros_like(cosine)
one_hot.scatter_(1, label.view(-1, 1), 1.0)
output = self.s * (cosine - one_hot * self.m)
return output
def __repr__(self):
return self.__class__.__name__ + '(' + 'in_features=' + str(self.
in_features) + ', out_features=' + str(self.out_features
) + ', s=' + str(self.s) + ', m=' + str(self.m) + ')'
def get_inputs():
return [torch.rand([4, 4]), torch.ones([4], dtype=torch.int64)]
def get_init_inputs():
return [[], {'in_features': 4, 'out_features': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
from torch.nn import Parameter
import torch.utils.data
import torch.optim
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0(in_out_ptr0
, in_ptr0, in_ptr1, in_ptr2, in_ptr3, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr2 + 4 * x0, xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr2 + (1 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr2 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp21 = tl.load(in_ptr2 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp29 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp14 = tmp13 * tmp13
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = tmp12 * tmp24
tmp26 = 1e-08
tmp27 = triton_helpers.maximum(tmp25, tmp26)
tmp28 = tmp0 / tmp27
tmp30 = x0
tmp31 = tmp29 == tmp30
tmp32 = 1.0
tmp33 = 0.0
tmp34 = tl.where(tmp31, tmp32, tmp33)
tmp35 = 0.4
tmp36 = tmp34 * tmp35
tmp37 = tmp28 - tmp36
tmp38 = 30.0
tmp39 = tmp37 * tmp38
tl.store(in_out_ptr0 + x2, tmp39, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_2, reinterpret_tensor(primals_1, (4, 4),
(1, 4), 0), out=buf0)
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf2 = buf1
del buf1
get_raw_stream(0)
triton_poi_fused_clamp_div_linalg_vector_norm_mul_scatter_sub_0[grid
(16)](buf2, buf0, primals_2, primals_1, primals_3, 16, XBLOCK=
16, num_warps=1, num_stages=1)
del primals_3
return buf2, primals_1, primals_2, buf0
def cosine_sim(x1, x2, dim=1, eps=1e-08):
ip = torch.mm(x1, x2.t())
w1 = torch.norm(x1, 2, dim)
w2 = torch.norm(x2, 2, dim)
return ip / torch.ger(w1, w2).clamp(min=eps)
class MarginCosineProductNew(nn.Module):
"""Implement of large margin cosine distance: :
Args:
in_features: size of each input sample
out_features: size of each output sample
s: norm of input feature
m: margin
"""
def __init__(self, in_features, out_features, s=30.0, m=0.4):
super(MarginCosineProductNew, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.s = s
self.m = m
self.weight = Parameter(torch.Tensor(out_features, in_features))
nn.init.xavier_uniform_(self.weight)
def __repr__(self):
return self.__class__.__name__ + '(' + 'in_features=' + str(self.
in_features) + ', out_features=' + str(self.out_features
) + ', s=' + str(self.s) + ', m=' + str(self.m) + ')'
def forward(self, input_0, input_1):
primals_1 = self.weight
primals_2 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3])
return output[0]
|
lindsey98/CosFace_pytorch
|
MarginCosineProduct
| false | 10,397 |
[
"MIT"
] | 0 |
39bddf763e06c7ccd21fbf45d0c7f1f4a9d8d24d
|
https://github.com/lindsey98/CosFace_pytorch/tree/39bddf763e06c7ccd21fbf45d0c7f1f4a9d8d24d
|
SplitDim
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/mt/cmttqkny6sfnxpzdxvh5ov2uac5527dmp3f2ypze3tkvqwdngpc2.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# output => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_3, %view, %slice_5], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp16 = tl.load(in_ptr1 + (0))
tmp17 = tl.broadcast_to(tmp16, [XBLOCK])
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4*x1), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 2, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr0 + (1 + (4*x1)), tmp9 & xmask, eviction_policy='evict_last', other=0.0)
tmp11 = 20.0
tmp12 = tmp10 > tmp11
tmp13 = tl_math.exp(tmp10)
tmp14 = libdevice.log1p(tmp13)
tmp15 = tl.where(tmp12, tmp10, tmp14)
tmp18 = tmp17 > tmp11
tmp19 = tl_math.exp(tmp17)
tmp20 = libdevice.log1p(tmp19)
tmp21 = tl.where(tmp18, tmp17, tmp20)
tmp22 = tmp15 + tmp21
tmp23 = tl.full(tmp22.shape, 0.0, tmp22.dtype)
tmp24 = tl.where(tmp9, tmp22, tmp23)
tmp25 = tmp0 >= tmp7
tmp26 = tl.full([1], 4, tl.int64)
tmp27 = tmp0 < tmp26
tmp28 = tl.load(in_ptr0 + (2 + (4*x1) + ((-2) + x0)), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp29 = tl.where(tmp9, tmp24, tmp28)
tmp30 = tl.where(tmp4, tmp5, tmp29)
tl.store(out_ptr0 + (x2), tmp30, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, primals_2, buf0, 16, grid=grid(16), stream=stream0)
del primals_1
return (buf0, primals_2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn as nn
import torch.utils.data
class SplitDim(nn.Module):
def __init__(self, nonlin_col=1, nonlin_type=torch.nn.functional.
softplus, correction=True):
super(SplitDim, self).__init__()
self.nonlinearity = nonlin_type
self.col = nonlin_col
if correction:
self.var = torch.nn.Parameter(torch.zeros(1))
else:
self.register_buffer('var', torch.ones(1, requires_grad=False) *
-15.0)
self.correction = correction
def forward(self, input):
transformed_output = self.nonlinearity(input[:, self.col])
transformed_output = transformed_output + self.nonlinearity(self.var)
stack_list = [input[:, :self.col], transformed_output.view(-1, 1)]
if self.col + 1 < input.size(1):
stack_list.append(input[:, self.col + 1:])
output = torch.cat(stack_list, 1)
return output
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch import nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp16 = tl.load(in_ptr1 + 0)
tmp17 = tl.broadcast_to(tmp16, [XBLOCK])
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + 4 * x1, tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 2, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr0 + (1 + 4 * x1), tmp9 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp11 = 20.0
tmp12 = tmp10 > tmp11
tmp13 = tl_math.exp(tmp10)
tmp14 = libdevice.log1p(tmp13)
tmp15 = tl.where(tmp12, tmp10, tmp14)
tmp18 = tmp17 > tmp11
tmp19 = tl_math.exp(tmp17)
tmp20 = libdevice.log1p(tmp19)
tmp21 = tl.where(tmp18, tmp17, tmp20)
tmp22 = tmp15 + tmp21
tmp23 = tl.full(tmp22.shape, 0.0, tmp22.dtype)
tmp24 = tl.where(tmp9, tmp22, tmp23)
tmp25 = tmp0 >= tmp7
tl.full([1], 4, tl.int64)
tmp28 = tl.load(in_ptr0 + (2 + 4 * x1 + (-2 + x0)), tmp25 & xmask,
eviction_policy='evict_last', other=0.0)
tmp29 = tl.where(tmp9, tmp24, tmp28)
tmp30 = tl.where(tmp4, tmp5, tmp29)
tl.store(out_ptr0 + x2, tmp30, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(16)](primals_1, primals_2, buf0, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_1
return buf0, primals_2
class SplitDimNew(nn.Module):
def __init__(self, nonlin_col=1, nonlin_type=torch.nn.functional.
softplus, correction=True):
super(SplitDimNew, self).__init__()
self.nonlinearity = nonlin_type
self.col = nonlin_col
if correction:
self.var = torch.nn.Parameter(torch.zeros(1))
else:
self.register_buffer('var', torch.ones(1, requires_grad=False) *
-15.0)
self.correction = correction
def forward(self, input_0):
primals_2 = self.var
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
|
junmokane/rlkit_jm
|
SplitDim
| false | 10,398 |
[
"MIT"
] | 0 |
34a1bcf47706d4c98e9ce3b7edfd96fee6f2dd70
|
https://github.com/junmokane/rlkit_jm/tree/34a1bcf47706d4c98e9ce3b7edfd96fee6f2dd70
|
StyledConv
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wi/cwiyl3lwwtancorrifw77xt3aqb4lermdintht45zvkj3bg54nbl.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, 0.5), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2o/c2oqkq7zaubqmw7vuixxlseb2ff5jzqqbyczicxlmsahuxwdpdyp.py
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, 1.0), kwargs = {})
triton_poi_fused_mul_1 = async_compile.triton('triton_poi_fused_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xb/cxbwntk4p6rlewwyy2lsbaujvvugjyjfuideulqmjkkz6oa23ili.py
# Topologically Sorted Source Nodes: [mul_2, weight_1, pow_1, sum_1, add, rsqrt, weight_2], Original ATen: [aten.mul, aten.pow, aten.sum, aten.add, aten.rsqrt]
# Source node to ATen node mapping:
# add => add
# mul_2 => mul_2
# pow_1 => pow_1
# rsqrt => rsqrt
# sum_1 => sum_1
# weight_1 => mul_3
# weight_2 => mul_4
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_5, 0.125), kwargs = {})
# %mul_3 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_2, %slice_4), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mul_3, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [2, 3, 4], True), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 1e-08), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_3, %rsqrt), kwargs = {})
triton_per_fused_add_mul_pow_rsqrt_sum_2 = async_compile.triton('triton_per_fused_add_mul_pow_rsqrt_sum_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 64],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mul_pow_rsqrt_sum_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_mul_pow_rsqrt_sum_2(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r5 = rindex
x0 = xindex % 4
r3 = (rindex // 16)
x1 = (xindex // 4)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (r5 + (64*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp3 = tl.load(in_ptr1 + (r3 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp1 = 0.125
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tmp5 = tmp4 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = 1e-08
tmp11 = tmp9 + tmp10
tmp12 = libdevice.rsqrt(tmp11)
tmp13 = tmp4 * tmp12
tl.debug_barrier()
tl.store(in_out_ptr0 + (x4), tmp12, xmask)
tl.store(out_ptr0 + (r5 + (64*x4)), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sv/csv7dxittkbvogm6j67l67lwlrg4sqq5fffvklyqhs4hv3hahubs.py
# Topologically Sorted Source Nodes: [mul_5, out_3, add_2, leaky_relu, out_4], Original ATen: [aten.mul, aten.add, aten.leaky_relu, aten.leaky_relu_backward]
# Source node to ATen node mapping:
# add_2 => add_2
# leaky_relu => gt, mul_6, where
# mul_5 => mul_5
# out_3 => add_1
# out_4 => mul_7
# Graph fragment:
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_6, %normal_functional), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_5, %mul_5), kwargs = {})
# %add_2 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %view_6), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%add_2, 0), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_2, 0.2), kwargs = {})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt, %add_2, %mul_6), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%where, 1.4142135623730951), kwargs = {})
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%where, 0), kwargs = {})
triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3 = async_compile.triton('triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*i1', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 25
x2 = (xindex // 100)
x1 = (xindex // 25) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tl.load(in_ptr2 + (x0 + (25*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp4 = tmp2 * tmp3
tmp5 = tmp0 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 0.0
tmp9 = tmp7 > tmp8
tmp10 = 0.2
tmp11 = tmp7 * tmp10
tmp12 = tl.where(tmp9, tmp7, tmp11)
tmp13 = 1.4142135623730951
tmp14 = tmp12 * tmp13
tmp15 = tmp12 > tmp8
tl.store(out_ptr0 + (x3), tmp14, xmask)
tl.store(out_ptr1 + (x3), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
assert_size_stride(primals_6, (1, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_2, buf0, 16, grid=grid(16), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
triton_poi_fused_mul_1.run(primals_3, buf1, 4, grid=grid(4), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, out], Original ATen: [aten.mul, aten.addmm]
extern_kernels.addmm(buf1, reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf1
buf3 = reinterpret_tensor(buf0, (4, 4, 1, 1, 1), (4, 1, 16, 16, 16), 0); del buf0 # reuse
buf4 = reinterpret_tensor(buf3, (4, 4, 1, 1, 1), (4, 1, 1, 1, 1), 0); del buf3 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, weight_1, pow_1, sum_1, add, rsqrt, weight_2], Original ATen: [aten.mul, aten.pow, aten.sum, aten.add, aten.rsqrt]
triton_per_fused_add_mul_pow_rsqrt_sum_2.run(buf4, primals_5, buf2, buf5, 16, 64, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4, 4, 4), (64, 16, 4, 1), 0), stride=(1, 1), padding=(2, 2), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=4, bias=None)
assert_size_stride(buf6, (1, 16, 5, 5), (400, 25, 5, 1))
buf7 = empty_strided_cuda((4, 1, 5, 5), (25, 25, 5, 1), torch.float32)
# Topologically Sorted Source Nodes: [noise], Original ATen: [aten.normal_functional]
buf8 = torch.ops.aten.normal_functional.default(buf7)
del buf7
buf9 = buf8
del buf8
buf10 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.float32)
buf11 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.bool)
# Topologically Sorted Source Nodes: [mul_5, out_3, add_2, leaky_relu, out_4], Original ATen: [aten.mul, aten.add, aten.leaky_relu, aten.leaky_relu_backward]
triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3.run(buf6, primals_6, buf9, primals_7, buf10, buf11, 400, grid=grid(400), stream=stream0)
del buf6
del primals_6
del primals_7
return (buf10, primals_5, reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), reinterpret_tensor(buf2, (4, 1, 4, 1, 1), (64, 64, 1, 1, 1), 0), buf4, reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4, 4, 4), (64, 16, 4, 1), 0), buf9, buf11, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, 4, 4, 4, 4), (256, 64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.autograd import Function
import math
import torch
from torch import nn
from torch.nn import functional as F
from torch.nn.functional import leaky_relu
def fused_leaky_relu(input_, bias, negative_slope=0.2, scale=2 ** 0.5):
return scale * leaky_relu(input_ + bias[:input_.shape[1]],
negative_slope, inplace=True)
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if k.ndim == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
k = torch.flip(k, [0, 1])
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, ch, _in_h, _in_w = input.shape
kernel_h, kernel_w = kernel.shape
assert up_y == up_x and up_y in [1, 2]
if up_y == 2:
w = input.new_zeros(2, 2)
w[0, 0] = 1
out = F.conv_transpose2d(input, w.view(1, 1, 2, 2).repeat(ch, 1, 1,
1), groups=ch, stride=2)
else:
out = input
out = F.pad(out, [pad_x0, pad_x1, pad_y0, pad_y1])
out = F.conv2d(out, kernel.view(1, 1, kernel_h, kernel_w).repeat(ch, 1,
1, 1), groups=ch)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
if input.device.type == 'cpu':
out = upfirdn2d_native(input, kernel, up, up, down, down, pad[0],
pad[1], pad[0], pad[1])
else:
out = UpFirDn2d.apply(input, kernel, (up, up), (down, down), (pad[0
], pad[1], pad[0], pad[1]))
return out
class FusedLeakyReLU(nn.Module):
def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5):
super().__init__()
self.bias = nn.Parameter(torch.zeros(channel))
self.negative_slope = negative_slope
self.scale = scale
def forward(self, x):
return self.scale * leaky_relu(x + self.bias.reshape((1, -1, 1, 1))
[:, :x.shape[1]], self.negative_slope, inplace=True)
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1.0,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = 1 / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, x):
if self.activation:
out = F.linear(x, self.weight * self.scale)
if self.activation == 'lrelu':
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
raise NotImplementedError
else:
out = F.linear(x, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class UpFirDn2dBackward(Function):
@staticmethod
def forward(ctx, grad_output, kernel, grad_kernel, up, down, pad, g_pad,
in_size, out_size):
up_x, up_y = up
down_x, down_y = down
g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1 = g_pad
grad_output = grad_output.reshape(-1, out_size[0], out_size[1], 1)
grad_input = upfirdn2d_op.upfirdn2d(grad_output, grad_kernel,
down_x, down_y, up_x, up_y, g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1)
grad_input = grad_input.view(in_size[0], in_size[1], in_size[2],
in_size[3])
ctx.save_for_backward(kernel)
pad_x0, pad_x1, pad_y0, pad_y1 = pad
ctx.up_x = up_x
ctx.up_y = up_y
ctx.down_x = down_x
ctx.down_y = down_y
ctx.pad_x0 = pad_x0
ctx.pad_x1 = pad_x1
ctx.pad_y0 = pad_y0
ctx.pad_y1 = pad_y1
ctx.in_size = in_size
ctx.out_size = out_size
return grad_input
@staticmethod
def backward(ctx, gradgrad_input):
kernel, = ctx.saved_tensors
gradgrad_input = gradgrad_input.reshape(-1, ctx.in_size[2], ctx.
in_size[3], 1)
gradgrad_out = upfirdn2d_op.upfirdn2d(gradgrad_input, kernel, ctx.
up_x, ctx.up_y, ctx.down_x, ctx.down_y, ctx.pad_x0, ctx.pad_x1,
ctx.pad_y0, ctx.pad_y1)
gradgrad_out = gradgrad_out.view(ctx.in_size[0], ctx.in_size[1],
ctx.out_size[0], ctx.out_size[1])
return gradgrad_out, None, None, None, None, None, None, None, None
class UpFirDn2d(Function):
@staticmethod
def forward(ctx, input, kernel, up, down, pad):
up_x, up_y = up
down_x, down_y = down
pad_x0, pad_x1, pad_y0, pad_y1 = pad
kernel_h, kernel_w = kernel.shape
_batch, channel, in_h, in_w = input.shape
ctx.in_size = input.shape
input = input.reshape(-1, in_h, in_w, 1)
ctx.save_for_backward(kernel, torch.flip(kernel, [0, 1]))
out_h = (in_h * up_y + pad_y0 + pad_y1 - kernel_h) // down_y + 1
out_w = (in_w * up_x + pad_x0 + pad_x1 - kernel_w) // down_x + 1
ctx.out_size = out_h, out_w
ctx.up = up_x, up_y
ctx.down = down_x, down_y
ctx.pad = pad_x0, pad_x1, pad_y0, pad_y1
g_pad_x0 = kernel_w - pad_x0 - 1
g_pad_y0 = kernel_h - pad_y0 - 1
g_pad_x1 = in_w * up_x - out_w * down_x + pad_x0 - up_x + 1
g_pad_y1 = in_h * up_y - out_h * down_y + pad_y0 - up_y + 1
ctx.g_pad = g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1
out = upfirdn2d_op.upfirdn2d(input, kernel, up_x, up_y, down_x,
down_y, pad_x0, pad_x1, pad_y0, pad_y1)
out = out.view(-1, channel, out_h, out_w)
return out
@staticmethod
def backward(ctx, grad_output):
kernel, grad_kernel = ctx.saved_tensors
grad_input = UpFirDn2dBackward.apply(grad_output, kernel,
grad_kernel, ctx.up, ctx.down, ctx.pad, ctx.g_pad, ctx.in_size,
ctx.out_size)
return grad_input, None, None, None, None
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, x):
out = upfirdn2d(x, self.kernel, pad=self.pad)
return out
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=(1,
3, 3, 1)):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
assert not downsample, 'Downsample is not implemented yet!'
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
self.blur = Blur(blur_kernel, pad=((p + 1) // 2 + factor - 1, p //
2 + 1), upsample_factor=factor)
self.scale = 1 / math.sqrt(in_channel * kernel_size ** 2)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, x, style):
batch, in_channel, height, width = x.shape
style = self.modulation(style)
style = style.view(batch, 1, -1, 1, 1)
first_k_oup = self.first_k_oup if hasattr(self, 'first_k_oup'
) and self.first_k_oup is not None else self.weight.shape[1]
assert first_k_oup <= self.weight.shape[1]
weight = self.weight
weight = weight[:, :first_k_oup, :in_channel].contiguous()
weight = self.scale * weight * style[:, :, :in_channel]
if self.demodulate:
weight = weight * torch.rsqrt(weight.pow(2).sum([2, 3, 4],
keepdim=True) + self.eps)
if self.upsample:
x = x.view(1, batch * in_channel, height, width)
weight = weight.transpose(1, 2)
weight = weight.reshape(weight.shape[0] * weight.shape[1],
weight.shape[2], weight.shape[3], weight.shape[4])
out = F.conv_transpose2d(x, weight, padding=0, stride=2, groups
=batch)
out = out.view(batch, -1, out.shape[-2], out.shape[-1])
out = self.blur(out)
else:
x = x.contiguous().view(1, batch * in_channel, height, width)
weight = weight.view(weight.shape[0] * weight.shape[1], weight.
shape[2], weight.shape[3], weight.shape[4])
out = F.conv2d(x, weight, padding=self.padding, groups=batch)
out = out.view(batch, -1, out.shape[-2], out.shape[-1])
return out
class NoiseInjection(nn.Module):
def __init__(self):
super().__init__()
self.weight = nn.Parameter(torch.zeros(1))
def forward(self, image, noise=None):
if noise is None:
batch, _, height, width = image.shape
noise = image.new_empty(batch, 1, height, width).normal_()
return image + self.weight * noise
class StyledConv(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
upsample=False, blur_kernel=(1, 3, 3, 1), demodulate=True,
activation='lrelu'):
super().__init__()
self.conv = ModulatedConv2d(in_channel, out_channel, kernel_size,
style_dim, upsample=upsample, blur_kernel=blur_kernel,
demodulate=demodulate)
self.noise = NoiseInjection()
if activation == 'lrelu':
self.activate = FusedLeakyReLU(out_channel)
else:
raise NotImplementedError
def forward(self, x, style, noise=None):
out = self.conv(x, style)
out = self.noise(out, noise=noise)
out = self.activate(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'out_channel': 4, 'kernel_size': 4,
'style_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch.autograd import Function
import math
from torch import nn
from torch.nn import functional as F
from torch.nn.functional import leaky_relu
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_per_fused_add_mul_pow_rsqrt_sum_2(in_out_ptr0, in_ptr0, in_ptr1,
out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r5 = rindex
x0 = xindex % 4
r3 = rindex // 16
x1 = xindex // 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (r5 + 64 * x0), xmask, eviction_policy=
'evict_last', other=0.0)
tmp3 = tl.load(in_ptr1 + (r3 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp1 = 0.125
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tmp5 = tmp4 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = 1e-08
tmp11 = tmp9 + tmp10
tmp12 = libdevice.rsqrt(tmp11)
tmp13 = tmp4 * tmp12
tl.debug_barrier()
tl.store(in_out_ptr0 + x4, tmp12, xmask)
tl.store(out_ptr0 + (r5 + 64 * x4), tmp13, xmask)
@triton.jit
def triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3(in_ptr0,
in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr
):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 25
x2 = xindex // 100
x1 = xindex // 25 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tl.load(in_ptr2 + (x0 + 25 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp4 = tmp2 * tmp3
tmp5 = tmp0 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 0.0
tmp9 = tmp7 > tmp8
tmp10 = 0.2
tmp11 = tmp7 * tmp10
tmp12 = tl.where(tmp9, tmp7, tmp11)
tmp13 = 1.4142135623730951
tmp14 = tmp12 * tmp13
tmp15 = tmp12 > tmp8
tl.store(out_ptr0 + x3, tmp14, xmask)
tl.store(out_ptr1 + x3, tmp15, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
assert_size_stride(primals_6, (1,), (1,))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(16)](primals_2, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_3, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(buf1, reinterpret_tensor(primals_4, (64, 4), (
4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1,
beta=1, out=buf2)
del buf1
buf3 = reinterpret_tensor(buf0, (4, 4, 1, 1, 1), (4, 1, 16, 16, 16), 0)
del buf0
buf4 = reinterpret_tensor(buf3, (4, 4, 1, 1, 1), (4, 1, 1, 1, 1), 0)
del buf3
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_per_fused_add_mul_pow_rsqrt_sum_2[grid(16)](buf4, primals_5,
buf2, buf5, 16, 64, XBLOCK=1, num_warps=2, num_stages=1)
buf6 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1,
16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4,
4, 4), (64, 16, 4, 1), 0), stride=(1, 1), padding=(2, 2),
dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=4, bias=None)
assert_size_stride(buf6, (1, 16, 5, 5), (400, 25, 5, 1))
buf7 = empty_strided_cuda((4, 1, 5, 5), (25, 25, 5, 1), torch.float32)
buf8 = torch.ops.aten.normal_functional.default(buf7)
del buf7
buf9 = buf8
del buf8
buf10 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.float32
)
buf11 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.bool)
triton_poi_fused_add_leaky_relu_leaky_relu_backward_mul_3[grid(400)](
buf6, primals_6, buf9, primals_7, buf10, buf11, 400, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
del primals_6
del primals_7
return buf10, primals_5, reinterpret_tensor(primals_4, (64, 4), (4, 1), 0
), reinterpret_tensor(buf2, (4, 1, 4, 1, 1), (64, 64, 1, 1, 1), 0
), buf4, reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 16, 4,
1), 0), reinterpret_tensor(buf5, (16, 4, 4, 4), (64, 16, 4, 1), 0
), buf9, buf11
def fused_leaky_relu(input_, bias, negative_slope=0.2, scale=2 ** 0.5):
return scale * leaky_relu(input_ + bias[:input_.shape[1]],
negative_slope, inplace=True)
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if k.ndim == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
k = torch.flip(k, [0, 1])
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, ch, _in_h, _in_w = input.shape
kernel_h, kernel_w = kernel.shape
assert up_y == up_x and up_y in [1, 2]
if up_y == 2:
w = input.new_zeros(2, 2)
w[0, 0] = 1
out = F.conv_transpose2d(input, w.view(1, 1, 2, 2).repeat(ch, 1, 1,
1), groups=ch, stride=2)
else:
out = input
out = F.pad(out, [pad_x0, pad_x1, pad_y0, pad_y1])
out = F.conv2d(out, kernel.view(1, 1, kernel_h, kernel_w).repeat(ch, 1,
1, 1), groups=ch)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
if input.device.type == 'cpu':
out = upfirdn2d_native(input, kernel, up, up, down, down, pad[0],
pad[1], pad[0], pad[1])
else:
out = UpFirDn2d.apply(input, kernel, (up, up), (down, down), (pad[0
], pad[1], pad[0], pad[1]))
return out
class FusedLeakyReLU(nn.Module):
def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5):
super().__init__()
self.bias = nn.Parameter(torch.zeros(channel))
self.negative_slope = negative_slope
self.scale = scale
def forward(self, x):
return self.scale * leaky_relu(x + self.bias.reshape((1, -1, 1, 1))
[:, :x.shape[1]], self.negative_slope, inplace=True)
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1.0,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = 1 / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, x):
if self.activation:
out = F.linear(x, self.weight * self.scale)
if self.activation == 'lrelu':
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
raise NotImplementedError
else:
out = F.linear(x, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class UpFirDn2dBackward(Function):
@staticmethod
def forward(ctx, grad_output, kernel, grad_kernel, up, down, pad, g_pad,
in_size, out_size):
up_x, up_y = up
down_x, down_y = down
g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1 = g_pad
grad_output = grad_output.reshape(-1, out_size[0], out_size[1], 1)
grad_input = upfirdn2d_op.upfirdn2d(grad_output, grad_kernel,
down_x, down_y, up_x, up_y, g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1)
grad_input = grad_input.view(in_size[0], in_size[1], in_size[2],
in_size[3])
ctx.save_for_backward(kernel)
pad_x0, pad_x1, pad_y0, pad_y1 = pad
ctx.up_x = up_x
ctx.up_y = up_y
ctx.down_x = down_x
ctx.down_y = down_y
ctx.pad_x0 = pad_x0
ctx.pad_x1 = pad_x1
ctx.pad_y0 = pad_y0
ctx.pad_y1 = pad_y1
ctx.in_size = in_size
ctx.out_size = out_size
return grad_input
@staticmethod
def backward(ctx, gradgrad_input):
kernel, = ctx.saved_tensors
gradgrad_input = gradgrad_input.reshape(-1, ctx.in_size[2], ctx.
in_size[3], 1)
gradgrad_out = upfirdn2d_op.upfirdn2d(gradgrad_input, kernel, ctx.
up_x, ctx.up_y, ctx.down_x, ctx.down_y, ctx.pad_x0, ctx.pad_x1,
ctx.pad_y0, ctx.pad_y1)
gradgrad_out = gradgrad_out.view(ctx.in_size[0], ctx.in_size[1],
ctx.out_size[0], ctx.out_size[1])
return gradgrad_out, None, None, None, None, None, None, None, None
class UpFirDn2d(Function):
@staticmethod
def forward(ctx, input, kernel, up, down, pad):
up_x, up_y = up
down_x, down_y = down
pad_x0, pad_x1, pad_y0, pad_y1 = pad
kernel_h, kernel_w = kernel.shape
_batch, channel, in_h, in_w = input.shape
ctx.in_size = input.shape
input = input.reshape(-1, in_h, in_w, 1)
ctx.save_for_backward(kernel, torch.flip(kernel, [0, 1]))
out_h = (in_h * up_y + pad_y0 + pad_y1 - kernel_h) // down_y + 1
out_w = (in_w * up_x + pad_x0 + pad_x1 - kernel_w) // down_x + 1
ctx.out_size = out_h, out_w
ctx.up = up_x, up_y
ctx.down = down_x, down_y
ctx.pad = pad_x0, pad_x1, pad_y0, pad_y1
g_pad_x0 = kernel_w - pad_x0 - 1
g_pad_y0 = kernel_h - pad_y0 - 1
g_pad_x1 = in_w * up_x - out_w * down_x + pad_x0 - up_x + 1
g_pad_y1 = in_h * up_y - out_h * down_y + pad_y0 - up_y + 1
ctx.g_pad = g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1
out = upfirdn2d_op.upfirdn2d(input, kernel, up_x, up_y, down_x,
down_y, pad_x0, pad_x1, pad_y0, pad_y1)
out = out.view(-1, channel, out_h, out_w)
return out
@staticmethod
def backward(ctx, grad_output):
kernel, grad_kernel = ctx.saved_tensors
grad_input = UpFirDn2dBackward.apply(grad_output, kernel,
grad_kernel, ctx.up, ctx.down, ctx.pad, ctx.g_pad, ctx.in_size,
ctx.out_size)
return grad_input, None, None, None, None
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, x):
out = upfirdn2d(x, self.kernel, pad=self.pad)
return out
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=(1,
3, 3, 1)):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
assert not downsample, 'Downsample is not implemented yet!'
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
self.blur = Blur(blur_kernel, pad=((p + 1) // 2 + factor - 1, p //
2 + 1), upsample_factor=factor)
self.scale = 1 / math.sqrt(in_channel * kernel_size ** 2)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, x, style):
batch, in_channel, height, width = x.shape
style = self.modulation(style)
style = style.view(batch, 1, -1, 1, 1)
first_k_oup = self.first_k_oup if hasattr(self, 'first_k_oup'
) and self.first_k_oup is not None else self.weight.shape[1]
assert first_k_oup <= self.weight.shape[1]
weight = self.weight
weight = weight[:, :first_k_oup, :in_channel].contiguous()
weight = self.scale * weight * style[:, :, :in_channel]
if self.demodulate:
weight = weight * torch.rsqrt(weight.pow(2).sum([2, 3, 4],
keepdim=True) + self.eps)
if self.upsample:
x = x.view(1, batch * in_channel, height, width)
weight = weight.transpose(1, 2)
weight = weight.reshape(weight.shape[0] * weight.shape[1],
weight.shape[2], weight.shape[3], weight.shape[4])
out = F.conv_transpose2d(x, weight, padding=0, stride=2, groups
=batch)
out = out.view(batch, -1, out.shape[-2], out.shape[-1])
out = self.blur(out)
else:
x = x.contiguous().view(1, batch * in_channel, height, width)
weight = weight.view(weight.shape[0] * weight.shape[1], weight.
shape[2], weight.shape[3], weight.shape[4])
out = F.conv2d(x, weight, padding=self.padding, groups=batch)
out = out.view(batch, -1, out.shape[-2], out.shape[-1])
return out
class NoiseInjection(nn.Module):
def __init__(self):
super().__init__()
self.weight = nn.Parameter(torch.zeros(1))
def forward(self, image, noise=None):
if noise is None:
batch, _, height, width = image.shape
noise = image.new_empty(batch, 1, height, width).normal_()
return image + self.weight * noise
class StyledConvNew(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
upsample=False, blur_kernel=(1, 3, 3, 1), demodulate=True,
activation='lrelu'):
super().__init__()
self.conv = ModulatedConv2d(in_channel, out_channel, kernel_size,
style_dim, upsample=upsample, blur_kernel=blur_kernel,
demodulate=demodulate)
self.noise = NoiseInjection()
if activation == 'lrelu':
self.activate = FusedLeakyReLU(out_channel)
else:
raise NotImplementedError
def forward(self, input_0, input_1):
primals_5 = self.conv.weight
primals_2 = self.conv.modulation.weight
primals_3 = self.conv.modulation.bias
primals_6 = self.noise.weight
primals_7 = self.activate.bias
primals_1 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
jchetboun/anycost-gan
|
StyledConv
| false | 10,399 |
[
"MIT"
] | 0 |
7e0005e50b915e2dfeb90fe7a9846c5df38d7c06
|
https://github.com/jchetboun/anycost-gan/tree/7e0005e50b915e2dfeb90fe7a9846c5df38d7c06
|
DiceLoss
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/f2/cf2m6qhcmiwbtav7gxsyv32awtl7srngldst5zjqkg6jzhbuk257.py
# Topologically Sorted Source Nodes: [mul, intersect, mul_1, add, denominator, clamp, per_channel_dice, sub, mean], Original ATen: [aten.mul, aten.sum, aten.add, aten.clamp, aten.div, aten.rsub, aten.mean]
# Source node to ATen node mapping:
# add => add
# clamp => clamp_min
# denominator => sum_2
# intersect => sum_1
# mean => mean
# mul => mul
# mul_1 => mul_1
# per_channel_dice => div
# sub => sub
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %view_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [-1]), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 2.0), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view, %view_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add, [-1]), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sum_2, 1e-05), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_1, %clamp_min), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %div), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub,), kwargs = {})
triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0 = async_compile.triton('triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp2 = tl.load(in_ptr1 + (r0), None)
tmp4 = tl.load(in_ptr0 + (4 + r0), None)
tmp6 = tl.load(in_ptr1 + (4 + r0), None)
tmp9 = tl.load(in_ptr0 + (8 + r0), None)
tmp11 = tl.load(in_ptr1 + (8 + r0), None)
tmp14 = tl.load(in_ptr0 + (12 + r0), None)
tmp16 = tl.load(in_ptr1 + (12 + r0), None)
tmp1 = tl.sigmoid(tmp0)
tmp3 = tmp1 * tmp2
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp8 = tmp3 + tmp7
tmp10 = tl.sigmoid(tmp9)
tmp12 = tmp10 * tmp11
tmp13 = tmp8 + tmp12
tmp15 = tl.sigmoid(tmp14)
tmp17 = tmp15 * tmp16
tmp18 = tmp13 + tmp17
tmp19 = 2.0
tmp20 = tmp18 * tmp19
tmp21 = tmp1 + tmp2
tmp22 = tmp5 + tmp6
tmp23 = tmp21 + tmp22
tmp24 = tmp10 + tmp11
tmp25 = tmp23 + tmp24
tmp26 = tmp15 + tmp16
tmp27 = tmp25 + tmp26
tmp28 = 1e-05
tmp29 = triton_helpers.maximum(tmp27, tmp28)
tmp30 = tmp20 / tmp29
tmp31 = 1.0
tmp32 = tmp31 - tmp30
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = tl.sum(tmp33, 1)[:, None]
tmp36 = 4.0
tmp37 = tmp35 / tmp36
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp37, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [mul, intersect, mul_1, add, denominator, clamp, per_channel_dice, sub, mean], Original ATen: [aten.mul, aten.sum, aten.add, aten.clamp, aten.div, aten.rsub, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0.run(buf2, arg0_1, arg1_1, 1, 4, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
from torch.autograd import Variable
def flatten(tensor):
"""Flattens a given tensor such that the channel axis is first.
The shapes are transformed as follows:
(N, C, D, H, W) -> (C, N * D * H * W)
"""
C = tensor.size(1)
axis_order = (1, 0) + tuple(range(2, tensor.dim()))
transposed = tensor.permute(axis_order)
return transposed.view(C, -1)
def compute_per_channel_dice(input, target, epsilon=1e-05, ignore_index=
None, weight=None):
assert input.size() == target.size(
), "'input' and 'target' must have the same shape"
if ignore_index is not None:
mask = target.clone().ne_(ignore_index)
mask.requires_grad = False
input = input * mask
target = target * mask
input = flatten(input)
target = flatten(target)
target = target.float()
intersect = (input * target).sum(-1)
if weight is not None:
intersect = weight * intersect
denominator = (input + target).sum(-1)
return 2.0 * intersect / denominator.clamp(min=epsilon)
class DiceLoss(nn.Module):
"""Computes Dice Loss, which just 1 - DiceCoefficient described above.
Additionally allows per-class weights to be provided.
"""
def __init__(self, epsilon=1e-05, weight=None, ignore_index=None,
sigmoid_normalization=True, skip_last_target=False):
super(DiceLoss, self).__init__()
self.epsilon = epsilon
self.register_buffer('weight', weight)
self.ignore_index = ignore_index
if sigmoid_normalization:
self.normalization = nn.Sigmoid()
else:
self.normalization = nn.Softmax(dim=1)
self.skip_last_target = skip_last_target
def forward(self, input, target):
input = self.normalization(input)
if self.weight is not None:
weight = Variable(self.weight, requires_grad=False)
else:
weight = None
if self.skip_last_target:
target = target[:, :-1, ...]
per_channel_dice = compute_per_channel_dice(input, target, epsilon=
self.epsilon, ignore_index=self.ignore_index, weight=weight)
return torch.mean(1.0 - per_channel_dice)
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0(in_out_ptr0, in_ptr0,
in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp2 = tl.load(in_ptr1 + r0, None)
tmp4 = tl.load(in_ptr0 + (4 + r0), None)
tmp6 = tl.load(in_ptr1 + (4 + r0), None)
tmp9 = tl.load(in_ptr0 + (8 + r0), None)
tmp11 = tl.load(in_ptr1 + (8 + r0), None)
tmp14 = tl.load(in_ptr0 + (12 + r0), None)
tmp16 = tl.load(in_ptr1 + (12 + r0), None)
tmp1 = tl.sigmoid(tmp0)
tmp3 = tmp1 * tmp2
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp8 = tmp3 + tmp7
tmp10 = tl.sigmoid(tmp9)
tmp12 = tmp10 * tmp11
tmp13 = tmp8 + tmp12
tmp15 = tl.sigmoid(tmp14)
tmp17 = tmp15 * tmp16
tmp18 = tmp13 + tmp17
tmp19 = 2.0
tmp20 = tmp18 * tmp19
tmp21 = tmp1 + tmp2
tmp22 = tmp5 + tmp6
tmp23 = tmp21 + tmp22
tmp24 = tmp10 + tmp11
tmp25 = tmp23 + tmp24
tmp26 = tmp15 + tmp16
tmp27 = tmp25 + tmp26
tmp28 = 1e-05
tmp29 = triton_helpers.maximum(tmp27, tmp28)
tmp30 = tmp20 / tmp29
tmp31 = 1.0
tmp32 = tmp31 - tmp30
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = tl.sum(tmp33, 1)[:, None]
tmp36 = 4.0
tmp37 = tmp35 / tmp36
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp37, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
get_raw_stream(0)
triton_per_fused_add_clamp_div_mean_mul_rsub_sum_0[grid(1)](buf2,
arg0_1, arg1_1, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf2,
def flatten(tensor):
"""Flattens a given tensor such that the channel axis is first.
The shapes are transformed as follows:
(N, C, D, H, W) -> (C, N * D * H * W)
"""
C = tensor.size(1)
axis_order = (1, 0) + tuple(range(2, tensor.dim()))
transposed = tensor.permute(axis_order)
return transposed.view(C, -1)
def compute_per_channel_dice(input, target, epsilon=1e-05, ignore_index=
None, weight=None):
assert input.size() == target.size(
), "'input' and 'target' must have the same shape"
if ignore_index is not None:
mask = target.clone().ne_(ignore_index)
mask.requires_grad = False
input = input * mask
target = target * mask
input = flatten(input)
target = flatten(target)
target = target.float()
intersect = (input * target).sum(-1)
if weight is not None:
intersect = weight * intersect
denominator = (input + target).sum(-1)
return 2.0 * intersect / denominator.clamp(min=epsilon)
class DiceLossNew(nn.Module):
"""Computes Dice Loss, which just 1 - DiceCoefficient described above.
Additionally allows per-class weights to be provided.
"""
def __init__(self, epsilon=1e-05, weight=None, ignore_index=None,
sigmoid_normalization=True, skip_last_target=False):
super(DiceLossNew, self).__init__()
self.epsilon = epsilon
self.register_buffer('weight', weight)
self.ignore_index = ignore_index
if sigmoid_normalization:
self.normalization = nn.Sigmoid()
else:
self.normalization = nn.Softmax(dim=1)
self.skip_last_target = skip_last_target
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
joowlim/pytorch-3dunet
|
DiceLoss
| false | 10,400 |
[
"MIT"
] | 0 |
d08049f60b619627521efd0fb171247e1536b262
|
https://github.com/joowlim/pytorch-3dunet/tree/d08049f60b619627521efd0fb171247e1536b262
|
InferenceNetLSTMCell
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ys/cys3jebwwb7bbssorkkzehii3rj2g54ysfj5qycyaqaktbfse5sq.py
# Topologically Sorted Source Nodes: [i_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# i_1 => sigmoid
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem,), kwargs = {})
triton_poi_fused_sigmoid_0 = async_compile.triton('triton_poi_fused_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + (16*x1)), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (x0 + (16*x1)), xmask)
tmp7 = tl.load(in_ptr4 + (x0 + (16*x1)), xmask)
tmp8 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + (x2), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7h/c7hxidbimuulhwlkxrqzggjodto5os62ortmymrajg4mwpsflkh7.py
# Topologically Sorted Source Nodes: [f_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# f_1 => sigmoid_1
# Graph fragment:
# %sigmoid_1 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_1,), kwargs = {})
triton_poi_fused_sigmoid_1 = async_compile.triton('triton_poi_fused_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (4 + x0 + (16*x1)), xmask)
tmp3 = tl.load(in_ptr2 + (4 + x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (4 + x0 + (16*x1)), xmask)
tmp7 = tl.load(in_ptr4 + (4 + x0 + (16*x1)), xmask)
tmp8 = tl.load(in_ptr5 + (4 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + (x2), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/bd/cbdel7ieyqyotjj3l56pbhbkdykcl7u4krkncs4ch6tzxqucyzky.py
# Topologically Sorted Source Nodes: [o_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# o_1 => sigmoid_3
# Graph fragment:
# %sigmoid_3 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_3,), kwargs = {})
triton_poi_fused_sigmoid_2 = async_compile.triton('triton_poi_fused_sigmoid_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (12 + x0 + (16*x1)), xmask)
tmp3 = tl.load(in_ptr2 + (12 + x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (12 + x0 + (16*x1)), xmask)
tmp7 = tl.load(in_ptr4 + (12 + x0 + (16*x1)), xmask)
tmp8 = tl.load(in_ptr5 + (12 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + (x2), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/um/cumvkqrcq4v7n6s35jolkn2wu5c7x5ct3me73ev7qzlxjowcziea.py
# Topologically Sorted Source Nodes: [mul_2, mul_3, new_c, tanh, new_h, new_h_2, new_c_1], Original ATen: [aten.mul, aten.add, aten.tanh, aten.native_layer_norm]
# Source node to ATen node mapping:
# mul_2 => mul_2
# mul_3 => mul_3
# new_c => add_2
# new_c_1 => var_mean_1
# new_h => mul_4
# new_h_2 => var_mean
# tanh => tanh
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %primals_18), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_18), kwargs = {})
# %add_2 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %mul_3), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_2,), kwargs = {})
# %mul_4 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_3, %tanh), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%mul_4, [3]), kwargs = {correction: 0, keepdim: True})
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add_2, [3]), kwargs = {correction: 0, keepdim: True})
triton_poi_fused_add_mul_native_layer_norm_tanh_3 = async_compile.triton('triton_poi_fused_add_mul_native_layer_norm_tanh_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_native_layer_norm_tanh_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_native_layer_norm_tanh_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr1, out_ptr2, out_ptr3, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (4*x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + (4*x0), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr3 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp21 = tl.load(in_ptr2 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr3 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr2 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp33 = tl.load(in_ptr3 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp5 = tmp4 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = libdevice.tanh(tmp6)
tmp8 = tmp0 * tmp7
tmp12 = tmp10 * tmp11
tmp14 = tmp13 * tmp11
tmp15 = tmp12 + tmp14
tmp16 = libdevice.tanh(tmp15)
tmp17 = tmp9 * tmp16
tmp18 = tmp8 + tmp17
tmp22 = tmp20 * tmp21
tmp24 = tmp23 * tmp21
tmp25 = tmp22 + tmp24
tmp26 = libdevice.tanh(tmp25)
tmp27 = tmp19 * tmp26
tmp28 = tmp18 + tmp27
tmp32 = tmp30 * tmp31
tmp34 = tmp33 * tmp31
tmp35 = tmp32 + tmp34
tmp36 = libdevice.tanh(tmp35)
tmp37 = tmp29 * tmp36
tmp38 = tmp28 + tmp37
tmp39 = 4.0
tmp40 = tmp38 / tmp39
tmp41 = tmp8 - tmp40
tmp42 = tmp41 * tmp41
tmp43 = tmp17 - tmp40
tmp44 = tmp43 * tmp43
tmp45 = tmp42 + tmp44
tmp46 = tmp27 - tmp40
tmp47 = tmp46 * tmp46
tmp48 = tmp45 + tmp47
tmp49 = tmp37 - tmp40
tmp50 = tmp49 * tmp49
tmp51 = tmp48 + tmp50
tmp52 = tmp51 / tmp39
tmp53 = tmp6 + tmp15
tmp54 = tmp53 + tmp25
tmp55 = tmp54 + tmp35
tmp56 = tmp55 / tmp39
tmp57 = tmp6 - tmp56
tmp58 = tmp57 * tmp57
tmp59 = tmp15 - tmp56
tmp60 = tmp59 * tmp59
tmp61 = tmp58 + tmp60
tmp62 = tmp25 - tmp56
tmp63 = tmp62 * tmp62
tmp64 = tmp61 + tmp63
tmp65 = tmp35 - tmp56
tmp66 = tmp65 * tmp65
tmp67 = tmp64 + tmp66
tmp68 = tmp67 / tmp39
tl.store(out_ptr0 + (x0), tmp40, xmask)
tl.store(out_ptr1 + (x0), tmp52, xmask)
tl.store(out_ptr2 + (x0), tmp56, xmask)
tl.store(out_ptr3 + (x0), tmp68, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wr/cwrax6jmtc4cdewawrgq5qodxervs2qo3z2th4qqtdt4jjlavqes.py
# Topologically Sorted Source Nodes: [mul_2, mul_3, new_c, tanh, new_h, new_h_2, new_c_1], Original ATen: [aten.mul, aten.add, aten.tanh, aten.native_layer_norm]
# Source node to ATen node mapping:
# mul_2 => mul_2
# mul_3 => mul_3
# new_c => add_2
# new_c_1 => add_5, add_6, mul_7, mul_8, rsqrt_1, sub_1
# new_h => mul_4
# new_h_2 => add_3, add_4, mul_5, mul_6, rsqrt, sub
# tanh => tanh
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %primals_18), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_18), kwargs = {})
# %add_2 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %mul_3), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_2,), kwargs = {})
# %mul_4 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_3, %tanh), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_4, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_3,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_4, %getitem_5), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_5, %primals_19), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_6, %primals_20), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_6, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_5,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_2, %getitem_7), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %rsqrt_1), kwargs = {})
# %mul_8 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_7, %primals_21), kwargs = {})
# %add_6 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_8, %primals_22), kwargs = {})
triton_poi_fused_add_mul_native_layer_norm_tanh_4 = async_compile.triton('triton_poi_fused_add_mul_native_layer_norm_tanh_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_native_layer_norm_tanh_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_native_layer_norm_tanh_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10, in_ptr11, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (x2), xmask)
tmp4 = tl.load(in_ptr3 + (x2), xmask)
tmp9 = tl.load(in_ptr4 + (x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr5 + (x1), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr6 + (x0), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr7 + (x0), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr8 + (x1), xmask, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr9 + (x1), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr10 + (x0), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr11 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp5 = tmp4 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = libdevice.tanh(tmp6)
tmp8 = tmp0 * tmp7
tmp10 = tmp8 - tmp9
tmp12 = 1e-05
tmp13 = tmp11 + tmp12
tmp14 = libdevice.rsqrt(tmp13)
tmp15 = tmp10 * tmp14
tmp17 = tmp15 * tmp16
tmp19 = tmp17 + tmp18
tmp21 = tmp6 - tmp20
tmp23 = tmp22 + tmp12
tmp24 = libdevice.rsqrt(tmp23)
tmp25 = tmp21 * tmp24
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tl.store(out_ptr0 + (x2), tmp19, xmask)
tl.store(out_ptr1 + (x2), tmp29, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (16, 4), (4, 1))
assert_size_stride(primals_9, (16, 4), (4, 1))
assert_size_stride(primals_10, (16, 4), (4, 1))
assert_size_stride(primals_11, (16, ), (1, ))
assert_size_stride(primals_12, (16, 4), (4, 1))
assert_size_stride(primals_13, (16, ), (1, ))
assert_size_stride(primals_14, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_15, (16, 4), (4, 1))
assert_size_stride(primals_16, (16, ), (1, ))
assert_size_stride(primals_17, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_18, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_19, (4, ), (1, ))
assert_size_stride(primals_20, (4, ), (1, ))
assert_size_stride(primals_21, (4, ), (1, ))
assert_size_stride(primals_22, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [z_h], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [z_x], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [z_bias], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [d_z_h], Original ATen: [aten.mm]
extern_kernels.mm(buf0, reinterpret_tensor(primals_8, (4, 16), (1, 4), 0), out=buf3)
buf4 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [d_z_x], Original ATen: [aten.mm]
extern_kernels.mm(buf1, reinterpret_tensor(primals_9, (4, 16), (1, 4), 0), out=buf4)
buf5 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf2, reinterpret_tensor(primals_10, (4, 16), (1, 4), 0), out=buf5)
buf6 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_6], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_13, reinterpret_tensor(primals_14, (64, 4), (4, 1), 0), reinterpret_tensor(primals_12, (4, 16), (1, 4), 0), alpha=1, beta=1, out=buf6)
del primals_12
del primals_13
buf7 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_7], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_16, reinterpret_tensor(primals_17, (64, 4), (4, 1), 0), reinterpret_tensor(primals_15, (4, 16), (1, 4), 0), alpha=1, beta=1, out=buf7)
del primals_15
del primals_16
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_1], Original ATen: [aten.sigmoid]
stream0 = get_raw_stream(0)
triton_poi_fused_sigmoid_0.run(buf3, buf6, buf4, buf7, buf5, primals_11, buf8, 256, grid=grid(256), stream=stream0)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [f_1], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_1.run(buf3, buf6, buf4, buf7, buf5, primals_11, buf9, 256, grid=grid(256), stream=stream0)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [o_1], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_2.run(buf3, buf6, buf4, buf7, buf5, primals_11, buf10, 256, grid=grid(256), stream=stream0)
del buf5
del primals_11
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf12 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf14 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, mul_3, new_c, tanh, new_h, new_h_2, new_c_1], Original ATen: [aten.mul, aten.add, aten.tanh, aten.native_layer_norm]
triton_poi_fused_add_mul_native_layer_norm_tanh_3.run(buf10, buf9, primals_18, buf8, buf11, buf12, buf14, buf15, 64, grid=grid(64), stream=stream0)
buf13 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf16 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, mul_3, new_c, tanh, new_h, new_h_2, new_c_1], Original ATen: [aten.mul, aten.add, aten.tanh, aten.native_layer_norm]
triton_poi_fused_add_mul_native_layer_norm_tanh_4.run(buf10, buf9, primals_18, buf8, buf11, buf12, primals_19, primals_20, buf14, buf15, primals_21, primals_22, buf13, buf16, 256, grid=grid(256), stream=stream0)
del buf11
del buf12
del buf14
del buf15
del primals_20
del primals_22
return (buf13, buf16, primals_18, primals_19, primals_21, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf0, buf3, buf1, buf4, buf2, reinterpret_tensor(primals_14, (64, 4), (4, 1), 0), buf6, reinterpret_tensor(primals_17, (64, 4), (4, 1), 0), buf7, buf8, buf9, buf10, primals_10, primals_9, primals_8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class InferenceNetLSTMCell(nn.Module):
def __init__(self, z_dim: 'int', input_dim: 'int', hidden_hat_dim:
'int', hidden_dim: 'int'):
super(InferenceNetLSTMCell, self).__init__()
self.w_hh = nn.Linear(hidden_hat_dim, z_dim)
self.w_hx = nn.Linear(hidden_hat_dim, z_dim)
self.w_hb = nn.Linear(hidden_hat_dim, z_dim)
self.W_hz = nn.Linear(z_dim, 4 * hidden_dim, bias=False)
self.W_xz = nn.Linear(z_dim, 4 * hidden_dim, bias=False)
self.b = nn.Linear(z_dim, 4 * hidden_dim)
self.Wh = nn.Linear(hidden_dim, 4 * hidden_dim)
self.Wx = nn.Linear(input_dim, 4 * hidden_dim)
self.dropout = nn.Dropout(p=0.1, inplace=True)
self.norm_h = nn.LayerNorm(hidden_dim)
self.norm_c = nn.LayerNorm(hidden_dim)
def forward(self, h_t, c, h_t_hat, inf_inputs):
z_h = self.w_hh(h_t_hat)
z_x = self.w_hx(h_t_hat)
z_bias = self.w_hb(h_t_hat)
d_z_h = self.W_hz(z_h)
d_z_x = self.W_xz(z_x)
b_z_b = self.b(z_bias)
ifgo = d_z_h * self.Wh(h_t) + d_z_x * self.Wx(inf_inputs) + b_z_b
i, f, g, o = torch.chunk(ifgo, 4, -1)
i = torch.sigmoid(i)
f = torch.sigmoid(f)
g = torch.sigmoid(g)
o = torch.sigmoid(o)
new_c = f * c + i * c
new_h = o * torch.tanh(new_c)
new_h = self.dropout(new_h)
new_h = self.norm_h(new_h)
new_c = self.norm_c(new_c)
return new_h, new_c
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'z_dim': 4, 'input_dim': 4, 'hidden_hat_dim': 4,
'hidden_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_sigmoid_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + 16 * x1), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + (x0 + 16 * x1), xmask)
tmp7 = tl.load(in_ptr4 + (x0 + 16 * x1), xmask)
tmp8 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + x2, tmp11, xmask)
@triton.jit
def triton_poi_fused_sigmoid_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (4 + x0 + 16 * x1), xmask)
tmp3 = tl.load(in_ptr2 + (4 + x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + (4 + x0 + 16 * x1), xmask)
tmp7 = tl.load(in_ptr4 + (4 + x0 + 16 * x1), xmask)
tmp8 = tl.load(in_ptr5 + (4 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + x2, tmp11, xmask)
@triton.jit
def triton_poi_fused_sigmoid_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (12 + x0 + 16 * x1), xmask)
tmp3 = tl.load(in_ptr2 + (12 + x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + (12 + x0 + 16 * x1), xmask)
tmp7 = tl.load(in_ptr4 + (12 + x0 + 16 * x1), xmask)
tmp8 = tl.load(in_ptr5 + (12 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp11 = tl.sigmoid(tmp10)
tl.store(out_ptr0 + x2, tmp11, xmask)
@triton.jit
def triton_poi_fused_add_mul_native_layer_norm_tanh_3(in_ptr0, in_ptr1,
in_ptr2, in_ptr3, out_ptr0, out_ptr1, out_ptr2, out_ptr3, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + 4 * x0, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + 4 * x0, xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp11 = tl.load(in_ptr2 + (1 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp13 = tl.load(in_ptr3 + (1 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp19 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp20 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp21 = tl.load(in_ptr2 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp23 = tl.load(in_ptr3 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp29 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp30 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp31 = tl.load(in_ptr2 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp33 = tl.load(in_ptr3 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp3 = tmp1 * tmp2
tmp5 = tmp4 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = libdevice.tanh(tmp6)
tmp8 = tmp0 * tmp7
tmp12 = tmp10 * tmp11
tmp14 = tmp13 * tmp11
tmp15 = tmp12 + tmp14
tmp16 = libdevice.tanh(tmp15)
tmp17 = tmp9 * tmp16
tmp18 = tmp8 + tmp17
tmp22 = tmp20 * tmp21
tmp24 = tmp23 * tmp21
tmp25 = tmp22 + tmp24
tmp26 = libdevice.tanh(tmp25)
tmp27 = tmp19 * tmp26
tmp28 = tmp18 + tmp27
tmp32 = tmp30 * tmp31
tmp34 = tmp33 * tmp31
tmp35 = tmp32 + tmp34
tmp36 = libdevice.tanh(tmp35)
tmp37 = tmp29 * tmp36
tmp38 = tmp28 + tmp37
tmp39 = 4.0
tmp40 = tmp38 / tmp39
tmp41 = tmp8 - tmp40
tmp42 = tmp41 * tmp41
tmp43 = tmp17 - tmp40
tmp44 = tmp43 * tmp43
tmp45 = tmp42 + tmp44
tmp46 = tmp27 - tmp40
tmp47 = tmp46 * tmp46
tmp48 = tmp45 + tmp47
tmp49 = tmp37 - tmp40
tmp50 = tmp49 * tmp49
tmp51 = tmp48 + tmp50
tmp52 = tmp51 / tmp39
tmp53 = tmp6 + tmp15
tmp54 = tmp53 + tmp25
tmp55 = tmp54 + tmp35
tmp56 = tmp55 / tmp39
tmp57 = tmp6 - tmp56
tmp58 = tmp57 * tmp57
tmp59 = tmp15 - tmp56
tmp60 = tmp59 * tmp59
tmp61 = tmp58 + tmp60
tmp62 = tmp25 - tmp56
tmp63 = tmp62 * tmp62
tmp64 = tmp61 + tmp63
tmp65 = tmp35 - tmp56
tmp66 = tmp65 * tmp65
tmp67 = tmp64 + tmp66
tmp68 = tmp67 / tmp39
tl.store(out_ptr0 + x0, tmp40, xmask)
tl.store(out_ptr1 + x0, tmp52, xmask)
tl.store(out_ptr2 + x0, tmp56, xmask)
tl.store(out_ptr3 + x0, tmp68, xmask)
@triton.jit
def triton_poi_fused_add_mul_native_layer_norm_tanh_4(in_ptr0, in_ptr1,
in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9,
in_ptr10, in_ptr11, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + x2, xmask)
tmp4 = tl.load(in_ptr3 + x2, xmask)
tmp9 = tl.load(in_ptr4 + x1, xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr5 + x1, xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr6 + x0, xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr7 + x0, xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr8 + x1, xmask, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr9 + x1, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr10 + x0, xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr11 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp5 = tmp4 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = libdevice.tanh(tmp6)
tmp8 = tmp0 * tmp7
tmp10 = tmp8 - tmp9
tmp12 = 1e-05
tmp13 = tmp11 + tmp12
tmp14 = libdevice.rsqrt(tmp13)
tmp15 = tmp10 * tmp14
tmp17 = tmp15 * tmp16
tmp19 = tmp17 + tmp18
tmp21 = tmp6 - tmp20
tmp23 = tmp22 + tmp12
tmp24 = libdevice.rsqrt(tmp23)
tmp25 = tmp21 * tmp24
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tl.store(out_ptr0 + x2, tmp19, xmask)
tl.store(out_ptr1 + x2, tmp29, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (16, 4), (4, 1))
assert_size_stride(primals_9, (16, 4), (4, 1))
assert_size_stride(primals_10, (16, 4), (4, 1))
assert_size_stride(primals_11, (16,), (1,))
assert_size_stride(primals_12, (16, 4), (4, 1))
assert_size_stride(primals_13, (16,), (1,))
assert_size_stride(primals_14, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_15, (16, 4), (4, 1))
assert_size_stride(primals_16, (16,), (1,))
assert_size_stride(primals_17, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_18, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_19, (4,), (1,))
assert_size_stride(primals_20, (4,), (1,))
assert_size_stride(primals_21, (4,), (1,))
assert_size_stride(primals_22, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.mm(buf0, reinterpret_tensor(primals_8, (4, 16), (1,
4), 0), out=buf3)
buf4 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_9, (4, 16), (1,
4), 0), out=buf4)
buf5 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.mm(buf2, reinterpret_tensor(primals_10, (4, 16), (1,
4), 0), out=buf5)
buf6 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.addmm(primals_13, reinterpret_tensor(primals_14, (64,
4), (4, 1), 0), reinterpret_tensor(primals_12, (4, 16), (1, 4),
0), alpha=1, beta=1, out=buf6)
del primals_12
del primals_13
buf7 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.addmm(primals_16, reinterpret_tensor(primals_17, (64,
4), (4, 1), 0), reinterpret_tensor(primals_15, (4, 16), (1, 4),
0), alpha=1, beta=1, out=buf7)
del primals_15
del primals_16
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_sigmoid_0[grid(256)](buf3, buf6, buf4, buf7, buf5,
primals_11, buf8, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_sigmoid_1[grid(256)](buf3, buf6, buf4, buf7, buf5,
primals_11, buf9, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_sigmoid_2[grid(256)](buf3, buf6, buf4, buf7, buf5,
primals_11, buf10, 256, XBLOCK=256, num_warps=4, num_stages=1)
del buf5
del primals_11
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf12 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf14 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused_add_mul_native_layer_norm_tanh_3[grid(64)](buf10,
buf9, primals_18, buf8, buf11, buf12, buf14, buf15, 64, XBLOCK=
64, num_warps=1, num_stages=1)
buf13 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf16 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_native_layer_norm_tanh_4[grid(256)](buf10,
buf9, primals_18, buf8, buf11, buf12, primals_19, primals_20,
buf14, buf15, primals_21, primals_22, buf13, buf16, 256, XBLOCK
=256, num_warps=4, num_stages=1)
del buf11
del buf12
del buf14
del buf15
del primals_20
del primals_22
return (buf13, buf16, primals_18, primals_19, primals_21,
reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf0, buf3, buf1,
buf4, buf2, reinterpret_tensor(primals_14, (64, 4), (4, 1), 0),
buf6, reinterpret_tensor(primals_17, (64, 4), (4, 1), 0), buf7,
buf8, buf9, buf10, primals_10, primals_9, primals_8)
class InferenceNetLSTMCellNew(nn.Module):
def __init__(self, z_dim: 'int', input_dim: 'int', hidden_hat_dim:
'int', hidden_dim: 'int'):
super(InferenceNetLSTMCellNew, self).__init__()
self.w_hh = nn.Linear(hidden_hat_dim, z_dim)
self.w_hx = nn.Linear(hidden_hat_dim, z_dim)
self.w_hb = nn.Linear(hidden_hat_dim, z_dim)
self.W_hz = nn.Linear(z_dim, 4 * hidden_dim, bias=False)
self.W_xz = nn.Linear(z_dim, 4 * hidden_dim, bias=False)
self.b = nn.Linear(z_dim, 4 * hidden_dim)
self.Wh = nn.Linear(hidden_dim, 4 * hidden_dim)
self.Wx = nn.Linear(input_dim, 4 * hidden_dim)
self.dropout = nn.Dropout(p=0.1, inplace=True)
self.norm_h = nn.LayerNorm(hidden_dim)
self.norm_c = nn.LayerNorm(hidden_dim)
def forward(self, input_0, input_1, input_2, input_3):
primals_1 = self.w_hh.weight
primals_2 = self.w_hh.bias
primals_4 = self.w_hx.weight
primals_5 = self.w_hx.bias
primals_6 = self.w_hb.weight
primals_7 = self.w_hb.bias
primals_8 = self.W_hz.weight
primals_9 = self.W_xz.weight
primals_10 = self.b.weight
primals_11 = self.b.bias
primals_12 = self.Wh.weight
primals_13 = self.Wh.bias
primals_15 = self.Wx.weight
primals_16 = self.Wx.bias
primals_19 = self.norm_h.weight
primals_20 = self.norm_h.bias
primals_21 = self.norm_c.weight
primals_22 = self.norm_c.bias
primals_3 = input_0
primals_14 = input_1
primals_17 = input_2
primals_18 = input_3
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22])
return output[0], output[1]
|
kingofpigeon/hypernlp
|
InferenceNetLSTMCell
| false | 10,401 |
[
"MIT"
] | 0 |
1270ae318e698775160a6299db35752823fda7c7
|
https://github.com/kingofpigeon/hypernlp/tree/1270ae318e698775160a6299db35752823fda7c7
|
MinMaxNorm
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/32/c32i2334e2cubbxcfo3qfe3hcxrbbjwe3jyvzl2kvxsr6gdadsdu.py
# Topologically Sorted Source Nodes: [sub, mul, truediv, add], Original ATen: [aten.sub, aten.mul, aten.div, aten.add]
# Source node to ATen node mapping:
# add => add
# mul => mul
# sub => sub
# truediv => div
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, 4), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, 1), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul, 0), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div, 0), kwargs = {})
triton_poi_fused_add_div_mul_sub_0 = async_compile.triton('triton_poi_fused_add_div_mul_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_mul_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_mul_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 4.0
tmp2 = tmp0 - tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp5 = float("inf")
tmp6 = tmp4 * tmp5
tmp7 = 0.0
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, mul, truediv, add], Original ATen: [aten.sub, aten.mul, aten.div, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_mul_sub_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class MinMaxNorm(nn.Module):
def __init__(self, min, max, a=0, b=1):
super(MinMaxNorm, self).__init__()
self.min, self.max = min, max
self.a, self.b = a, b
def forward(self, x):
return self.a + (x - self.min) * (self.b - self.a) / (self.max -
self.min)
def inverse(self, x):
return self.min + (x - self.a) * (self.max - self.min) / (self.b -
self.a)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'min': 4, 'max': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_div_mul_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 4.0
tmp2 = tmp0 - tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp5 = float('inf')
tmp6 = tmp4 * tmp5
tmp7 = 0.0
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_mul_sub_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class MinMaxNormNew(nn.Module):
def __init__(self, min, max, a=0, b=1):
super(MinMaxNormNew, self).__init__()
self.min, self.max = min, max
self.a, self.b = a, b
def inverse(self, x):
return self.min + (x - self.a) * (self.max - self.min) / (self.b -
self.a)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
iclementine/speedyspeech
|
MinMaxNorm
| false | 10,402 |
[
"BSD-3-Clause"
] | 0 |
db527587a3699b71082d61c9e9fad7ed795d1980
|
https://github.com/iclementine/speedyspeech/tree/db527587a3699b71082d61c9e9fad7ed795d1980
|
CCAMDec
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/3m/c3mxgkf4weymbmbgydi4j4i6eycdz2flzbf3jce3eapte2aqyfta.py
# Topologically Sorted Source Nodes: [energy_new], Original ATen: [aten.sub]
# Source node to ATen node mapping:
# energy_new => sub
# Graph fragment:
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%expand, %bmm), kwargs = {})
triton_poi_fused_sub_0 = async_compile.triton('triton_poi_fused_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (x2), xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp8 = tmp6 - tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hz/chzi3aam26mikdhljz5x7jlqazm7kpktzeptsf36thgfhsg7ub6a.py
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention => amax, exp, sub_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%sub, [-1], True), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_1,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/em/cem6qbxwbiqnjqybzk5arf2obt5uggy4qs7otwwpovvnrhvdc6h4.py
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/5e/c5e7z5qmoiqut4wygb4iv6xmv65bbiotnb64o5cgidinohzcyout.py
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
# Source node to ATen node mapping:
# mul => mul
# out_2 => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, %view_3), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_1, %mul), kwargs = {})
triton_poi_fused_add_mul_3 = async_compile.triton('triton_poi_fused_add_mul_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tl.load(in_ptr2 + (x0), xmask)
tmp4 = tmp2 * tmp3
tmp5 = tmp0 + tmp4
tl.store(out_ptr0 + (x0), tmp5, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(primals_1, (4, 4, 16), (64, 16, 1), 0), reinterpret_tensor(primals_2, (4, 16, 4), (64, 1, 16), 0), out=buf0)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy_new], Original ATen: [aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_sub_0.run(buf0, buf1, 64, grid=grid(64), stream=stream0)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf2, buf3, 64, grid=grid(64), stream=stream0)
del buf2
buf4 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention, out], Original ATen: [aten._softmax, aten.bmm]
extern_kernels.bmm(buf3, reinterpret_tensor(primals_2, (4, 4, 16), (64, 16, 1), 0), out=buf4)
del buf3
del primals_2
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
triton_poi_fused_add_mul_3.run(primals_1, primals_3, buf4, buf5, 256, grid=grid(256), stream=stream0)
del primals_1
del primals_3
return (buf5, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn.parameter import Parameter
class CCAMDec(Module):
"""
CCAM decoding module
"""
def __init__(self):
super(CCAMDec, self).__init__()
self.softmax = Softmax(dim=-1)
self.scale = Parameter(torch.zeros(1))
def forward(self, x, y):
"""
inputs :
x : input feature(N,C,H,W) y:gathering centers(N,K,H,W)
returns :
out : compact channel attention feature
attention map: K*C
"""
m_batchsize, C, width, height = x.size()
x_reshape = x.view(m_batchsize, C, -1)
B, K, _W, _H = y.size()
y_reshape = y.view(B, K, -1)
proj_query = x_reshape
proj_key = y_reshape.permute(0, 2, 1)
energy = torch.bmm(proj_query, proj_key)
energy_new = torch.max(energy, -1, keepdim=True)[0].expand_as(energy
) - energy
attention = self.softmax(energy_new)
proj_value = y.view(B, K, -1)
out = torch.bmm(attention, proj_value)
out = out.view(m_batchsize, C, width, height)
out = x + self.scale * out
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.nn import Module
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn.parameter import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + x2, xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp8 = tmp6 - tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tl.load(in_ptr2 + x0, xmask)
tmp4 = tmp2 * tmp3
tmp5 = tmp0 + tmp4
tl.store(out_ptr0 + x0, tmp5, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(primals_1, (4, 4, 16), (64,
16, 1), 0), reinterpret_tensor(primals_2, (4, 16, 4), (64, 1,
16), 0), out=buf0)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_sub_0[grid(64)](buf0, buf1, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf2 = buf0
del buf0
triton_poi_fused__softmax_1[grid(64)](buf1, buf2, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf3 = buf1
del buf1
triton_poi_fused__softmax_2[grid(64)](buf2, buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del buf2
buf4 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
extern_kernels.bmm(buf3, reinterpret_tensor(primals_2, (4, 4, 16),
(64, 16, 1), 0), out=buf4)
del buf3
del primals_2
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_3[grid(256)](primals_1, primals_3, buf4,
buf5, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_3
return buf5, buf4
class CCAMDecNew(Module):
"""
CCAM decoding module
"""
def __init__(self):
super(CCAMDecNew, self).__init__()
self.softmax = Softmax(dim=-1)
self.scale = Parameter(torch.zeros(1))
def forward(self, input_0, input_1):
primals_3 = self.scale
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3])
return output[0]
|
bfjei2825401/siamban
|
CCAMDec
| false | 10,403 |
[
"Apache-2.0"
] | 0 |
c41d58742b146dfc8960053453227c6e9fec1bac
|
https://github.com/bfjei2825401/siamban/tree/c41d58742b146dfc8960053453227c6e9fec1bac
|
PAM_Module
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/sr/csrg6irduolxnaubd5v3tlh5eeuhw27sxkg3o56t4veh47sq6ce3.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 2
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ko/ckow7ci7f3mygm6ujdzdisip6tet25h4hj6uestesqalhkarwrrw.py
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention => amax, div, exp, sub, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_per_fused__softmax_1 = async_compile.triton('triton_per_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[64, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__softmax_1(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, float("-inf"))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + (16*x0)), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/w5/cw5gytijzzkwnfpq2a2axdsj4pfxgxmwiuzizuyd4bw5uwnanzw7.py
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/j4/cj4f6qdb45emg4zrdv5vzxtw2vswpyt2rqyalr6mxgomzeyk55j5.py
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
# Source node to ATen node mapping:
# mul => mul
# out_2 => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_8, %view_3), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %primals_1), kwargs = {})
triton_poi_fused_add_mul_3 = async_compile.triton('triton_poi_fused_add_mul_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp2 = tl.load(in_ptr1 + (x0), xmask)
tmp4 = tl.load(in_ptr2 + (x0), xmask)
tmp3 = tmp1 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + (x0), tmp5, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (2, ), (1, ))
assert_size_stride(primals_4, (2, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (2, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 2, 4, 4), (32, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_3, 128, grid=grid(128), stream=stream0)
del primals_3
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_1, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 2, 4, 4), (32, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf3, primals_5, 128, grid=grid(128), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf1, (4, 16, 2), (32, 1, 16), 0), reinterpret_tensor(buf3, (4, 2, 16), (32, 16, 1), 0), out=buf4)
buf7 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
triton_per_fused__softmax_1.run(buf4, buf7, 64, 16, grid=grid(64), stream=stream0)
del buf4
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf8 = extern_kernels.convolution(primals_1, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 4, 4, 4), (64, 16, 4, 1))
buf9 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf9, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
buf10 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf9, (4, 4, 16), (64, 16, 1), 0), reinterpret_tensor(buf7, (4, 16, 16), (256, 1, 16), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
triton_poi_fused_add_mul_3.run(primals_8, buf10, primals_1, buf11, 256, grid=grid(256), stream=stream0)
return (buf11, primals_1, primals_2, primals_4, primals_6, primals_8, buf7, buf10, reinterpret_tensor(buf9, (4, 16, 4), (64, 1, 16), 0), reinterpret_tensor(buf1, (4, 2, 16), (32, 16, 1), 0), reinterpret_tensor(buf3, (4, 16, 2), (32, 1, 16), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((2, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((2, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((2, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((2, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
from torch.nn import Conv2d
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn.parameter import Parameter
class PAM_Module(Module):
""" Position attention module"""
def __init__(self, in_dim):
super(PAM_Module, self).__init__()
self.channel_in = in_dim
out_channels = max(in_dim // 8, min(in_dim, 2))
self.query_conv = Conv2d(in_channels=in_dim, out_channels=
out_channels, kernel_size=1)
self.key_conv = Conv2d(in_channels=in_dim, out_channels=
out_channels, kernel_size=1)
self.value_conv = Conv2d(in_channels=in_dim, out_channels=in_dim,
kernel_size=1)
self.gamma = Parameter(torch.zeros(1))
self.softmax = Softmax(dim=-1)
def forward(self, x):
"""
inputs :
x : input feature maps( B X C X H X W)
returns :
out : attention value + input feature
attention: B X (HxW) X (HxW)
"""
m_batchsize, C, height, width = x.size()
proj_query = self.query_conv(x).view(m_batchsize, -1, width * height
).permute(0, 2, 1)
proj_key = self.key_conv(x).view(m_batchsize, -1, width * height)
energy = torch.bmm(proj_query, proj_key)
attention = self.softmax(energy)
proj_value = self.value_conv(x).view(m_batchsize, -1, width * height)
out = torch.bmm(proj_value, attention.permute(0, 2, 1))
out = out.view(m_batchsize, C, height, width)
out = self.gamma * out + x
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.nn import Module
from torch.nn import Conv2d
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn.parameter import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 2
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_per_fused__softmax_1(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, float('-inf'))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + 16 * x0), tmp11, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp2 = tl.load(in_ptr1 + x0, xmask)
tmp4 = tl.load(in_ptr2 + x0, xmask)
tmp3 = tmp1 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + x0, tmp5, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (2,), (1,))
assert_size_stride(primals_4, (2, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (2,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 2, 4, 4), (32, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(128)](buf1, primals_3, 128,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
buf2 = extern_kernels.convolution(primals_1, primals_4, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 2, 4, 4), (32, 16, 4, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_0[grid(128)](buf3, primals_5, 128,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf1, (4, 16, 2), (32, 1, 16),
0), reinterpret_tensor(buf3, (4, 2, 16), (32, 16, 1), 0), out=buf4)
buf7 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
triton_per_fused__softmax_1[grid(64)](buf4, buf7, 64, 16, XBLOCK=8,
num_warps=2, num_stages=1)
del buf4
buf8 = extern_kernels.convolution(primals_1, primals_6, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 4, 4, 4), (64, 16, 4, 1))
buf9 = buf8
del buf8
triton_poi_fused_convolution_2[grid(256)](buf9, primals_7, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_7
buf10 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf9, (4, 4, 16), (64, 16, 1),
0), reinterpret_tensor(buf7, (4, 16, 16), (256, 1, 16), 0), out
=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_3[grid(256)](primals_8, buf10, primals_1,
buf11, 256, XBLOCK=256, num_warps=4, num_stages=1)
return (buf11, primals_1, primals_2, primals_4, primals_6, primals_8,
buf7, buf10, reinterpret_tensor(buf9, (4, 16, 4), (64, 1, 16), 0),
reinterpret_tensor(buf1, (4, 2, 16), (32, 16, 1), 0),
reinterpret_tensor(buf3, (4, 16, 2), (32, 1, 16), 0))
class PAM_ModuleNew(Module):
""" Position attention module"""
def __init__(self, in_dim):
super(PAM_ModuleNew, self).__init__()
self.channel_in = in_dim
out_channels = max(in_dim // 8, min(in_dim, 2))
self.query_conv = Conv2d(in_channels=in_dim, out_channels=
out_channels, kernel_size=1)
self.key_conv = Conv2d(in_channels=in_dim, out_channels=
out_channels, kernel_size=1)
self.value_conv = Conv2d(in_channels=in_dim, out_channels=in_dim,
kernel_size=1)
self.gamma = Parameter(torch.zeros(1))
self.softmax = Softmax(dim=-1)
def forward(self, input_0):
primals_8 = self.gamma
primals_2 = self.query_conv.weight
primals_3 = self.query_conv.bias
primals_4 = self.key_conv.weight
primals_5 = self.key_conv.bias
primals_6 = self.value_conv.weight
primals_7 = self.value_conv.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8])
return output[0]
|
bfjei2825401/siamban
|
PAM_Module
| false | 10,404 |
[
"Apache-2.0"
] | 0 |
c41d58742b146dfc8960053453227c6e9fec1bac
|
https://github.com/bfjei2825401/siamban/tree/c41d58742b146dfc8960053453227c6e9fec1bac
|
Encoder
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/dk/cdkfylhbmav6z7nn6vy4nwynu3ry6ilbxpjspabpseiuux7csyql.py
# Topologically Sorted Source Nodes: [input_3], Original ATen: [aten.native_group_norm]
# Source node to ATen node mapping:
# input_3 => add, rsqrt, var_mean
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
triton_poi_fused_native_group_norm_0 = async_compile.triton('triton_poi_fused_native_group_norm_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_group_norm_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_group_norm_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp5 = tmp2 + tmp4
tmp7 = triton_helpers.maximum(tmp1, tmp6)
tmp8 = tmp5 + tmp7
tmp10 = triton_helpers.maximum(tmp1, tmp9)
tmp11 = tmp8 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tmp2 - tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp4 - tmp13
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp7 - tmp13
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp10 - tmp13
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp24 / tmp12
tmp26 = 1e-05
tmp27 = tmp25 + tmp26
tmp28 = libdevice.rsqrt(tmp27)
tl.store(out_ptr0 + (x0), tmp13, xmask)
tl.store(out_ptr1 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/kd/ckdswb4lufwiij2wklkypvchblxzpif4ph4uxoqoklvcbijmlski.py
# Topologically Sorted Source Nodes: [input_3], Original ATen: [aten.native_group_norm]
# Source node to ATen node mapping:
# input_3 => add_1, mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %unsqueeze_7), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %unsqueeze_4), kwargs = {})
triton_poi_fused_native_group_norm_1 = async_compile.triton('triton_poi_fused_native_group_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_group_norm_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_group_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = (xindex // 4)
x1 = (xindex // 4) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp3 = tl.load(in_ptr1 + (x4), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr2 + (x4), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr4 + (x1), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = tmp2 - tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 * tmp7
tmp10 = tmp8 + tmp9
tl.store(out_ptr0 + (x3), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 8, 4, 4), (128, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.max_pool3d_with_indices]
buf0 = torch.ops.aten.max_pool3d_with_indices.default(primals_1, [2, 2, 2], [2, 2, 2])
del primals_1
buf1 = buf0[0]
del buf0
# Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(reinterpret_tensor(buf1, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1), 0), primals_2, stride=(1, 1, 1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False, output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf3, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1))
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_3], Original ATen: [aten.native_group_norm]
stream0 = get_raw_stream(0)
triton_poi_fused_native_group_norm_0.run(buf3, buf4, buf5, 16, grid=grid(16), stream=stream0)
buf6 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_3], Original ATen: [aten.native_group_norm]
triton_poi_fused_native_group_norm_1.run(buf3, buf4, buf5, primals_3, primals_4, buf6, 64, grid=grid(64), stream=stream0)
del primals_4
# Topologically Sorted Source Nodes: [input_4], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(reinterpret_tensor(buf6, (1, 4, 4, 2, 2), (0, 16, 4, 2, 1), 0), primals_5, stride=(1, 1, 1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False, output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf7, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1))
buf8 = buf5; del buf5 # reuse
buf9 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [input_6], Original ATen: [aten.native_group_norm]
triton_poi_fused_native_group_norm_0.run(buf7, buf8, buf9, 16, grid=grid(16), stream=stream0)
buf10 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_6], Original ATen: [aten.native_group_norm]
triton_poi_fused_native_group_norm_1.run(buf7, buf8, buf9, primals_6, primals_7, buf10, 64, grid=grid(64), stream=stream0)
del buf8
del buf9
del primals_7
return (buf10, primals_2, primals_3, primals_5, primals_6, reinterpret_tensor(buf1, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1), 0), buf3, reinterpret_tensor(buf6, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1), 0), buf7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 8, 4, 4), (128, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3, 3, 3), (108, 27, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 3, 3, 3), (108, 27, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
def conv3d(in_channels, out_channels, kernel_size, bias, padding=1):
return nn.Conv3d(in_channels, out_channels, kernel_size, padding=
padding, bias=bias)
def create_conv(in_channels, out_channels, kernel_size, order, num_groups,
padding=1):
"""
Create a list of modules with together constitute a single conv layer with non-linearity
and optional batchnorm/groupnorm.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
order (string): order of things, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
padding (int): add zero-padding to the input
Return:
list of tuple (name, module)
"""
assert 'c' in order, 'Conv layer MUST be present'
assert order[0
] not in 'rle', 'Non-linearity cannot be the first operation in the layer'
modules = []
for i, char in enumerate(order):
if char == 'r':
modules.append(('ReLU', nn.ReLU(inplace=True)))
elif char == 'l':
modules.append(('LeakyReLU', nn.LeakyReLU(negative_slope=0.1,
inplace=True)))
elif char == 'e':
modules.append(('ELU', nn.ELU(inplace=True)))
elif char == 'c':
bias = not ('g' in order or 'b' in order)
modules.append(('conv', conv3d(in_channels, out_channels,
kernel_size, bias, padding=padding)))
elif char == 'g':
is_before_conv = i < order.index('c')
assert not is_before_conv, 'GroupNorm MUST go after the Conv3d'
if out_channels < num_groups:
num_groups = out_channels
modules.append(('groupnorm', nn.GroupNorm(num_groups=num_groups,
num_channels=out_channels)))
elif char == 'b':
is_before_conv = i < order.index('c')
if is_before_conv:
modules.append(('batchnorm', nn.BatchNorm3d(in_channels)))
else:
modules.append(('batchnorm', nn.BatchNorm3d(out_channels)))
else:
raise ValueError(
f"Unsupported layer type '{char}'. MUST be one of ['b', 'g', 'r', 'l', 'e', 'c']"
)
return modules
class SingleConv(nn.Sequential):
"""
Basic convolutional module consisting of a Conv3d, non-linearity and optional batchnorm/groupnorm. The order
of operations can be specified via the `order` parameter
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'crg', num_groups=8, padding=1):
super(SingleConv, self).__init__()
for name, module in create_conv(in_channels, out_channels,
kernel_size, order, num_groups, padding=padding):
self.add_module(name, module)
class DoubleConv(nn.Sequential):
"""
A module consisting of two consecutive convolution layers (e.g. BatchNorm3d+ReLU+Conv3d).
We use (Conv3d+ReLU+GroupNorm3d) by default.
This can be changed however by providing the 'order' argument, e.g. in order
to change to Conv3d+BatchNorm3d+ELU use order='cbe'.
Use padded convolutions to make sure that the output (H_out, W_out) is the same
as (H_in, W_in), so that you don't have to crop in the decoder path.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
encoder (bool): if True we're in the encoder path, otherwise we're in the decoder
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, encoder, kernel_size=3,
order='crg', num_groups=8):
super(DoubleConv, self).__init__()
if encoder:
conv1_in_channels = in_channels
conv1_out_channels = out_channels // 2
if conv1_out_channels < in_channels:
conv1_out_channels = in_channels
conv2_in_channels, conv2_out_channels = (conv1_out_channels,
out_channels)
else:
conv1_in_channels, conv1_out_channels = in_channels, out_channels
conv2_in_channels, conv2_out_channels = out_channels, out_channels
self.add_module('SingleConv1', SingleConv(conv1_in_channels,
conv1_out_channels, kernel_size, order, num_groups))
self.add_module('SingleConv2', SingleConv(conv2_in_channels,
conv2_out_channels, kernel_size, order, num_groups))
class Encoder(nn.Module):
"""
A single module from the encoder path consisting of the optional max
pooling layer (one may specify the MaxPool kernel_size to be different
than the standard (2,2,2), e.g. if the volumetric data is anisotropic
(make sure to use complementary scale_factor in the decoder path) followed by
a DoubleConv module.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
conv_kernel_size (int): size of the convolving kernel
apply_pooling (bool): if True use MaxPool3d before DoubleConv
pool_kernel_size (tuple): the size of the window to take a max over
pool_type (str): pooling layer: 'max' or 'avg'
basic_module(nn.Module): either ResNetBlock or DoubleConv
conv_layer_order (string): determines the order of layers
in `DoubleConv` module. See `DoubleConv` for more info.
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, conv_kernel_size=3,
apply_pooling=True, pool_kernel_size=(2, 2, 2), pool_type='max',
basic_module=DoubleConv, conv_layer_order='crg', num_groups=8):
super(Encoder, self).__init__()
assert pool_type in ['max', 'avg']
if apply_pooling:
if pool_type == 'max':
self.pooling = nn.MaxPool3d(kernel_size=pool_kernel_size)
else:
self.pooling = nn.AvgPool3d(kernel_size=pool_kernel_size)
else:
self.pooling = None
self.basic_module = basic_module(in_channels, out_channels, encoder
=True, kernel_size=conv_kernel_size, order=conv_layer_order,
num_groups=num_groups)
def forward(self, x):
if self.pooling is not None:
x = self.pooling(x)
x = self.basic_module(x)
return x
def get_inputs():
return [torch.rand([4, 8, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_native_group_norm_0(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp5 = tmp2 + tmp4
tmp7 = triton_helpers.maximum(tmp1, tmp6)
tmp8 = tmp5 + tmp7
tmp10 = triton_helpers.maximum(tmp1, tmp9)
tmp11 = tmp8 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tmp2 - tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp4 - tmp13
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp7 - tmp13
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp10 - tmp13
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp24 / tmp12
tmp26 = 1e-05
tmp27 = tmp25 + tmp26
tmp28 = libdevice.rsqrt(tmp27)
tl.store(out_ptr0 + x0, tmp13, xmask)
tl.store(out_ptr1 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_native_group_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = xindex // 4
x1 = xindex // 4 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp3 = tl.load(in_ptr1 + x4, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr2 + x4, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr4 + x1, xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = tmp2 - tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 * tmp7
tmp10 = tmp8 + tmp9
tl.store(out_ptr0 + x3, tmp10, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 8, 4, 4), (128, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = torch.ops.aten.max_pool3d_with_indices.default(primals_1, [2,
2, 2], [2, 2, 2])
del primals_1
buf1 = buf0[0]
del buf0
buf3 = extern_kernels.convolution(reinterpret_tensor(buf1, (1, 4, 4,
2, 2), (64, 16, 4, 2, 1), 0), primals_2, stride=(1, 1, 1),
padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False,
output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf3, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1))
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
get_raw_stream(0)
triton_poi_fused_native_group_norm_0[grid(16)](buf3, buf4, buf5, 16,
XBLOCK=16, num_warps=1, num_stages=1)
buf6 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
triton_poi_fused_native_group_norm_1[grid(64)](buf3, buf4, buf5,
primals_3, primals_4, buf6, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del primals_4
buf7 = extern_kernels.convolution(reinterpret_tensor(buf6, (1, 4, 4,
2, 2), (0, 16, 4, 2, 1), 0), primals_5, stride=(1, 1, 1),
padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False,
output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf7, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1))
buf8 = buf5
del buf5
buf9 = buf4
del buf4
triton_poi_fused_native_group_norm_0[grid(16)](buf7, buf8, buf9, 16,
XBLOCK=16, num_warps=1, num_stages=1)
buf10 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
triton_poi_fused_native_group_norm_1[grid(64)](buf7, buf8, buf9,
primals_6, primals_7, buf10, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf8
del buf9
del primals_7
return (buf10, primals_2, primals_3, primals_5, primals_6,
reinterpret_tensor(buf1, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1), 0),
buf3, reinterpret_tensor(buf6, (1, 4, 4, 2, 2), (64, 16, 4, 2, 1),
0), buf7)
def conv3d(in_channels, out_channels, kernel_size, bias, padding=1):
return nn.Conv3d(in_channels, out_channels, kernel_size, padding=
padding, bias=bias)
def create_conv(in_channels, out_channels, kernel_size, order, num_groups,
padding=1):
"""
Create a list of modules with together constitute a single conv layer with non-linearity
and optional batchnorm/groupnorm.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
order (string): order of things, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
padding (int): add zero-padding to the input
Return:
list of tuple (name, module)
"""
assert 'c' in order, 'Conv layer MUST be present'
assert order[0
] not in 'rle', 'Non-linearity cannot be the first operation in the layer'
modules = []
for i, char in enumerate(order):
if char == 'r':
modules.append(('ReLU', nn.ReLU(inplace=True)))
elif char == 'l':
modules.append(('LeakyReLU', nn.LeakyReLU(negative_slope=0.1,
inplace=True)))
elif char == 'e':
modules.append(('ELU', nn.ELU(inplace=True)))
elif char == 'c':
bias = not ('g' in order or 'b' in order)
modules.append(('conv', conv3d(in_channels, out_channels,
kernel_size, bias, padding=padding)))
elif char == 'g':
is_before_conv = i < order.index('c')
assert not is_before_conv, 'GroupNorm MUST go after the Conv3d'
if out_channels < num_groups:
num_groups = out_channels
modules.append(('groupnorm', nn.GroupNorm(num_groups=num_groups,
num_channels=out_channels)))
elif char == 'b':
is_before_conv = i < order.index('c')
if is_before_conv:
modules.append(('batchnorm', nn.BatchNorm3d(in_channels)))
else:
modules.append(('batchnorm', nn.BatchNorm3d(out_channels)))
else:
raise ValueError(
f"Unsupported layer type '{char}'. MUST be one of ['b', 'g', 'r', 'l', 'e', 'c']"
)
return modules
class SingleConv(nn.Sequential):
"""
Basic convolutional module consisting of a Conv3d, non-linearity and optional batchnorm/groupnorm. The order
of operations can be specified via the `order` parameter
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'crg', num_groups=8, padding=1):
super(SingleConv, self).__init__()
for name, module in create_conv(in_channels, out_channels,
kernel_size, order, num_groups, padding=padding):
self.add_module(name, module)
class DoubleConv(nn.Sequential):
"""
A module consisting of two consecutive convolution layers (e.g. BatchNorm3d+ReLU+Conv3d).
We use (Conv3d+ReLU+GroupNorm3d) by default.
This can be changed however by providing the 'order' argument, e.g. in order
to change to Conv3d+BatchNorm3d+ELU use order='cbe'.
Use padded convolutions to make sure that the output (H_out, W_out) is the same
as (H_in, W_in), so that you don't have to crop in the decoder path.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
encoder (bool): if True we're in the encoder path, otherwise we're in the decoder
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, encoder, kernel_size=3,
order='crg', num_groups=8):
super(DoubleConv, self).__init__()
if encoder:
conv1_in_channels = in_channels
conv1_out_channels = out_channels // 2
if conv1_out_channels < in_channels:
conv1_out_channels = in_channels
conv2_in_channels, conv2_out_channels = (conv1_out_channels,
out_channels)
else:
conv1_in_channels, conv1_out_channels = in_channels, out_channels
conv2_in_channels, conv2_out_channels = out_channels, out_channels
self.add_module('SingleConv1', SingleConv(conv1_in_channels,
conv1_out_channels, kernel_size, order, num_groups))
self.add_module('SingleConv2', SingleConv(conv2_in_channels,
conv2_out_channels, kernel_size, order, num_groups))
class EncoderNew(nn.Module):
"""
A single module from the encoder path consisting of the optional max
pooling layer (one may specify the MaxPool kernel_size to be different
than the standard (2,2,2), e.g. if the volumetric data is anisotropic
(make sure to use complementary scale_factor in the decoder path) followed by
a DoubleConv module.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
conv_kernel_size (int): size of the convolving kernel
apply_pooling (bool): if True use MaxPool3d before DoubleConv
pool_kernel_size (tuple): the size of the window to take a max over
pool_type (str): pooling layer: 'max' or 'avg'
basic_module(nn.Module): either ResNetBlock or DoubleConv
conv_layer_order (string): determines the order of layers
in `DoubleConv` module. See `DoubleConv` for more info.
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, conv_kernel_size=3,
apply_pooling=True, pool_kernel_size=(2, 2, 2), pool_type='max',
basic_module=DoubleConv, conv_layer_order='crg', num_groups=8):
super(EncoderNew, self).__init__()
assert pool_type in ['max', 'avg']
if apply_pooling:
if pool_type == 'max':
self.pooling = nn.MaxPool3d(kernel_size=pool_kernel_size)
else:
self.pooling = nn.AvgPool3d(kernel_size=pool_kernel_size)
else:
self.pooling = None
self.basic_module = basic_module(in_channels, out_channels, encoder
=True, kernel_size=conv_kernel_size, order=conv_layer_order,
num_groups=num_groups)
def forward(self, input_0):
primals_2 = self.basic_module.SingleConv1.conv.weight
primals_3 = self.basic_module.SingleConv1.groupnorm.weight
primals_4 = self.basic_module.SingleConv1.groupnorm.bias
primals_5 = self.basic_module.SingleConv2.conv.weight
primals_6 = self.basic_module.SingleConv2.groupnorm.weight
primals_7 = self.basic_module.SingleConv2.groupnorm.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
joowlim/pytorch-3dunet
|
Encoder
| false | 10,405 |
[
"MIT"
] | 0 |
d08049f60b619627521efd0fb171247e1536b262
|
https://github.com/joowlim/pytorch-3dunet/tree/d08049f60b619627521efd0fb171247e1536b262
|
StandardNorm
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yi/cyiakmznzllqklcnfmh5hu3smacswbrh4drlxafkndcx6god674a.py
# Topologically Sorted Source Nodes: [sub, truediv], Original ATen: [aten.sub, aten.div]
# Source node to ATen node mapping:
# sub => sub
# truediv => div
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, 4), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, 4), kwargs = {})
triton_poi_fused_div_sub_0 = async_compile.triton('triton_poi_fused_div_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 4.0
tmp2 = tmp0 - tmp1
tmp3 = 0.25
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, truediv], Original ATen: [aten.sub, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_div_sub_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class StandardNorm(nn.Module):
def __init__(self, mean, std):
super(StandardNorm, self).__init__()
self.mean = mean
self.std = std
def forward(self, x):
return (x - self.mean) / self.std
def inverse(self, x):
return x * self.std + self.mean
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'mean': 4, 'std': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_div_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 4.0
tmp2 = tmp0 - tmp1
tmp3 = 0.25
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + x0, tmp4, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_div_sub_0[grid(256)](arg0_1, buf0, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class StandardNormNew(nn.Module):
def __init__(self, mean, std):
super(StandardNormNew, self).__init__()
self.mean = mean
self.std = std
def inverse(self, x):
return x * self.std + self.mean
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
iclementine/speedyspeech
|
StandardNorm
| false | 10,406 |
[
"BSD-3-Clause"
] | 0 |
db527587a3699b71082d61c9e9fad7ed795d1980
|
https://github.com/iclementine/speedyspeech/tree/db527587a3699b71082d61c9e9fad7ed795d1980
|
EuclideanComparator_1
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/kf/ckf5m3mk5oxwljoolseccy4i3hna4r2zdcsgb4fkgn7aaawvw5zq.py
# Topologically Sorted Source Nodes: [dist, truediv], Original ATen: [aten.dist, aten.div]
# Source node to ATen node mapping:
# dist => pow_1, pow_2, sub, sum_1
# truediv => div
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %arg0_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, None), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%pow_2, 4), kwargs = {})
triton_per_fused_dist_div_0 = async_compile.triton('triton_per_fused_dist_div_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_dist_div_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_dist_div_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr1 + (r0), None)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = libdevice.sqrt(tmp6)
tmp8 = 0.25
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp9, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [dist, truediv], Original ATen: [aten.dist, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_dist_div_0.run(buf1, arg1_1, arg0_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from dataclasses import dataclass
from collections import defaultdict
import torch.optim
from torch import nn
class Base(nn.Module):
registered = defaultdict(dict)
@dataclass
class Config:
pass
@property
def config(self):
return self._config
def __init__(self, *args, config: Config=None, **kwargs):
super().__init__(*args, **kwargs)
self._config = config
def __str__(self) ->str:
return self.__name__
@classmethod
def module(Child, Impl):
try:
Impl.name
except AttributeError:
msg = 'Class {Impl} has no attribute .name'
raise irtm.IRTMError(msg)
Base.registered[Child.__name__][Impl.name] = Impl
return Impl
@classmethod
def init(Child, *, name: str=None, **kwargs):
try:
if name is None:
name = 'noop'
A = Base.registered[Child.__name__][name]
except KeyError:
dicrep = yaml.dump(Base.registered, default_flow_style=False)
msg = (
f'could not find module "{name}"\n\navailable modules:\n{dicrep}'
)
raise irtm.IRTMError(msg)
config = A.Config(**kwargs)
log.info(f'! initializing {A.__name__} with {config}')
return A(config=config)
class Comparator(Base):
pass
@Comparator.module
class EuclideanComparator_1(Comparator):
name = 'euclidean 1'
def forward(self, X, Y):
return torch.dist(X, Y, p=2) / X.shape[0]
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
from dataclasses import dataclass
from collections import defaultdict
import torch.optim
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_dist_div_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr1 + r0, None)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = libdevice.sqrt(tmp6)
tmp8 = 0.25
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp9, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_dist_div_0[grid(1)](buf1, arg1_1, arg0_1, 1, 256,
num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf1,
class Base(nn.Module):
registered = defaultdict(dict)
@dataclass
class Config:
pass
@property
def config(self):
return self._config
def __init__(self, *args, config: Config=None, **kwargs):
super().__init__(*args, **kwargs)
self._config = config
def __str__(self) ->str:
return self.__name__
@classmethod
def module(Child, Impl):
try:
Impl.name
except AttributeError:
msg = 'Class {Impl} has no attribute .name'
raise irtm.IRTMError(msg)
Base.registered[Child.__name__][Impl.name] = Impl
return Impl
@classmethod
def init(Child, *, name: str=None, **kwargs):
try:
if name is None:
name = 'noop'
A = Base.registered[Child.__name__][name]
except KeyError:
dicrep = yaml.dump(Base.registered, default_flow_style=False)
msg = (
f'could not find module "{name}"\n\navailable modules:\n{dicrep}'
)
raise irtm.IRTMError(msg)
config = A.Config(**kwargs)
log.info(f'! initializing {A.__name__} with {config}')
return A(config=config)
class Comparator(Base):
pass
@Comparator.module
class EuclideanComparator_1New(Comparator):
name = 'euclidean 1'
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
lavis-nlp/irtm
|
EuclideanComparator_1
| false | 10,407 |
[
"MIT"
] | 0 |
e6c96519918795cfaa0c09ef2d4164f451265518
|
https://github.com/lavis-nlp/irtm/tree/e6c96519918795cfaa0c09ef2d4164f451265518
|
AffineConstantFlow
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/n3/cn3wlek7bl5jirlp73pieiarvuadnzcb6uy5z3ieztq35hnq6trv.py
# Topologically Sorted Source Nodes: [exp, mul, z], Original ATen: [aten.exp, aten.mul, aten.add]
# Source node to ATen node mapping:
# exp => exp
# mul => mul
# z => add
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%primals_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %exp), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %primals_3), kwargs = {})
triton_poi_fused_add_exp_mul_0 = async_compile.triton('triton_poi_fused_add_exp_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_exp_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_exp_mul_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp3 = tmp0 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + (x2), tmp5, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zj/czjrqmkyp2ezurdzw3bheq5i5tw7cwlzcc36wzadkzk7h5or6ugs.py
# Topologically Sorted Source Nodes: [log_det, repeat], Original ATen: [aten.sum, aten.repeat]
# Source node to ATen node mapping:
# log_det => sum_1
# repeat => repeat
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%primals_1, [1]), kwargs = {})
# %repeat : [num_users=1] = call_function[target=torch.ops.aten.repeat.default](args = (%sum_1, [4, 1]), kwargs = {})
triton_per_fused_repeat_sum_1 = async_compile.triton('triton_per_fused_repeat_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_repeat_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_repeat_sum_1(in_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp3, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [exp, mul, z], Original ATen: [aten.exp, aten.mul, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_exp_mul_0.run(primals_2, primals_1, primals_3, buf0, 256, grid=grid(256), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_det, repeat], Original ATen: [aten.sum, aten.repeat]
triton_per_fused_repeat_sum_1.run(primals_1, buf2, 1, 4, grid=grid(1), stream=stream0)
return (buf0, buf2, primals_1, primals_2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import Tensor
from torch import nn
class FlowBlock(nn.Module):
"""
Abstract base class for any flow blocks.
"""
def __init__(self, dimension):
super(FlowBlock, self).__init__()
self.dimension = dimension
def forward(self, x: 'Tensor') ->(Tensor, Tensor):
"""
When implemented, forward method will represent z = f(x) and log |det f'(x)/dx|
x: (*, dimension), z: (*, dimension) and log_det: (*, 1)
"""
raise NotImplementedError('Forward not implemented')
def inverse(self, z: 'Tensor') ->(Tensor, Tensor):
"""
When implemented, inverse method will represent x = f^-(z) and log |det f^-'(z)/dz|
z: (*, dimension), x: (*, dimension) and log_det: (*, 1)
"""
raise NotImplementedError('Inverse not implemented')
class AffineConstantFlow(FlowBlock):
"""
Scales + Shifts the flow by (learned) constants per dimension.
In NICE paper there is a Scaling layer which is a special case of this where t is None
"""
def __init__(self, dimension, scale=True, shift=True):
super().__init__(dimension)
zeros = torch.zeros(size=(1, dimension))
self.s = nn.Parameter(torch.randn(1, dimension, requires_grad=True)
) if scale else zeros
self.t = nn.Parameter(torch.randn(1, dimension, requires_grad=True)
) if shift else zeros
def forward(self, x) ->(Tensor, Tensor):
z = x * torch.exp(self.s) + self.t
log_det = torch.sum(self.s, dim=1)
return z, log_det.repeat(x.shape[0], 1)
def inverse(self, z) ->(Tensor, Tensor):
x = (z - self.t) * torch.exp(-self.s)
log_det = torch.sum(-self.s, dim=1)
return x, log_det.repeat(z.shape[0], 1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dimension': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch import Tensor
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_exp_mul_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp3 = tmp0 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + x2, tmp5, xmask)
@triton.jit
def triton_per_fused_repeat_sum_1(in_ptr0, out_ptr1, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp3, None)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_exp_mul_0[grid(256)](primals_2, primals_1,
primals_3, buf0, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
triton_per_fused_repeat_sum_1[grid(1)](primals_1, buf2, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
return buf0, buf2, primals_1, primals_2
class FlowBlock(nn.Module):
"""
Abstract base class for any flow blocks.
"""
def __init__(self, dimension):
super(FlowBlock, self).__init__()
self.dimension = dimension
def forward(self, x: 'Tensor') ->(Tensor, Tensor):
"""
When implemented, forward method will represent z = f(x) and log |det f'(x)/dx|
x: (*, dimension), z: (*, dimension) and log_det: (*, 1)
"""
raise NotImplementedError('Forward not implemented')
def inverse(self, z: 'Tensor') ->(Tensor, Tensor):
"""
When implemented, inverse method will represent x = f^-(z) and log |det f^-'(z)/dz|
z: (*, dimension), x: (*, dimension) and log_det: (*, 1)
"""
raise NotImplementedError('Inverse not implemented')
class AffineConstantFlowNew(FlowBlock):
"""
Scales + Shifts the flow by (learned) constants per dimension.
In NICE paper there is a Scaling layer which is a special case of this where t is None
"""
def __init__(self, dimension, scale=True, shift=True):
super().__init__(dimension)
zeros = torch.zeros(size=(1, dimension))
self.s = nn.Parameter(torch.randn(1, dimension, requires_grad=True)
) if scale else zeros
self.t = nn.Parameter(torch.randn(1, dimension, requires_grad=True)
) if shift else zeros
def inverse(self, z) ->(Tensor, Tensor):
x = (z - self.t) * torch.exp(-self.s)
log_det = torch.sum(-self.s, dim=1)
return x, log_det.repeat(z.shape[0], 1)
def forward(self, input_0):
primals_1 = self.s
primals_3 = self.t
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0], output[1]
|
lleonart1984/generative_modeling
|
AffineConstantFlow
| false | 10,408 |
[
"MIT"
] | 0 |
d47c53d34b9eb704b6e8b2c334262b53fe7f4f32
|
https://github.com/lleonart1984/generative_modeling/tree/d47c53d34b9eb704b6e8b2c334262b53fe7f4f32
|
MaxPoolingAggregator_1
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/2l/c2lm5wvy5varadxpp77k6lvi6yjwzernwi4uqg6gmabg2nygeeur.py
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
# Source node to ATen node mapping:
# max_1 => getitem
# Graph fragment:
# %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%max_1, 0), kwargs = {})
triton_poi_fused_max_0 = async_compile.triton('triton_poi_fused_max_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
stream0 = get_raw_stream(0)
triton_poi_fused_max_0.run(arg0_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from dataclasses import dataclass
from collections import defaultdict
import torch.optim
from torch import nn
class Base(nn.Module):
registered = defaultdict(dict)
@dataclass
class Config:
pass
@property
def config(self):
return self._config
def __init__(self, *args, config: Config=None, **kwargs):
super().__init__(*args, **kwargs)
self._config = config
def __str__(self) ->str:
return self.__name__
@classmethod
def module(Child, Impl):
try:
Impl.name
except AttributeError:
msg = 'Class {Impl} has no attribute .name'
raise irtm.IRTMError(msg)
Base.registered[Child.__name__][Impl.name] = Impl
return Impl
@classmethod
def init(Child, *, name: str=None, **kwargs):
try:
if name is None:
name = 'noop'
A = Base.registered[Child.__name__][name]
except KeyError:
dicrep = yaml.dump(Base.registered, default_flow_style=False)
msg = (
f'could not find module "{name}"\n\navailable modules:\n{dicrep}'
)
raise irtm.IRTMError(msg)
config = A.Config(**kwargs)
log.info(f'! initializing {A.__name__} with {config}')
return A(config=config)
class Aggregator(Base):
pass
@Aggregator.module
class MaxPoolingAggregator_1(Aggregator):
name = 'max 1'
def forward(self, X: 'torch.Tensor') ->torch.Tensor:
return X.max(axis=1).values
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from dataclasses import dataclass
from collections import defaultdict
import torch.optim
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_max_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tl.store(out_ptr0 + x2, tmp6, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_max_0[grid(64)](arg0_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del arg0_1
return buf0,
class Base(nn.Module):
registered = defaultdict(dict)
@dataclass
class Config:
pass
@property
def config(self):
return self._config
def __init__(self, *args, config: Config=None, **kwargs):
super().__init__(*args, **kwargs)
self._config = config
def __str__(self) ->str:
return self.__name__
@classmethod
def module(Child, Impl):
try:
Impl.name
except AttributeError:
msg = 'Class {Impl} has no attribute .name'
raise irtm.IRTMError(msg)
Base.registered[Child.__name__][Impl.name] = Impl
return Impl
@classmethod
def init(Child, *, name: str=None, **kwargs):
try:
if name is None:
name = 'noop'
A = Base.registered[Child.__name__][name]
except KeyError:
dicrep = yaml.dump(Base.registered, default_flow_style=False)
msg = (
f'could not find module "{name}"\n\navailable modules:\n{dicrep}'
)
raise irtm.IRTMError(msg)
config = A.Config(**kwargs)
log.info(f'! initializing {A.__name__} with {config}')
return A(config=config)
class Aggregator(Base):
pass
@Aggregator.module
class MaxPoolingAggregator_1New(Aggregator):
name = 'max 1'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
lavis-nlp/irtm
|
MaxPoolingAggregator_1
| false | 10,409 |
[
"MIT"
] | 0 |
e6c96519918795cfaa0c09ef2d4164f451265518
|
https://github.com/lavis-nlp/irtm/tree/e6c96519918795cfaa0c09ef2d4164f451265518
|
CPAMDec
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/zo/czobpmlyr5atbcpsuque6vcmk7nafmb3smtbzoqilz46drm7zbkm.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_3, %primals_4, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/j4/cj4f6qdb45emg4zrdv5vzxtw2vswpyt2rqyalr6mxgomzeyk55j5.py
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
# Source node to ATen node mapping:
# mul => mul
# out_2 => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_9, %view_6), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %primals_1), kwargs = {})
triton_poi_fused_add_mul_3 = async_compile.triton('triton_poi_fused_add_mul_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp2 = tl.load(in_ptr1 + (x0), xmask)
tmp4 = tl.load(in_ptr2 + (x0), xmask)
tmp3 = tmp1 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + (x0), tmp5, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_4, (1, ), (1, ))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1, ), (1, ))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_3, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_6, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_5
del primals_6
buf3 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf3, primals_4, 64, grid=grid(64), stream=stream0)
del primals_4
buf4 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 0), 0), reinterpret_tensor(buf2, (4, 1, 4), (4, 1, 1), 0), out=buf4)
buf5 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf4, buf5, 256, grid=grid(256), stream=stream0)
buf6 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [attention], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_8, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf7)
del primals_7
del primals_8
buf8 = reinterpret_tensor(buf5, (4, 4, 16), (64, 16, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf7, (4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(buf6, (4, 4, 16), (64, 1, 4), 0), out=buf8)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, out_2], Original ATen: [aten.mul, aten.add]
triton_poi_fused_add_mul_3.run(primals_9, buf8, primals_1, buf9, 256, grid=grid(256), stream=stream0)
return (buf9, primals_1, primals_3, primals_9, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), buf6, buf8, reinterpret_tensor(buf7, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf3, (4, 1, 16), (16, 16, 1), 0), reinterpret_tensor(buf2, (4, 4, 1), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
from torch.nn import Conv2d
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn import Linear
from torch.nn.parameter import Parameter
class CPAMDec(Module):
"""
CPAM decoding module
"""
def __init__(self, in_channels):
super(CPAMDec, self).__init__()
self.softmax = Softmax(dim=-1)
self.scale = Parameter(torch.zeros(1))
self.conv_query = Conv2d(in_channels=in_channels, out_channels=
in_channels // 4, kernel_size=1)
self.conv_key = Linear(in_channels, in_channels // 4)
self.conv_value = Linear(in_channels, in_channels)
def forward(self, x, y):
"""
inputs :
x : input feature(N,C,H,W) y:gathering centers(N,K,M)
returns :
out : compact position attention feature
attention map: (H*W)*M
"""
m_batchsize, C, width, height = x.size()
m_batchsize, K, _M = y.size()
proj_query = self.conv_query(x).view(m_batchsize, -1, width * height
).permute(0, 2, 1)
proj_key = self.conv_key(y).view(m_batchsize, K, -1).permute(0, 2, 1)
energy = torch.bmm(proj_query, proj_key)
attention = self.softmax(energy)
proj_value = self.conv_value(y).permute(0, 2, 1)
out = torch.bmm(proj_value, attention.permute(0, 2, 1))
out = out.view(m_batchsize, C, width, height)
out = self.scale * out + x
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.nn import Module
from torch.nn import Conv2d
from torch.nn import Parameter
from torch.nn import Softmax
from torch.nn import Linear
from torch.nn.parameter import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_mul_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp2 = tl.load(in_ptr1 + x0, xmask)
tmp4 = tl.load(in_ptr2 + x0, xmask)
tmp3 = tmp1 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + x0, tmp5, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_4, (1,), (1,))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1,), (1,))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_3, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_6, reinterpret_tensor(primals_2, (16,
4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 1), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_5
del primals_6
buf3 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(64)](buf3, primals_4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_4
buf4 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 0),
0), reinterpret_tensor(buf2, (4, 1, 4), (4, 1, 1), 0), out=buf4)
buf5 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf4, buf5, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf6 = buf4
del buf4
triton_poi_fused__softmax_2[grid(256)](buf5, buf6, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf7 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_8, reinterpret_tensor(primals_2, (16,
4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf7)
del primals_7
del primals_8
buf8 = reinterpret_tensor(buf5, (4, 4, 16), (64, 16, 1), 0)
del buf5
extern_kernels.bmm(reinterpret_tensor(buf7, (4, 4, 4), (16, 1, 4),
0), reinterpret_tensor(buf6, (4, 4, 16), (64, 1, 4), 0), out=buf8)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_3[grid(256)](primals_9, buf8, primals_1,
buf9, 256, XBLOCK=256, num_warps=4, num_stages=1)
return buf9, primals_1, primals_3, primals_9, reinterpret_tensor(primals_2,
(16, 4), (4, 1), 0), buf6, buf8, reinterpret_tensor(buf7, (4, 4, 4),
(16, 4, 1), 0), reinterpret_tensor(buf3, (4, 1, 16), (16, 16, 1), 0
), reinterpret_tensor(buf2, (4, 4, 1), (4, 1, 1), 0)
class CPAMDecNew(Module):
"""
CPAM decoding module
"""
def __init__(self, in_channels):
super(CPAMDecNew, self).__init__()
self.softmax = Softmax(dim=-1)
self.scale = Parameter(torch.zeros(1))
self.conv_query = Conv2d(in_channels=in_channels, out_channels=
in_channels // 4, kernel_size=1)
self.conv_key = Linear(in_channels, in_channels // 4)
self.conv_value = Linear(in_channels, in_channels)
def forward(self, input_0, input_1):
primals_4 = self.scale
primals_3 = self.conv_query.weight
primals_6 = self.conv_query.bias
primals_5 = self.conv_key.weight
primals_9 = self.conv_key.bias
primals_7 = self.conv_value.weight
primals_8 = self.conv_value.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
|
bfjei2825401/siamban
|
CPAMDec
| false | 10,410 |
[
"Apache-2.0"
] | 0 |
c41d58742b146dfc8960053453227c6e9fec1bac
|
https://github.com/bfjei2825401/siamban/tree/c41d58742b146dfc8960053453227c6e9fec1bac
|
LearnedPositionalEncoding
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/up/cuphmudfbm5xmuymbvibhn5q6e5y55prj6knqkmj43mx7ovcvcgz.py
# Topologically Sorted Source Nodes: [pos], Original ATen: [aten.arange]
# Source node to ATen node mapping:
# pos => iota
# Graph fragment:
# %iota : [num_users=1] = call_function[target=torch.ops.prims.iota.default](args = (4,), kwargs = {start: 0, step: 1, dtype: torch.int64, device: cuda:0, requires_grad: False})
triton_poi_fused_arange_0 = async_compile.triton('triton_poi_fused_arange_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0,), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_arange_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_arange_0(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7g/c7ggkspjiryu6lrnyqmt3ltk6cvy3xv7hlniyify4whtbxpu3vcz.py
# Topologically Sorted Source Nodes: [embedding], Original ATen: [aten.embedding]
# Source node to ATen node mapping:
# embedding => embedding
# Graph fragment:
# %embedding : [num_users=2] = call_function[target=torch.ops.aten.embedding.default](args = (%primals_2, %expand), kwargs = {})
triton_poi_fused_embedding_1 = async_compile.triton('triton_poi_fused_embedding_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_embedding_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_embedding_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 16)
x0 = xindex % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x2), xmask, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 100, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tl.device_assert(((0 <= tmp4) & (tmp4 < 100)) | ~(xmask), "index out of bounds: 0 <= tmp4 < 100")
tmp6 = tl.load(in_ptr1 + (x0 + (4*tmp4)), xmask)
tl.store(out_ptr0 + (x4), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/y3/cy35gsxm3xa2mi4ahfwlz6f5exunl2ig4pka6onlgazxeb4uuhe3.py
# Topologically Sorted Source Nodes: [x, u, sub, pow_1, s], Original ATen: [aten.add, aten.mean, aten.sub, aten.pow]
# Source node to ATen node mapping:
# pow_1 => pow_1
# s => mean_1
# sub => sub
# u => mean
# x => add
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_1, %embedding), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%add, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %mean), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_1, [-1], True), kwargs = {})
triton_poi_fused_add_mean_pow_sub_2 = async_compile.triton('triton_poi_fused_add_mean_pow_sub_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mean_pow_sub_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mean_pow_sub_2(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_ptr0 + (4*x2), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + (4*x2)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x2)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + (x2), tmp16, xmask)
tl.store(out_ptr1 + (x2), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hy/chysp3ia3h42md5xyhzqsgskl34zzc6iy7yp2pqz3zla4xgfsa47.py
# Topologically Sorted Source Nodes: [x, u, sub, add_1, sqrt, x_1, mul, add_2], Original ATen: [aten.add, aten.mean, aten.sub, aten.sqrt, aten.div, aten.mul]
# Source node to ATen node mapping:
# add_1 => add_1
# add_2 => add_2
# mul => mul
# sqrt => sqrt
# sub => sub
# u => mean
# x => add
# x_1 => div
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_1, %embedding), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%add, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %mean), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean_1, 1e-12), kwargs = {})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%add_1,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %sqrt), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, %div), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %primals_4), kwargs = {})
triton_poi_fused_add_div_mean_mul_sqrt_sub_3 = async_compile.triton('triton_poi_fused_add_div_mean_mul_sqrt_sub_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_mean_mul_sqrt_sub_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_mean_mul_sqrt_sub_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x3 = xindex
x4 = xindex % 64
x5 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x3), xmask)
tmp2 = tl.load(in_ptr2 + (x4), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + (x5), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr4 + (x5), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 - tmp4
tmp7 = 1e-12
tmp8 = tmp6 + tmp7
tmp9 = libdevice.sqrt(tmp8)
tmp10 = tmp5 / tmp9
tmp11 = tmp0 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + (x3), tmp13, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (100, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, ), (1, ), torch.int64)
# Topologically Sorted Source Nodes: [pos], Original ATen: [aten.arange]
stream0 = get_raw_stream(0)
triton_poi_fused_arange_0.run(buf0, 4, grid=grid(4), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [embedding], Original ATen: [aten.embedding]
triton_poi_fused_embedding_1.run(buf0, primals_2, buf1, 64, grid=grid(64), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [x, u, sub, pow_1, s], Original ATen: [aten.add, aten.mean, aten.sub, aten.pow]
triton_poi_fused_add_mean_pow_sub_2.run(primals_1, buf1, buf2, buf3, 64, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x, u, sub, add_1, sqrt, x_1, mul, add_2], Original ATen: [aten.add, aten.mean, aten.sub, aten.sqrt, aten.div, aten.mul]
triton_poi_fused_add_div_mean_mul_sqrt_sub_3.run(primals_3, primals_1, buf1, buf2, buf3, primals_4, buf4, 256, grid=grid(256), stream=stream0)
del buf2
del buf3
del primals_4
return (buf4, primals_1, primals_3, reinterpret_tensor(buf0, (4, 1), (1, 1), 0), buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((100, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class LayerNorm(nn.Module):
"""A layernorm module in the TF style (epsilon inside the square root)."""
def __init__(self, d_model, variance_epsilon=1e-12):
super().__init__()
self.gamma = nn.Parameter(torch.ones(d_model))
self.beta = nn.Parameter(torch.zeros(d_model))
self.variance_epsilon = variance_epsilon
def forward(self, x):
u = x.mean(-1, keepdim=True)
s = (x - u).pow(2).mean(-1, keepdim=True)
x = (x - u) / torch.sqrt(s + self.variance_epsilon)
return self.gamma * x + self.beta
class LearnedPositionalEncoding(nn.Module):
def __init__(self, d_model, dropout=0.1, max_len=100):
super(LearnedPositionalEncoding, self).__init__()
self.dropout = nn.Dropout(p=dropout)
self.pos_embed = nn.Embedding(max_len, d_model)
self.layernorm = LayerNorm(d_model)
def forward(self, x):
seq_len = x.size(0)
pos = torch.arange(seq_len, dtype=torch.long, device=x.device)
pos = pos.unsqueeze(-1).expand(x.size()[:2])
x = x + self.pos_embed(pos)
return self.dropout(self.layernorm(x))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'d_model': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_arange_0(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tl.store(out_ptr0 + x0, tmp0, xmask)
@triton.jit
def triton_poi_fused_embedding_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 16
x0 = xindex % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + x2, xmask, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 100, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tl.device_assert((0 <= tmp4) & (tmp4 < 100) | ~xmask,
'index out of bounds: 0 <= tmp4 < 100')
tmp6 = tl.load(in_ptr1 + (x0 + 4 * tmp4), xmask)
tl.store(out_ptr0 + x4, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_mean_pow_sub_2(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_ptr0 + 4 * x2, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x2), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + 4 * x2), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + x2, tmp16, xmask)
tl.store(out_ptr1 + x2, tmp28, xmask)
@triton.jit
def triton_poi_fused_add_div_mean_mul_sqrt_sub_3(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x3 = xindex
x4 = xindex % 64
x5 = xindex // 4
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x3, xmask)
tmp2 = tl.load(in_ptr2 + x4, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + x5, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr4 + x5, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 - tmp4
tmp7 = 1e-12
tmp8 = tmp6 + tmp7
tmp9 = libdevice.sqrt(tmp8)
tmp10 = tmp5 / tmp9
tmp11 = tmp0 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + x3, tmp13, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (100, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4,), (1,), torch.int64)
get_raw_stream(0)
triton_poi_fused_arange_0[grid(4)](buf0, 4, XBLOCK=4, num_warps=1,
num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_embedding_1[grid(64)](buf0, primals_2, buf1, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused_add_mean_pow_sub_2[grid(64)](primals_1, buf1, buf2,
buf3, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_div_mean_mul_sqrt_sub_3[grid(256)](primals_3,
primals_1, buf1, buf2, buf3, primals_4, buf4, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf2
del buf3
del primals_4
return buf4, primals_1, primals_3, reinterpret_tensor(buf0, (4, 1), (1,
1), 0), buf1
class LayerNorm(nn.Module):
"""A layernorm module in the TF style (epsilon inside the square root)."""
def __init__(self, d_model, variance_epsilon=1e-12):
super().__init__()
self.gamma = nn.Parameter(torch.ones(d_model))
self.beta = nn.Parameter(torch.zeros(d_model))
self.variance_epsilon = variance_epsilon
def forward(self, x):
u = x.mean(-1, keepdim=True)
s = (x - u).pow(2).mean(-1, keepdim=True)
x = (x - u) / torch.sqrt(s + self.variance_epsilon)
return self.gamma * x + self.beta
class LearnedPositionalEncodingNew(nn.Module):
def __init__(self, d_model, dropout=0.1, max_len=100):
super(LearnedPositionalEncodingNew, self).__init__()
self.dropout = nn.Dropout(p=dropout)
self.pos_embed = nn.Embedding(max_len, d_model)
self.layernorm = LayerNorm(d_model)
def forward(self, input_0):
primals_2 = self.pos_embed.weight
primals_3 = self.layernorm.gamma
primals_4 = self.layernorm.beta
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
|
longnsl1998/vietocr
|
LearnedPositionalEncoding
| false | 10,411 |
[
"Apache-2.0"
] | 0 |
686dd6c9d897e0401c20e7dcadb07a07c1dbc284
|
https://github.com/longnsl1998/vietocr/tree/686dd6c9d897e0401c20e7dcadb07a07c1dbc284
|
CrossNet
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/52/c52kaji6eyiltsqxp7w6aivnhxv5zwp3fyswckudklooo5a6gore.py
# Topologically Sorted Source Nodes: [xl_w], Original ATen: [aten.clone, aten._unsafe_view]
# Source node to ATen node mapping:
# xl_w => clone, view
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute,), kwargs = {memory_format: torch.contiguous_format})
# %view : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%clone, [64, 4]), kwargs = {})
triton_poi_fused__unsafe_view_clone_0 = async_compile.triton('triton_poi_fused__unsafe_view_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_view_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_view_clone_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + ((16*x1) + (64*(y0 // 16)) + (y0 % 16)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/c5/cc5niyk56fa3m2oc45eqgnf3eeposiu56eefgo6q2ozqlq3nwxnc.py
# Topologically Sorted Source Nodes: [dot_], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# dot_ => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = (xindex // 64)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/p5/cp5pohlvud22ry7jg7rk2cbrxg2zgd5m6ftrmu3zisebczw32ykt.py
# Topologically Sorted Source Nodes: [dot_], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# dot_ => clone_2
# Graph fragment:
# %clone_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_1,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_2 = async_compile.triton('triton_poi_fused_clone_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = (xindex // 64)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/rj/crjfvijs5tvs7lmrx3tnlbqpiqf6fftdmggpgew3khzaaiwhtbmp.py
# Topologically Sorted Source Nodes: [xl_w_1], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# xl_w_1 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x4 = (xindex // 256)
x5 = (xindex // 16) % 16
x2 = (xindex // 16) % 4
x6 = (xindex // 4) % 16
x7 = xindex
tmp0 = tl.load(in_ptr0 + (x5 + (16*x0) + (64*x4)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x6 + (16*x0) + (64*x4)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(out_ptr0 + (x7), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ho/cho3ytdwjrelha3yb4kblaihfa7pdkxvm7w76zy4lipwg7jvufdx.py
# Topologically Sorted Source Nodes: [add, x_l, add_2, x_l_1, x_l_2], Original ATen: [aten.add, aten.squeeze]
# Source node to ATen node mapping:
# add => add
# add_2 => add_2
# x_l => add_1
# x_l_1 => add_3
# x_l_2 => squeeze
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_5, %select_1), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %unsqueeze), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_11, %select_3), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_2, %add_1), kwargs = {})
# %squeeze : [num_users=1] = call_function[target=torch.ops.aten.squeeze.dim](args = (%add_3, 2), kwargs = {})
triton_poi_fused_add_squeeze_4 = async_compile.triton('triton_poi_fused_add_squeeze_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_squeeze_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_squeeze_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = (xindex // 4)
x1 = (xindex // 4) % 4
x3 = (xindex // 64)
x6 = xindex % 16
x7 = xindex
tmp0 = tl.load(in_ptr0 + (x4), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4 + x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x4), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x6 + (16*x3)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp2 + tmp7
tl.store(out_ptr0 + (x7), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 1), (4, 1, 1))
assert_size_stride(primals_3, (2, 4, 1), (4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [xl_w], Original ATen: [aten.clone, aten._unsafe_view]
stream0 = get_raw_stream(0)
triton_poi_fused__unsafe_view_clone_0.run(primals_1, buf0, 64, 4, grid=grid(64, 4), stream=stream0)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [xl_w], Original ATen: [aten.mm]
extern_kernels.mm(buf0, reinterpret_tensor(primals_2, (4, 1), (1, 1), 0), out=buf1)
buf2 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [dot_], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(primals_1, buf2, 1024, grid=grid(1024), stream=stream0)
buf3 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [dot_], Original ATen: [aten.clone]
triton_poi_fused_clone_2.run(buf1, buf3, 256, grid=grid(256), stream=stream0)
del buf1
buf4 = empty_strided_cuda((64, 4, 1), (4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [dot_], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf3, (64, 4, 1), (4, 1, 0), 0), out=buf4)
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [xl_w_1], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(buf4, primals_3, primals_1, buf5, 1024, grid=grid(1024), stream=stream0)
buf6 = reinterpret_tensor(buf3, (256, 1), (1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [xl_w_1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf5, (256, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 1), (1, 1), 4), out=buf6)
buf7 = empty_strided_cuda((64, 4, 1), (4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [dot__1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf6, (64, 4, 1), (4, 1, 1), 0), out=buf7)
del buf6
buf8 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, x_l, add_2, x_l_1, x_l_2], Original ATen: [aten.add, aten.squeeze]
triton_poi_fused_add_squeeze_4.run(buf7, primals_3, buf4, primals_1, buf8, 1024, grid=grid(1024), stream=stream0)
del buf4
del buf7
del primals_1
del primals_3
return (buf8, reinterpret_tensor(buf2, (64, 4, 4), (16, 1, 4), 0), reinterpret_tensor(buf5, (4, 256), (1, 4), 0), reinterpret_tensor(primals_2, (1, 4), (1, 1), 4), reinterpret_tensor(buf0, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((2, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((2, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from sklearn.metrics import *
import torch.onnx
import torch as torch
class CrossNet(nn.Module):
"""The Cross Network part of Deep&Cross Network model,
which leans both low and high degree cross feature.
Input shape
- 2D tensor with shape: ``(batch_size, units)``.
Output shape
- 2D tensor with shape: ``(batch_size, units)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **input_feature_num**: Positive integer, shape(Input tensor)[-1]
- **layer_num**: Positive integer, the cross layer number
- **parameterization**: string, ``"vector"`` or ``"matrix"`` , way to parameterize the cross network.
- **l2_reg**: float between 0 and 1. L2 regularizer strength applied to the kernel weights matrix
- **seed**: A Python integer to use as random seed.
References
- [Wang R, Fu B, Fu G, et al. Deep & cross network for ad click predictions[C]//Proceedings of the ADKDD'17. ACM, 2017: 12.](https://arxiv.org/abs/1708.05123)
- [Wang R, Shivanna R, Cheng D Z, et al. DCN-M: Improved Deep & Cross Network for Feature Cross Learning in Web-scale Learning to Rank Systems[J]. 2020.](https://arxiv.org/abs/2008.13535)
"""
def __init__(self, in_features, layer_num=2, parameterization='vector',
seed=1024, device='cpu'):
super(CrossNet, self).__init__()
self.layer_num = layer_num
self.parameterization = parameterization
if self.parameterization == 'vector':
self.kernels = nn.Parameter(torch.Tensor(self.layer_num,
in_features, 1))
elif self.parameterization == 'matrix':
self.kernels = nn.Parameter(torch.Tensor(self.layer_num,
in_features, in_features))
else:
raise ValueError("parameterization should be 'vector' or 'matrix'")
self.bias = nn.Parameter(torch.Tensor(self.layer_num, in_features, 1))
for i in range(self.kernels.shape[0]):
nn.init.xavier_normal_(self.kernels[i])
for i in range(self.bias.shape[0]):
nn.init.zeros_(self.bias[i])
self
def forward(self, inputs):
x_0 = inputs.unsqueeze(2)
x_l = x_0
for i in range(self.layer_num):
if self.parameterization == 'vector':
xl_w = torch.tensordot(x_l, self.kernels[i], dims=([1], [0]))
dot_ = torch.matmul(x_0, xl_w)
x_l = dot_ + self.bias[i] + x_l
elif self.parameterization == 'matrix':
xl_w = torch.matmul(self.kernels[i], x_l)
dot_ = xl_w + self.bias[i]
x_l = x_0 * dot_ + x_l
else:
raise ValueError(
"parameterization should be 'vector' or 'matrix'")
x_l = torch.squeeze(x_l, dim=2)
return x_l
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_features': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
from sklearn.metrics import *
import torch.onnx
import torch as torch
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused__unsafe_view_clone_0(in_ptr0, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (16 * x1 + 64 * (y0 // 16) + y0 % 16), xmask &
ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex // 64
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex // 64
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x4 = xindex // 256
x5 = xindex // 16 % 16
x2 = xindex // 16 % 4
x6 = xindex // 4 % 16
x7 = xindex
tmp0 = tl.load(in_ptr0 + (x5 + 16 * x0 + 64 * x4), xmask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x6 + 16 * x0 + 64 * x4), xmask,
eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(out_ptr0 + x7, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_squeeze_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex // 4
x1 = xindex // 4 % 4
x3 = xindex // 64
x6 = xindex % 16
x7 = xindex
tmp0 = tl.load(in_ptr0 + x4, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4 + x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x4, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x6 + 16 * x3), xmask, eviction_policy=
'evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp2 + tmp7
tl.store(out_ptr0 + x7, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 1), (4, 1, 1))
assert_size_stride(primals_3, (2, 4, 1), (4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__unsafe_view_clone_0[grid(64, 4)](primals_1, buf0,
64, 4, XBLOCK=4, YBLOCK=32, num_warps=4, num_stages=1)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.mm(buf0, reinterpret_tensor(primals_2, (4, 1), (1, 1
), 0), out=buf1)
buf2 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_poi_fused_clone_1[grid(1024)](primals_1, buf2, 1024, XBLOCK=
256, num_warps=4, num_stages=1)
buf3 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
triton_poi_fused_clone_2[grid(256)](buf1, buf3, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf1
buf4 = empty_strided_cuda((64, 4, 1), (4, 1, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf3, (64, 4, 1), (4, 1, 0), 0), out=buf4)
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_poi_fused_clone_3[grid(1024)](buf4, primals_3, primals_1,
buf5, 1024, XBLOCK=256, num_warps=4, num_stages=1)
buf6 = reinterpret_tensor(buf3, (256, 1), (1, 1), 0)
del buf3
extern_kernels.mm(reinterpret_tensor(buf5, (256, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 1), (1, 1), 4), out=buf6)
buf7 = empty_strided_cuda((64, 4, 1), (4, 1, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf6, (64, 4, 1), (4, 1, 1), 0), out=buf7)
del buf6
buf8 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_poi_fused_add_squeeze_4[grid(1024)](buf7, primals_3, buf4,
primals_1, buf8, 1024, XBLOCK=128, num_warps=4, num_stages=1)
del buf4
del buf7
del primals_1
del primals_3
return buf8, reinterpret_tensor(buf2, (64, 4, 4), (16, 1, 4), 0
), reinterpret_tensor(buf5, (4, 256), (1, 4), 0), reinterpret_tensor(
primals_2, (1, 4), (1, 1), 4), reinterpret_tensor(buf0, (4, 64), (1,
4), 0)
class CrossNetNew(nn.Module):
"""The Cross Network part of Deep&Cross Network model,
which leans both low and high degree cross feature.
Input shape
- 2D tensor with shape: ``(batch_size, units)``.
Output shape
- 2D tensor with shape: ``(batch_size, units)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **input_feature_num**: Positive integer, shape(Input tensor)[-1]
- **layer_num**: Positive integer, the cross layer number
- **parameterization**: string, ``"vector"`` or ``"matrix"`` , way to parameterize the cross network.
- **l2_reg**: float between 0 and 1. L2 regularizer strength applied to the kernel weights matrix
- **seed**: A Python integer to use as random seed.
References
- [Wang R, Fu B, Fu G, et al. Deep & cross network for ad click predictions[C]//Proceedings of the ADKDD'17. ACM, 2017: 12.](https://arxiv.org/abs/1708.05123)
- [Wang R, Shivanna R, Cheng D Z, et al. DCN-M: Improved Deep & Cross Network for Feature Cross Learning in Web-scale Learning to Rank Systems[J]. 2020.](https://arxiv.org/abs/2008.13535)
"""
def __init__(self, in_features, layer_num=2, parameterization='vector',
seed=1024, device='cpu'):
super(CrossNetNew, self).__init__()
self.layer_num = layer_num
self.parameterization = parameterization
if self.parameterization == 'vector':
self.kernels = nn.Parameter(torch.Tensor(self.layer_num,
in_features, 1))
elif self.parameterization == 'matrix':
self.kernels = nn.Parameter(torch.Tensor(self.layer_num,
in_features, in_features))
else:
raise ValueError("parameterization should be 'vector' or 'matrix'")
self.bias = nn.Parameter(torch.Tensor(self.layer_num, in_features, 1))
for i in range(self.kernels.shape[0]):
nn.init.xavier_normal_(self.kernels[i])
for i in range(self.bias.shape[0]):
nn.init.zeros_(self.bias[i])
self
def forward(self, input_0):
primals_2 = self.kernels
primals_3 = self.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
dulvqingyunLT/DeepCTR-Torch
|
CrossNet
| false | 10,412 |
[
"Apache-2.0"
] | 0 |
f40cf08f3469aa471f9ca69e44c5de51180341cc
|
https://github.com/dulvqingyunLT/DeepCTR-Torch/tree/f40cf08f3469aa471f9ca69e44c5de51180341cc
|
ExtResNetBlock
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/cn/ccnjw5i2bmojhqd43gtvy2mo5rvcvx4kyqocflzm32zw36biefd7.py
# Topologically Sorted Source Nodes: [input_2, input_3], Original ATen: [aten.native_group_norm, aten.elu]
# Source node to ATen node mapping:
# input_2 => add, add_1, mul_1, rsqrt, var_mean
# input_3 => expm1, gt, mul_2, mul_4, where
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %unsqueeze_6), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %unsqueeze_3), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%add_1, 0), kwargs = {})
# %mul_2 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, 1.0), kwargs = {})
# %expm1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul_2,), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1, 1.0), kwargs = {})
# %where : [num_users=3] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul_2, %mul_4), kwargs = {})
triton_per_fused_elu_native_group_norm_0 = async_compile.triton('triton_per_fused_elu_native_group_norm_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_elu_native_group_norm_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_elu_native_group_norm_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
x2 = xindex % 4
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + (x2), xmask, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 0.0
tmp29 = tmp27 > tmp28
tmp30 = 1.0
tmp31 = tmp27 * tmp30
tmp32 = libdevice.expm1(tmp31)
tmp33 = tmp32 * tmp30
tmp34 = tl.where(tmp29, tmp31, tmp33)
tl.store(in_out_ptr0 + (r1 + (16*x0)), tmp34, xmask)
tl.store(out_ptr2 + (x0), tmp22, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ql/cql5y77ikb3qyybctpvmmqkg56j3yahci6d3fpavhe5gjjt6labu.py
# Topologically Sorted Source Nodes: [input_8, out, out_1], Original ATen: [aten.native_group_norm, aten.add, aten.elu]
# Source node to ATen node mapping:
# input_8 => add_4, add_5, mul_11, rsqrt_2, var_mean_2
# out => add_6
# out_1 => expm1_2, gt_2, mul_12, mul_14, where_2
# Graph fragment:
# %var_mean_2 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_4, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_4, 1e-05), kwargs = {})
# %rsqrt_2 : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
# %mul_11 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_5, %unsqueeze_22), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_11, %unsqueeze_19), kwargs = {})
# %add_6 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_5, %where), kwargs = {})
# %gt_2 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%add_6, 0), kwargs = {})
# %mul_12 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_6, 1.0), kwargs = {})
# %expm1_2 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul_12,), kwargs = {})
# %mul_14 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1_2, 1.0), kwargs = {})
# %where_2 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt_2, %mul_12, %mul_14), kwargs = {})
triton_per_fused_add_elu_native_group_norm_1 = async_compile.triton('triton_per_fused_add_elu_native_group_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_elu_native_group_norm_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_elu_native_group_norm_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
x2 = xindex % 4
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + (x2), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr3 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp29 = tmp27 + tmp28
tmp30 = 0.0
tmp31 = tmp29 > tmp30
tmp32 = 1.0
tmp33 = tmp29 * tmp32
tmp34 = libdevice.expm1(tmp33)
tmp35 = tmp34 * tmp32
tmp36 = tl.where(tmp31, tmp33, tmp35)
tl.store(in_out_ptr0 + (r1 + (16*x0)), tmp36, xmask)
tl.store(out_ptr2 + (x0), tmp22, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(reinterpret_tensor(primals_2, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1), 0), primals_1, stride=(1, 1, 1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False, output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf0, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = buf4; del buf4 # reuse
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_2, input_3], Original ATen: [aten.native_group_norm, aten.elu]
stream0 = get_raw_stream(0)
triton_per_fused_elu_native_group_norm_0.run(buf6, buf0, primals_3, primals_4, buf1, buf5, 16, 16, grid=grid(16), stream=stream0)
del primals_4
# Topologically Sorted Source Nodes: [input_4], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(reinterpret_tensor(buf6, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1), 0), primals_5, stride=(1, 1, 1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False, output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf7, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf8 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf13 = buf11; del buf11 # reuse
buf12 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_5, input_6], Original ATen: [aten.native_group_norm, aten.elu]
triton_per_fused_elu_native_group_norm_0.run(buf13, buf7, primals_6, primals_7, buf8, buf12, 16, 16, grid=grid(16), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [input_7], Original ATen: [aten.convolution]
buf14 = extern_kernels.convolution(reinterpret_tensor(buf13, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1), 0), primals_8, stride=(1, 1, 1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False, output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf14, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf15 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf19 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf20 = buf19; del buf19 # reuse
buf18 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_8, out, out_1], Original ATen: [aten.native_group_norm, aten.add, aten.elu]
triton_per_fused_add_elu_native_group_norm_1.run(buf20, buf14, primals_9, primals_10, buf6, buf15, buf18, 16, 16, grid=grid(16), stream=stream0)
del primals_10
return (buf20, primals_1, primals_3, primals_5, primals_6, primals_8, primals_9, reinterpret_tensor(primals_2, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1), 0), buf0, reinterpret_tensor(buf1, (4, 4), (4, 1), 0), reinterpret_tensor(buf5, (4, 4), (4, 1), 0), buf6, buf7, reinterpret_tensor(buf8, (4, 4), (4, 1), 0), reinterpret_tensor(buf12, (4, 4), (4, 1), 0), buf13, buf14, reinterpret_tensor(buf15, (4, 4), (4, 1), 0), reinterpret_tensor(buf18, (4, 4), (4, 1), 0), buf20, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3, 3), (108, 27, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 3, 3, 3), (108, 27, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 3, 3, 3), (108, 27, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
def conv3d(in_channels, out_channels, kernel_size, bias, padding=1):
return nn.Conv3d(in_channels, out_channels, kernel_size, padding=
padding, bias=bias)
def create_conv(in_channels, out_channels, kernel_size, order, num_groups,
padding=1):
"""
Create a list of modules with together constitute a single conv layer with non-linearity
and optional batchnorm/groupnorm.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
order (string): order of things, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
padding (int): add zero-padding to the input
Return:
list of tuple (name, module)
"""
assert 'c' in order, 'Conv layer MUST be present'
assert order[0
] not in 'rle', 'Non-linearity cannot be the first operation in the layer'
modules = []
for i, char in enumerate(order):
if char == 'r':
modules.append(('ReLU', nn.ReLU(inplace=True)))
elif char == 'l':
modules.append(('LeakyReLU', nn.LeakyReLU(negative_slope=0.1,
inplace=True)))
elif char == 'e':
modules.append(('ELU', nn.ELU(inplace=True)))
elif char == 'c':
bias = not ('g' in order or 'b' in order)
modules.append(('conv', conv3d(in_channels, out_channels,
kernel_size, bias, padding=padding)))
elif char == 'g':
is_before_conv = i < order.index('c')
assert not is_before_conv, 'GroupNorm MUST go after the Conv3d'
if out_channels < num_groups:
num_groups = out_channels
modules.append(('groupnorm', nn.GroupNorm(num_groups=num_groups,
num_channels=out_channels)))
elif char == 'b':
is_before_conv = i < order.index('c')
if is_before_conv:
modules.append(('batchnorm', nn.BatchNorm3d(in_channels)))
else:
modules.append(('batchnorm', nn.BatchNorm3d(out_channels)))
else:
raise ValueError(
f"Unsupported layer type '{char}'. MUST be one of ['b', 'g', 'r', 'l', 'e', 'c']"
)
return modules
class SingleConv(nn.Sequential):
"""
Basic convolutional module consisting of a Conv3d, non-linearity and optional batchnorm/groupnorm. The order
of operations can be specified via the `order` parameter
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'crg', num_groups=8, padding=1):
super(SingleConv, self).__init__()
for name, module in create_conv(in_channels, out_channels,
kernel_size, order, num_groups, padding=padding):
self.add_module(name, module)
class ExtResNetBlock(nn.Module):
"""
Basic UNet block consisting of a SingleConv followed by the residual block.
The SingleConv takes care of increasing/decreasing the number of channels and also ensures that the number
of output channels is compatible with the residual block that follows.
This block can be used instead of standard DoubleConv in the Encoder module.
Motivated by: https://arxiv.org/pdf/1706.00120.pdf
Notice we use ELU instead of ReLU (order='cge') and put non-linearity after the groupnorm.
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'cge', num_groups=8, **kwargs):
super(ExtResNetBlock, self).__init__()
self.conv1 = SingleConv(in_channels, out_channels, kernel_size=
kernel_size, order=order, num_groups=num_groups)
self.conv2 = SingleConv(out_channels, out_channels, kernel_size=
kernel_size, order=order, num_groups=num_groups)
n_order = order
for c in 'rel':
n_order = n_order.replace(c, '')
self.conv3 = SingleConv(out_channels, out_channels, kernel_size=
kernel_size, order=n_order, num_groups=num_groups)
if 'l' in order:
self.non_linearity = nn.LeakyReLU(negative_slope=0.1, inplace=True)
elif 'e' in order:
self.non_linearity = nn.ELU(inplace=True)
else:
self.non_linearity = nn.ReLU(inplace=True)
def forward(self, x):
out = self.conv1(x)
residual = out
out = self.conv2(out)
out = self.conv3(out)
out += residual
out = self.non_linearity(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_elu_native_group_norm_0(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr2, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
x2 = xindex % 4
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + x2, xmask, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 0.0
tmp29 = tmp27 > tmp28
tmp30 = 1.0
tmp31 = tmp27 * tmp30
tmp32 = libdevice.expm1(tmp31)
tmp33 = tmp32 * tmp30
tmp34 = tl.where(tmp29, tmp31, tmp33)
tl.store(in_out_ptr0 + (r1 + 16 * x0), tmp34, xmask)
tl.store(out_ptr2 + x0, tmp22, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
@triton.jit
def triton_per_fused_add_elu_native_group_norm_1(in_out_ptr0, in_ptr0,
in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr2, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
x2 = xindex % 4
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + x2, xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr3 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp29 = tmp27 + tmp28
tmp30 = 0.0
tmp31 = tmp29 > tmp30
tmp32 = 1.0
tmp33 = tmp29 * tmp32
tmp34 = libdevice.expm1(tmp33)
tmp35 = tmp34 * tmp32
tmp36 = tl.where(tmp31, tmp33, tmp35)
tl.store(in_out_ptr0 + (r1 + 16 * x0), tmp36, xmask)
tl.store(out_ptr2 + x0, tmp22, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 3, 3, 3), (108, 27, 9, 3, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(reinterpret_tensor(primals_2, (1,
4, 4, 4, 4), (256, 64, 16, 4, 1), 0), primals_1, stride=(1, 1,
1), padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False,
output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf0, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = buf4
del buf4
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
get_raw_stream(0)
triton_per_fused_elu_native_group_norm_0[grid(16)](buf6, buf0,
primals_3, primals_4, buf1, buf5, 16, 16, XBLOCK=8, num_warps=2,
num_stages=1)
del primals_4
buf7 = extern_kernels.convolution(reinterpret_tensor(buf6, (1, 4, 4,
4, 4), (256, 64, 16, 4, 1), 0), primals_5, stride=(1, 1, 1),
padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False,
output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf7, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf8 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf13 = buf11
del buf11
buf12 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
triton_per_fused_elu_native_group_norm_0[grid(16)](buf13, buf7,
primals_6, primals_7, buf8, buf12, 16, 16, XBLOCK=8, num_warps=
2, num_stages=1)
del primals_7
buf14 = extern_kernels.convolution(reinterpret_tensor(buf13, (1, 4,
4, 4, 4), (256, 64, 16, 4, 1), 0), primals_8, stride=(1, 1, 1),
padding=(1, 1, 1), dilation=(1, 1, 1), transposed=False,
output_padding=(0, 0, 0), groups=1, bias=None)
assert_size_stride(buf14, (1, 4, 4, 4, 4), (256, 64, 16, 4, 1))
buf15 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf19 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf20 = buf19
del buf19
buf18 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
triton_per_fused_add_elu_native_group_norm_1[grid(16)](buf20, buf14,
primals_9, primals_10, buf6, buf15, buf18, 16, 16, XBLOCK=8,
num_warps=2, num_stages=1)
del primals_10
return (buf20, primals_1, primals_3, primals_5, primals_6, primals_8,
primals_9, reinterpret_tensor(primals_2, (1, 4, 4, 4, 4), (256, 64,
16, 4, 1), 0), buf0, reinterpret_tensor(buf1, (4, 4), (4, 1), 0),
reinterpret_tensor(buf5, (4, 4), (4, 1), 0), buf6, buf7,
reinterpret_tensor(buf8, (4, 4), (4, 1), 0), reinterpret_tensor(
buf12, (4, 4), (4, 1), 0), buf13, buf14, reinterpret_tensor(buf15,
(4, 4), (4, 1), 0), reinterpret_tensor(buf18, (4, 4), (4, 1), 0), buf20
)
def conv3d(in_channels, out_channels, kernel_size, bias, padding=1):
return nn.Conv3d(in_channels, out_channels, kernel_size, padding=
padding, bias=bias)
def create_conv(in_channels, out_channels, kernel_size, order, num_groups,
padding=1):
"""
Create a list of modules with together constitute a single conv layer with non-linearity
and optional batchnorm/groupnorm.
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
order (string): order of things, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
padding (int): add zero-padding to the input
Return:
list of tuple (name, module)
"""
assert 'c' in order, 'Conv layer MUST be present'
assert order[0
] not in 'rle', 'Non-linearity cannot be the first operation in the layer'
modules = []
for i, char in enumerate(order):
if char == 'r':
modules.append(('ReLU', nn.ReLU(inplace=True)))
elif char == 'l':
modules.append(('LeakyReLU', nn.LeakyReLU(negative_slope=0.1,
inplace=True)))
elif char == 'e':
modules.append(('ELU', nn.ELU(inplace=True)))
elif char == 'c':
bias = not ('g' in order or 'b' in order)
modules.append(('conv', conv3d(in_channels, out_channels,
kernel_size, bias, padding=padding)))
elif char == 'g':
is_before_conv = i < order.index('c')
assert not is_before_conv, 'GroupNorm MUST go after the Conv3d'
if out_channels < num_groups:
num_groups = out_channels
modules.append(('groupnorm', nn.GroupNorm(num_groups=num_groups,
num_channels=out_channels)))
elif char == 'b':
is_before_conv = i < order.index('c')
if is_before_conv:
modules.append(('batchnorm', nn.BatchNorm3d(in_channels)))
else:
modules.append(('batchnorm', nn.BatchNorm3d(out_channels)))
else:
raise ValueError(
f"Unsupported layer type '{char}'. MUST be one of ['b', 'g', 'r', 'l', 'e', 'c']"
)
return modules
class SingleConv(nn.Sequential):
"""
Basic convolutional module consisting of a Conv3d, non-linearity and optional batchnorm/groupnorm. The order
of operations can be specified via the `order` parameter
Args:
in_channels (int): number of input channels
out_channels (int): number of output channels
kernel_size (int): size of the convolving kernel
order (string): determines the order of layers, e.g.
'cr' -> conv + ReLU
'crg' -> conv + ReLU + groupnorm
'cl' -> conv + LeakyReLU
'ce' -> conv + ELU
num_groups (int): number of groups for the GroupNorm
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'crg', num_groups=8, padding=1):
super(SingleConv, self).__init__()
for name, module in create_conv(in_channels, out_channels,
kernel_size, order, num_groups, padding=padding):
self.add_module(name, module)
class ExtResNetBlockNew(nn.Module):
"""
Basic UNet block consisting of a SingleConv followed by the residual block.
The SingleConv takes care of increasing/decreasing the number of channels and also ensures that the number
of output channels is compatible with the residual block that follows.
This block can be used instead of standard DoubleConv in the Encoder module.
Motivated by: https://arxiv.org/pdf/1706.00120.pdf
Notice we use ELU instead of ReLU (order='cge') and put non-linearity after the groupnorm.
"""
def __init__(self, in_channels, out_channels, kernel_size=3, order=
'cge', num_groups=8, **kwargs):
super(ExtResNetBlockNew, self).__init__()
self.conv1 = SingleConv(in_channels, out_channels, kernel_size=
kernel_size, order=order, num_groups=num_groups)
self.conv2 = SingleConv(out_channels, out_channels, kernel_size=
kernel_size, order=order, num_groups=num_groups)
n_order = order
for c in 'rel':
n_order = n_order.replace(c, '')
self.conv3 = SingleConv(out_channels, out_channels, kernel_size=
kernel_size, order=n_order, num_groups=num_groups)
if 'l' in order:
self.non_linearity = nn.LeakyReLU(negative_slope=0.1, inplace=True)
elif 'e' in order:
self.non_linearity = nn.ELU(inplace=True)
else:
self.non_linearity = nn.ReLU(inplace=True)
def forward(self, input_0):
primals_1 = self.conv1.conv.weight
primals_3 = self.conv1.groupnorm.weight
primals_4 = self.conv1.groupnorm.bias
primals_5 = self.conv2.conv.weight
primals_6 = self.conv2.groupnorm.weight
primals_7 = self.conv2.groupnorm.bias
primals_8 = self.conv3.conv.weight
primals_9 = self.conv3.groupnorm.weight
primals_10 = self.conv3.groupnorm.bias
primals_2 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9, primals_10])
return output[0]
|
joowlim/pytorch-3dunet
|
ExtResNetBlock
| false | 10,413 |
[
"MIT"
] | 0 |
d08049f60b619627521efd0fb171247e1536b262
|
https://github.com/joowlim/pytorch-3dunet/tree/d08049f60b619627521efd0fb171247e1536b262
|
QNetwork
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/2i/c2ibuo46ejiy7dbwrzgygpphwuemyxqooxyjcai24lecredzgyhh.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 18944
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 296
x2 = xindex % 1184
x3 = (xindex // 1184)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (1280*x3)), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (296, 4), (4, 1))
assert_size_stride(primals_2, (296, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (296, 296), (296, 1))
assert_size_stride(primals_5, (296, ), (1, ))
assert_size_stride(primals_6, (4, 296), (296, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 296), (296, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 296), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 296), (4736, 1184, 296, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 296), (5120, 1280, 296, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf6, 18944, grid=grid(18944), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 296), (296, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 296), (296, 1), 0), reinterpret_tensor(primals_4, (296, 296), (1, 296), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 296), (4736, 1184, 296, 1), 0); del buf2 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 296), (5120, 1280, 296, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf5, 18944, grid=grid(18944), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 296), (296, 1), 0), reinterpret_tensor(primals_6, (296, 4), (1, 296), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 296), (296, 1), 0), reinterpret_tensor(buf3, (64, 296), (296, 1), 0), primals_6, buf5, primals_4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((296, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((296, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((296, 296), (296, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((296, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 296), (296, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
import torch.nn as nn
class QNetwork(nn.Module):
"""Actor (Policy) Model."""
def __init__(self, state_size, action_size, seed, fc1_units=296,
fc2_units=296):
"""Initialize parameters and build model.
Params
======
state_size (int): Dimension of each state
action_size (int): Dimension of each action
seed (int): Random seed
fc1_units (int): Number of nodes in first hidden layer
fc2_units (int): Number of nodes in second hidden layer
"""
super(QNetwork, self).__init__()
self.seed = torch.manual_seed(seed)
self.fc1 = nn.Linear(state_size, fc1_units)
self.fc2 = nn.Linear(fc1_units, fc2_units)
self.fc3 = nn.Linear(fc2_units, action_size)
def forward(self, state):
"""Build a network that maps state -> action values."""
x = F.relu(self.fc1(state))
x = F.relu(self.fc2(x))
return self.fc3(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'state_size': 4, 'action_size': 4, 'seed': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 18944
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 296
x2 = xindex % 1184
x3 = xindex // 1184
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 1280 * x3), tmp6, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (296, 4), (4, 1))
assert_size_stride(primals_2, (296,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (296, 296), (296, 1))
assert_size_stride(primals_5, (296,), (1,))
assert_size_stride(primals_6, (4, 296), (296, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 296), (296, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 296), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 296), (4736, 1184, 296, 1), 0
)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 296), (5120, 1280, 296, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(18944)](buf1,
primals_2, buf6, 18944, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 296), (296, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 296), (296, 1), 0),
reinterpret_tensor(primals_4, (296, 296), (1, 296), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 296), (4736, 1184, 296, 1), 0
)
del buf2
buf5 = empty_strided_cuda((4, 4, 4, 296), (5120, 1280, 296, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(18944)](buf3,
primals_5, buf5, 18944, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 296),
(296, 1), 0), reinterpret_tensor(primals_6, (296, 4), (1, 296),
0), alpha=1, beta=1, out=buf4)
del primals_7
return reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 296), (296, 1), 0
), reinterpret_tensor(buf3, (64, 296), (296, 1), 0
), primals_6, buf5, primals_4, buf6
class QNetworkNew(nn.Module):
"""Actor (Policy) Model."""
def __init__(self, state_size, action_size, seed, fc1_units=296,
fc2_units=296):
"""Initialize parameters and build model.
Params
======
state_size (int): Dimension of each state
action_size (int): Dimension of each action
seed (int): Random seed
fc1_units (int): Number of nodes in first hidden layer
fc2_units (int): Number of nodes in second hidden layer
"""
super(QNetworkNew, self).__init__()
self.seed = torch.manual_seed(seed)
self.fc1 = nn.Linear(state_size, fc1_units)
self.fc2 = nn.Linear(fc1_units, fc2_units)
self.fc3 = nn.Linear(fc2_units, action_size)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.fc3.weight
primals_7 = self.fc3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
luiz-rocha94/navigation
|
QNetwork
| false | 10,414 |
[
"MIT"
] | 0 |
fd5e00d8b9051e82dfe15793e53f8d1f86e8ecbe
|
https://github.com/luiz-rocha94/navigation/tree/fd5e00d8b9051e82dfe15793e53f8d1f86e8ecbe
|
Coskx
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/7w/c7wbptj2xs4dg6al5exxesdedwezdo6ffqlgla7qkjcyjfntox4v.py
# Topologically Sorted Source Nodes: [mul, cos], Original ATen: [aten.mul, aten.cos]
# Source node to ATen node mapping:
# cos => cos
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 50), kwargs = {})
# %cos : [num_users=1] = call_function[target=torch.ops.aten.cos.default](args = (%mul,), kwargs = {})
triton_poi_fused_cos_mul_0 = async_compile.triton('triton_poi_fused_cos_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cos_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cos_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 50.0
tmp2 = tmp0 * tmp1
tmp3 = tl_math.cos(tmp2)
tl.store(out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, cos], Original ATen: [aten.mul, aten.cos]
stream0 = get_raw_stream(0)
triton_poi_fused_cos_mul_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class Coskx(nn.Module):
def __init__(self, k=50):
super(Coskx, self).__init__()
self.k = k
def forward(self, input):
return torch.cos(input * self.k)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cos_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 50.0
tmp2 = tmp0 * tmp1
tmp3 = tl_math.cos(tmp2)
tl.store(out_ptr0 + x0, tmp3, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cos_mul_0[grid(256)](arg0_1, buf0, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class CoskxNew(nn.Module):
def __init__(self, k=50):
super(CoskxNew, self).__init__()
self.k = k
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
jiaj15/SAIL
|
Coskx
| false | 10,415 |
[
"MIT"
] | 0 |
734be06a2b0ae70801f59c191b86332592da97cf
|
https://github.com/jiaj15/SAIL/tree/734be06a2b0ae70801f59c191b86332592da97cf
|
GroupNorm32
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/zy/czyl6lkgkemly2rorp7dytnq5bsikpucv7yhz2var2hddv5xnkj4.py
# Topologically Sorted Source Nodes: [y, mul, sigmoid, y_1], Original ATen: [aten.native_group_norm, aten.mul, aten.sigmoid]
# Source node to ATen node mapping:
# mul => mul_2
# sigmoid => sigmoid
# y => add, add_1, mul_1, rsqrt, var_mean
# y_1 => mul_3
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %unsqueeze_5), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %unsqueeze_2), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, 4.0), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%mul_2,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, %sigmoid), kwargs = {})
triton_per_fused_mul_native_group_norm_sigmoid_0 = async_compile.triton('triton_per_fused_mul_native_group_norm_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_native_group_norm_sigmoid_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_native_group_norm_sigmoid_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (r3), None, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + (r3), None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = 64.0
tmp18 = tmp16 / tmp17
tmp19 = 1e-05
tmp20 = tmp18 + tmp19
tmp21 = libdevice.rsqrt(tmp20)
tmp22 = tmp0 - tmp10
tmp23 = tmp22 * tmp21
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 4.0
tmp29 = tmp27 * tmp28
tmp30 = tl.sigmoid(tmp29)
tmp31 = tmp27 * tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp21, xmask)
tl.store(in_out_ptr1 + (r1 + (64*x0)), tmp31, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf1 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [y, mul, sigmoid, y_1], Original ATen: [aten.native_group_norm, aten.mul, aten.sigmoid]
stream0 = get_raw_stream(0)
triton_per_fused_mul_native_group_norm_sigmoid_0.run(buf3, buf5, primals_1, primals_2, primals_3, buf0, 4, 64, grid=grid(4), stream=stream0)
return (buf5, primals_1, primals_2, primals_3, buf0, buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
from torch import nn
class GroupNorm32(nn.GroupNorm):
def __init__(self, num_groups, num_channels, swish, eps=1e-05):
super().__init__(num_groups=num_groups, num_channels=num_channels,
eps=eps)
self.swish = swish
def forward(self, x):
y = super().forward(x.float())
if self.swish == 1.0:
y = F.silu(y)
elif self.swish:
y = y * F.sigmoid(y * float(self.swish))
return y
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_groups': 1, 'num_channels': 4, 'swish': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mul_native_group_norm_sigmoid_0(in_out_ptr0,
in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = rindex // 16
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + r3, None, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + r3, None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = 64.0
tmp18 = tmp16 / tmp17
tmp19 = 1e-05
tmp20 = tmp18 + tmp19
tmp21 = libdevice.rsqrt(tmp20)
tmp22 = tmp0 - tmp10
tmp23 = tmp22 * tmp21
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 4.0
tmp29 = tmp27 * tmp28
tmp30 = tl.sigmoid(tmp29)
tmp31 = tmp27 * tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp21, xmask)
tl.store(in_out_ptr1 + (r1 + 64 * x0), tmp31, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf1
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = buf4
del buf4
get_raw_stream(0)
triton_per_fused_mul_native_group_norm_sigmoid_0[grid(4)](buf3,
buf5, primals_1, primals_2, primals_3, buf0, 4, 64, XBLOCK=1,
num_warps=2, num_stages=1)
return buf5, primals_1, primals_2, primals_3, buf0, buf3
class GroupNorm32New(nn.GroupNorm):
def __init__(self, num_groups, num_channels, swish, eps=1e-05):
super().__init__(num_groups=num_groups, num_channels=num_channels,
eps=eps)
self.swish = swish
def forward(self, input_0):
primals_2 = self.weight
primals_3 = self.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
litevxx/glid-3
|
GroupNorm32
| false | 10,416 |
[
"MIT"
] | 0 |
d7bd53e671d642b0cbc8af81197170b585c7e624
|
https://github.com/litevxx/glid-3/tree/d7bd53e671d642b0cbc8af81197170b585c7e624
|
Qnet
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/md/cmd3ewacyhu5w5hausgbjbmtnt5rr66cgczh4ibdypq7dz6p4v7g.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (128, 128), (128, 1))
assert_size_stride(primals_5, (128, ), (1, ))
assert_size_stride(primals_6, (2, 128), (128, 1))
assert_size_stride(primals_7, (2, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf6, 8192, grid=grid(8192), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(primals_4, (128, 128), (1, 128), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf2 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf5, 8192, grid=grid(8192), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 2), (2, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 128), (128, 1), 0), reinterpret_tensor(primals_6, (128, 2), (1, 128), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (reinterpret_tensor(buf4, (4, 4, 4, 2), (32, 8, 2, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(buf3, (64, 128), (128, 1), 0), primals_6, buf5, primals_4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((128, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((2, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((2, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import random
import torch
import torch.nn as nn
import torch.nn.functional as F
class Qnet(nn.Module):
def __init__(self):
super(Qnet, self).__init__()
self.fc1 = nn.Linear(4, 128)
self.fc2 = nn.Linear(128, 128)
self.fc3 = nn.Linear(128, 2)
def forward(self, x):
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
def get_action(self, obs, epsilon):
out = self.forward(obs)
coin = random.random()
if coin < epsilon:
return random.randint(0, 1)
else:
return out.argmax().item()
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import random
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (128, 128), (128, 1))
assert_size_stride(primals_5, (128,), (1,))
assert_size_stride(primals_6, (2, 128), (128, 1))
assert_size_stride(primals_7, (2,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(8192)](buf1,
primals_2, buf6, 8192, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 128), (128, 1), 0),
reinterpret_tensor(primals_4, (128, 128), (1, 128), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf2
buf5 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(8192)](buf3,
primals_5, buf5, 8192, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 2), (2, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 128),
(128, 1), 0), reinterpret_tensor(primals_6, (128, 2), (1, 128),
0), alpha=1, beta=1, out=buf4)
del primals_7
return reinterpret_tensor(buf4, (4, 4, 4, 2), (32, 8, 2, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 128), (128, 1), 0
), reinterpret_tensor(buf3, (64, 128), (128, 1), 0
), primals_6, buf5, primals_4, buf6
class QnetNew(nn.Module):
def __init__(self):
super(QnetNew, self).__init__()
self.fc1 = nn.Linear(4, 128)
self.fc2 = nn.Linear(128, 128)
self.fc3 = nn.Linear(128, 2)
def get_action(self, obs, epsilon):
out = self.forward(obs)
coin = random.random()
if coin < epsilon:
return random.randint(0, 1)
else:
return out.argmax().item()
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.fc3.weight
primals_7 = self.fc3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
linklab/link_rl_book_codes
|
Qnet
| false | 10,417 |
[
"MIT"
] | 0 |
b272b46d5ecd2802f34648440ff53641c68cbbf0
|
https://github.com/linklab/link_rl_book_codes/tree/b272b46d5ecd2802f34648440ff53641c68cbbf0
|
ScaledDotAttention
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/iu/ciuxern2omgit5ovksuiwlddxkww6e3pkid4q2h3sauzn5rbd35z.py
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv1d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_2, %primals_3, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/au/cau4pihcaptiev5y2ewn2o2nvrwhk7hogc72cofmmtbyv4rxc2oy.py
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv1d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_2, %primals_3, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 4) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hg/chg3iq6bscxmmxv5f7tuzgwycb4mgrimwfhv2nauw5rj4tt5cmv2.py
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# weights_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zu/czuvep3dmpmqmhiiliwubh4ghdt2qr27va67sszkua7trziinwov.py
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# weights_1 => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_3 = async_compile.triton('triton_poi_fused__softmax_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_5, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_8, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_11, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(primals_1, buf0, 16, 4, grid=grid(16, 4), stream=stream0)
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4), (16, 4, 1))
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv1d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(primals_4, buf2, 16, 4, grid=grid(16, 4), stream=stream0)
# Topologically Sorted Source Nodes: [conv1d_1], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf2, primals_5, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4), (16, 4, 1))
buf4 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv1d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(primals_7, buf4, 16, 4, grid=grid(16, 4), stream=stream0)
# Topologically Sorted Source Nodes: [conv1d_2], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(buf4, primals_8, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4), (16, 4, 1))
buf6 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf6, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
buf7 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [conv1d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf7, primals_6, 64, grid=grid(64), stream=stream0)
del primals_6
buf8 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 4, 4), (16, 1, 4), 0), buf7, out=buf8)
buf9 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf8, buf9, 64, grid=grid(64), stream=stream0)
buf10 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf9, buf10, 64, grid=grid(64), stream=stream0)
buf11 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [conv1d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf11, primals_9, 64, grid=grid(64), stream=stream0)
del primals_9
buf12 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(buf10, reinterpret_tensor(buf11, (4, 4, 4), (16, 1, 4), 0), out=buf12)
buf13 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv1d_3], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf12, buf13, 16, 4, grid=grid(16, 4), stream=stream0)
# Topologically Sorted Source Nodes: [conv1d_3], Original ATen: [aten.convolution]
buf14 = extern_kernels.convolution(buf13, primals_10, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf14, (4, 4, 4), (16, 4, 1))
del buf13
buf15 = buf14; del buf14 # reuse
# Topologically Sorted Source Nodes: [conv1d_3], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf15, primals_11, 64, grid=grid(64), stream=stream0)
del primals_11
return (reinterpret_tensor(buf15, (4, 4, 4), (16, 1, 4), 0), buf10, primals_2, primals_5, primals_8, primals_10, reinterpret_tensor(primals_1, (4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_4, (4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_7, (4, 4, 4), (16, 1, 4), 0), buf10, reinterpret_tensor(buf12, (4, 4, 4), (16, 1, 4), 0), buf11, buf6, reinterpret_tensor(buf7, (4, 4, 4), (16, 1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import LayerNorm
def scaled_dot_attention(q, k, v, mask=None, noise=0, dropout=lambda x: x):
"""
:param q: queries, (batch, time1, channels1)
:param k: keys, (batch, time2, channels1)
:param v: values, (batch, time2, channels2)
:param mask: boolean mask, (batch, time1, time2)
:param dropout: a dropout function - this allows keeping dropout as a module -> better control when training/eval
:return: (batch, time1, channels2), (batch, time1, time2)
"""
weights = torch.matmul(q, k.transpose(2, 1))
if mask is not None:
weights = weights.masked_fill(~mask, float('-inf'))
if noise:
weights += noise * torch.randn(weights.shape)
weights = torch.softmax(weights, dim=-1)
weights = dropout(weights)
result = torch.matmul(weights, v)
return result, weights
def mask(x, lengths, dim=-1):
assert dim != 0, 'Masking not available for batch dimension'
assert len(lengths) == x.shape[0
], 'Lengths must contain as many elements as there are items in the batch'
lengths = torch.as_tensor(lengths)
to_expand = [1] * (x.ndim - 1) + [-1]
mask = torch.arange(x.shape[dim]).expand(to_expand).transpose(dim, -1
).expand(x.shape)
mask = mask < lengths.expand(to_expand).transpose(0, -1)
return mask
class Conv1d(nn.Conv1d):
"""A wrapper around nn.Conv1d, that works on (batch, time, channels)"""
def __init__(self, in_channels, out_channels, kernel_size=1, stride=1,
dilation=1, groups=1, bias=True, padding=0):
super(Conv1d, self).__init__(in_channels=in_channels, out_channels=
out_channels, kernel_size=kernel_size, stride=stride, dilation=
dilation, groups=groups, bias=bias, padding=padding)
def forward(self, x):
return super().forward(x.transpose(2, 1)).transpose(2, 1)
class ScaledDotAttention(nn.Module):
def __init__(self, in_channels, hidden_channels, out_channels, noise=0,
normalize=False, dropout=False):
super(ScaledDotAttention, self).__init__()
self.noise = noise
self.dropout = torch.nn.Dropout(p=dropout)
self.normalize = normalize
self.fc_query = Conv1d(in_channels, hidden_channels)
self.fc_keys = Conv1d(in_channels, hidden_channels)
if normalize:
self.qnorm = LayerNorm(in_channels)
self.knorm = LayerNorm(in_channels)
self.fc_keys.weight = torch.nn.Parameter(self.fc_query.weight.clone())
self.fc_keys.bias = torch.nn.Parameter(self.fc_query.bias.clone())
self.fc_values = Conv1d(in_channels, hidden_channels)
self.fc_out = Conv1d(hidden_channels, out_channels)
def forward(self, q, k, v, mask=None):
"""
:param q: queries, (batch, time1, channels1)
:param k: keys, (batch, time2, channels1)
:param v: values, (batch, time2, channels2)
:param mask: boolean mask, (batch, time1, time2)
:return: (batch, time1, channels2), (batch, time1, time2)
"""
noise = self.noise if self.training else 0
if self.normalize:
q = self.qnorm(q)
k = self.knorm(k)
alignment, weights = scaled_dot_attention(self.fc_query(q), self.
fc_keys(k), self.fc_values(v), mask, noise=noise, dropout=self.
dropout)
alignment = self.fc_out(alignment)
return alignment, weights
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'in_channels': 4, 'hidden_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
from torch.nn import LayerNorm
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_ptr0, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 4 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_5, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_8, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_11, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(16, 4)](primals_1, buf0, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4), (16, 4, 1))
buf2 = buf0
del buf0
triton_poi_fused_convolution_0[grid(16, 4)](primals_4, buf2, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
buf3 = extern_kernels.convolution(buf2, primals_5, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4), (16, 4, 1))
buf4 = buf2
del buf2
triton_poi_fused_convolution_0[grid(16, 4)](primals_7, buf4, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
buf5 = extern_kernels.convolution(buf4, primals_8, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4), (16, 4, 1))
buf6 = buf1
del buf1
triton_poi_fused_convolution_1[grid(64)](buf6, primals_3, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_3
buf7 = buf3
del buf3
triton_poi_fused_convolution_1[grid(64)](buf7, primals_6, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_6
buf8 = buf4
del buf4
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 4, 4), (16, 1, 4),
0), buf7, out=buf8)
buf9 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(64)](buf8, buf9, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf10 = buf8
del buf8
triton_poi_fused__softmax_3[grid(64)](buf9, buf10, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf11 = buf5
del buf5
triton_poi_fused_convolution_1[grid(64)](buf11, primals_9, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_9
buf12 = buf9
del buf9
extern_kernels.bmm(buf10, reinterpret_tensor(buf11, (4, 4, 4), (16,
1, 4), 0), out=buf12)
buf13 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_convolution_0[grid(16, 4)](buf12, buf13, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
buf14 = extern_kernels.convolution(buf13, primals_10, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf14, (4, 4, 4), (16, 4, 1))
del buf13
buf15 = buf14
del buf14
triton_poi_fused_convolution_1[grid(64)](buf15, primals_11, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_11
return reinterpret_tensor(buf15, (4, 4, 4), (16, 1, 4), 0
), buf10, primals_2, primals_5, primals_8, primals_10, reinterpret_tensor(
primals_1, (4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_4,
(4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_7, (4, 4, 4),
(16, 1, 4), 0), buf10, reinterpret_tensor(buf12, (4, 4, 4), (16, 1,
4), 0), buf11, buf6, reinterpret_tensor(buf7, (4, 4, 4), (16, 1, 4), 0)
def scaled_dot_attention(q, k, v, mask=None, noise=0, dropout=lambda x: x):
"""
:param q: queries, (batch, time1, channels1)
:param k: keys, (batch, time2, channels1)
:param v: values, (batch, time2, channels2)
:param mask: boolean mask, (batch, time1, time2)
:param dropout: a dropout function - this allows keeping dropout as a module -> better control when training/eval
:return: (batch, time1, channels2), (batch, time1, time2)
"""
weights = torch.matmul(q, k.transpose(2, 1))
if mask is not None:
weights = weights.masked_fill(~mask, float('-inf'))
if noise:
weights += noise * torch.randn(weights.shape)
weights = torch.softmax(weights, dim=-1)
weights = dropout(weights)
result = torch.matmul(weights, v)
return result, weights
def mask(x, lengths, dim=-1):
assert dim != 0, 'Masking not available for batch dimension'
assert len(lengths) == x.shape[0
], 'Lengths must contain as many elements as there are items in the batch'
lengths = torch.as_tensor(lengths)
to_expand = [1] * (x.ndim - 1) + [-1]
mask = torch.arange(x.shape[dim]).expand(to_expand).transpose(dim, -1
).expand(x.shape)
mask = mask < lengths.expand(to_expand).transpose(0, -1)
return mask
class Conv1d(nn.Conv1d):
"""A wrapper around nn.Conv1d, that works on (batch, time, channels)"""
def __init__(self, in_channels, out_channels, kernel_size=1, stride=1,
dilation=1, groups=1, bias=True, padding=0):
super(Conv1d, self).__init__(in_channels=in_channels, out_channels=
out_channels, kernel_size=kernel_size, stride=stride, dilation=
dilation, groups=groups, bias=bias, padding=padding)
def forward(self, x):
return super().forward(x.transpose(2, 1)).transpose(2, 1)
class ScaledDotAttentionNew(nn.Module):
def __init__(self, in_channels, hidden_channels, out_channels, noise=0,
normalize=False, dropout=False):
super(ScaledDotAttentionNew, self).__init__()
self.noise = noise
self.dropout = torch.nn.Dropout(p=dropout)
self.normalize = normalize
self.fc_query = Conv1d(in_channels, hidden_channels)
self.fc_keys = Conv1d(in_channels, hidden_channels)
if normalize:
self.qnorm = LayerNorm(in_channels)
self.knorm = LayerNorm(in_channels)
self.fc_keys.weight = torch.nn.Parameter(self.fc_query.weight.clone())
self.fc_keys.bias = torch.nn.Parameter(self.fc_query.bias.clone())
self.fc_values = Conv1d(in_channels, hidden_channels)
self.fc_out = Conv1d(hidden_channels, out_channels)
def forward(self, input_0, input_1, input_2):
primals_2 = self.fc_query.weight
primals_3 = self.fc_query.bias
primals_5 = self.fc_keys.weight
primals_6 = self.fc_keys.bias
primals_8 = self.fc_values.weight
primals_9 = self.fc_values.bias
primals_10 = self.fc_out.weight
primals_11 = self.fc_out.bias
primals_1 = input_0
primals_4 = input_1
primals_7 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0], output[1]
|
iclementine/speedyspeech
|
ScaledDotAttention
| false | 10,418 |
[
"BSD-3-Clause"
] | 0 |
db527587a3699b71082d61c9e9fad7ed795d1980
|
https://github.com/iclementine/speedyspeech/tree/db527587a3699b71082d61c9e9fad7ed795d1980
|
Decoder
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/md/cmd3ewacyhu5w5hausgbjbmtnt5rr66cgczh4ibdypq7dz6p4v7g.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (128, 128), (128, 1))
assert_size_stride(primals_5, (128, ), (1, ))
assert_size_stride(primals_6, (4, 128), (128, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf6, 8192, grid=grid(8192), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(primals_4, (128, 128), (1, 128), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf2 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf5, 8192, grid=grid(8192), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 128), (128, 1), 0), reinterpret_tensor(primals_6, (128, 4), (1, 128), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(buf3, (64, 128), (128, 1), 0), primals_6, buf5, primals_4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((128, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
from torch import nn
def weights_init_(m):
if isinstance(m, nn.Linear):
torch.nn.init.xavier_uniform_(m.weight, gain=1)
torch.nn.init.constant_(m.bias, 0)
class Decoder(torch.nn.Module):
def __init__(self, input_dim, out_dim, hidden_size=128):
super(Decoder, self).__init__()
self.linear1 = torch.nn.Linear(input_dim, hidden_size)
self.linear2 = torch.nn.Linear(hidden_size, hidden_size)
self.linear3 = torch.nn.Linear(hidden_size, out_dim)
self.apply(weights_init_)
def forward(self, x):
x = F.relu(self.linear1(x))
x = F.relu(self.linear2(x))
x = self.linear3(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'out_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (128, 128), (128, 1))
assert_size_stride(primals_5, (128,), (1,))
assert_size_stride(primals_6, (4, 128), (128, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(8192)](buf1,
primals_2, buf6, 8192, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 128), (128, 1), 0),
reinterpret_tensor(primals_4, (128, 128), (1, 128), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf2
buf5 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(8192)](buf3,
primals_5, buf5, 8192, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 128),
(128, 1), 0), reinterpret_tensor(primals_6, (128, 4), (1, 128),
0), alpha=1, beta=1, out=buf4)
del primals_7
return reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 128), (128, 1), 0
), reinterpret_tensor(buf3, (64, 128), (128, 1), 0
), primals_6, buf5, primals_4, buf6
def weights_init_(m):
if isinstance(m, nn.Linear):
torch.nn.init.xavier_uniform_(m.weight, gain=1)
torch.nn.init.constant_(m.bias, 0)
class DecoderNew(torch.nn.Module):
def __init__(self, input_dim, out_dim, hidden_size=128):
super(DecoderNew, self).__init__()
self.linear1 = torch.nn.Linear(input_dim, hidden_size)
self.linear2 = torch.nn.Linear(hidden_size, hidden_size)
self.linear3 = torch.nn.Linear(hidden_size, out_dim)
self.apply(weights_init_)
def forward(self, input_0):
primals_1 = self.linear1.weight
primals_2 = self.linear1.bias
primals_4 = self.linear2.weight
primals_5 = self.linear2.bias
primals_6 = self.linear3.weight
primals_7 = self.linear3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
jiaj15/SAIL
|
Decoder
| false | 10,419 |
[
"MIT"
] | 0 |
734be06a2b0ae70801f59c191b86332592da97cf
|
https://github.com/jiaj15/SAIL/tree/734be06a2b0ae70801f59c191b86332592da97cf
|
PolicyNetwork
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/6o/c6o7ainbzocsswla76yvmdsc5donraaar3dzlx2icwrueb7fc46u.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xk/cxkugsynlmnyrjhah42fewrhwovuvurnuv2qimo2qhxq27wjmq7q.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# x_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x3), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jf/cjfzp64ny4hf7wdw5wptah3hqv5fcsh5rrw4brz7uxcy6ad57n7h.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# x_1 => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (256, 4), (4, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 256), (256, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 256), (4096, 1024, 256, 1), 0); del buf0 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf5, 16384, grid=grid(16384), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 256), (256, 1), 0), reinterpret_tensor(primals_4, (256, 4), (1, 256), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf3
return (buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 256), (256, 1), 0), buf4, primals_4, buf5, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((256, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class PolicyNetwork(nn.Module):
def __init__(self, num_inputs, num_actions, hidden_size=256):
super(PolicyNetwork, self).__init__()
self.num_actions = num_actions
self.linear1 = nn.Linear(num_inputs, hidden_size)
self.linear2 = nn.Linear(hidden_size, num_actions)
def forward(self, state):
x = F.relu(self.linear1(state))
x = F.softmax(self.linear2(x), dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_inputs': 4, 'num_actions': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x3, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (256, 4), (4, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 256), (256, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 256), (4096, 1024, 256, 1), 0
)
del buf0
buf5 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(16384)](buf1,
primals_2, buf5, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 256),
(256, 1), 0), reinterpret_tensor(primals_4, (256, 4), (1, 256),
0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf2, buf3, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
triton_poi_fused__softmax_2[grid(256)](buf3, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf3
return buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 256), (256, 1), 0
), buf4, primals_4, buf5
class PolicyNetworkNew(nn.Module):
def __init__(self, num_inputs, num_actions, hidden_size=256):
super(PolicyNetworkNew, self).__init__()
self.num_actions = num_actions
self.linear1 = nn.Linear(num_inputs, hidden_size)
self.linear2 = nn.Linear(hidden_size, num_actions)
def forward(self, input_0):
primals_1 = self.linear1.weight
primals_2 = self.linear1.bias
primals_4 = self.linear2.weight
primals_5 = self.linear2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
linklab/link_rl_book_codes
|
PolicyNetwork
| false | 10,420 |
[
"MIT"
] | 0 |
b272b46d5ecd2802f34648440ff53641c68cbbf0
|
https://github.com/linklab/link_rl_book_codes/tree/b272b46d5ecd2802f34648440ff53641c68cbbf0
|
ActorCriticNetwork
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/6o/c6o7ainbzocsswla76yvmdsc5donraaar3dzlx2icwrueb7fc46u.py
# Topologically Sorted Source Nodes: [value], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# value => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xk/cxkugsynlmnyrjhah42fewrhwovuvurnuv2qimo2qhxq27wjmq7q.py
# Topologically Sorted Source Nodes: [policy_dist_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# policy_dist_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_7, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_7, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x3), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jf/cjfzp64ny4hf7wdw5wptah3hqv5fcsh5rrw4brz7uxcy6ad57n7h.py
# Topologically Sorted Source Nodes: [policy_dist_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# policy_dist_1 => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (256, 4), (4, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (1, 256), (256, 1))
assert_size_stride(primals_5, (1, ), (1, ))
assert_size_stride(primals_6, (256, 4), (4, 1))
assert_size_stride(primals_7, (256, ), (1, ))
assert_size_stride(primals_8, (4, 256), (256, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 256), (4096, 1024, 256, 1), 0); del buf0 # reuse
buf10 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
# Topologically Sorted Source Nodes: [value], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf10, 16384, grid=grid(16384), stream=stream0)
del primals_2
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [value_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 256), (256, 1), 0), reinterpret_tensor(primals_4, (256, 1), (1, 256), 0), alpha=1, beta=1, out=buf3)
del primals_5
buf4 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 256), (1, 4), 0), out=buf4)
del primals_6
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 256), (4096, 1024, 256, 1), 0); del buf4 # reuse
buf9 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
# Topologically Sorted Source Nodes: [policy_dist], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf5, primals_7, buf9, 16384, grid=grid(16384), stream=stream0)
del primals_7
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf5, (64, 256), (256, 1), 0), reinterpret_tensor(primals_8, (256, 4), (1, 256), 0), alpha=1, beta=1, out=buf6)
del primals_9
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [policy_dist_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
buf8 = reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf6 # reuse
# Topologically Sorted Source Nodes: [policy_dist_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf7, buf8, 256, grid=grid(256), stream=stream0)
del buf7
return (reinterpret_tensor(buf3, (4, 4, 4, 1), (16, 4, 1, 1), 0), buf8, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 256), (256, 1), 0), reinterpret_tensor(buf5, (64, 256), (256, 1), 0), buf8, primals_8, buf9, primals_4, buf10, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((256, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((256, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class ActorCriticNetwork(nn.Module):
def __init__(self, num_inputs, num_actions, hidden_size=256):
super(ActorCriticNetwork, self).__init__()
self.num_actions = num_actions
self.critic_linear1 = nn.Linear(num_inputs, hidden_size)
self.critic_linear2 = nn.Linear(hidden_size, 1)
self.actor_linear1 = nn.Linear(num_inputs, hidden_size)
self.actor_linear2 = nn.Linear(hidden_size, num_actions)
def forward(self, state_tensor):
value = F.relu(self.critic_linear1(state_tensor))
value = self.critic_linear2(value)
policy_dist = F.relu(self.actor_linear1(state_tensor))
policy_dist = F.softmax(self.actor_linear2(policy_dist), dim=1)
return value, policy_dist
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_inputs': 4, 'num_actions': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x3, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (256, 4), (4, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (1, 256), (256, 1))
assert_size_stride(primals_5, (1,), (1,))
assert_size_stride(primals_6, (256, 4), (4, 1))
assert_size_stride(primals_7, (256,), (1,))
assert_size_stride(primals_8, (4, 256), (256, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 256), (4096, 1024, 256, 1), 0
)
del buf0
buf10 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(16384)](buf1,
primals_2, buf10, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 256),
(256, 1), 0), reinterpret_tensor(primals_4, (256, 1), (1, 256),
0), alpha=1, beta=1, out=buf3)
del primals_5
buf4 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 256), (1, 4), 0), out=buf4)
del primals_6
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 256), (4096, 1024, 256, 1), 0
)
del buf4
buf9 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(16384)](buf5,
primals_7, buf9, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del primals_7
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_9, reinterpret_tensor(buf5, (64, 256),
(256, 1), 0), reinterpret_tensor(primals_8, (256, 4), (1, 256),
0), alpha=1, beta=1, out=buf6)
del primals_9
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf8 = reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf6
triton_poi_fused__softmax_2[grid(256)](buf7, buf8, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf7
return reinterpret_tensor(buf3, (4, 4, 4, 1), (16, 4, 1, 1), 0
), buf8, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 256), (256, 1), 0
), reinterpret_tensor(buf5, (64, 256), (256, 1), 0
), buf8, primals_8, buf9, primals_4, buf10
class ActorCriticNetworkNew(nn.Module):
def __init__(self, num_inputs, num_actions, hidden_size=256):
super(ActorCriticNetworkNew, self).__init__()
self.num_actions = num_actions
self.critic_linear1 = nn.Linear(num_inputs, hidden_size)
self.critic_linear2 = nn.Linear(hidden_size, 1)
self.actor_linear1 = nn.Linear(num_inputs, hidden_size)
self.actor_linear2 = nn.Linear(hidden_size, num_actions)
def forward(self, input_0):
primals_1 = self.critic_linear1.weight
primals_2 = self.critic_linear1.bias
primals_4 = self.critic_linear2.weight
primals_5 = self.critic_linear2.bias
primals_6 = self.actor_linear1.weight
primals_7 = self.actor_linear1.bias
primals_8 = self.actor_linear2.weight
primals_9 = self.actor_linear2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0], output[1]
|
linklab/link_rl_book_codes
|
ActorCriticNetwork
| false | 10,421 |
[
"MIT"
] | 0 |
b272b46d5ecd2802f34648440ff53641c68cbbf0
|
https://github.com/linklab/link_rl_book_codes/tree/b272b46d5ecd2802f34648440ff53641c68cbbf0
|
SE
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/l3/cl35tzbhrd24dhunkbb6gjs54aklpyr46oikqhoylcgmkcmhujil.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# out => mean
# Graph fragment:
# %mean : [num_users=2] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [-1, -2], True), kwargs = {})
triton_per_fused_mean_0 = async_compile.triton('triton_per_fused_mean_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 16.0
tmp6 = tmp4 / tmp5
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6u/c6utrmzep5ziyaxjrlcwvhjg7wjjd7tfipeakfphf2az5qxhw3yf.py
# Topologically Sorted Source Nodes: [conv2d, sigmoid, out_1], Original ATen: [aten.convolution, aten.sigmoid, aten.mul]
# Source node to ATen node mapping:
# conv2d => convolution
# out_1 => mul
# sigmoid => sigmoid
# Graph fragment:
# %convolution : [num_users=3] = call_function[target=torch.ops.aten.convolution.default](args = (%mean, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution,), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, %sigmoid), kwargs = {})
triton_poi_fused_convolution_mul_sigmoid_1 = async_compile.triton('triton_poi_fused_convolution_mul_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_mul_sigmoid_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_mul_sigmoid_1(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp4 = tmp2 * tmp3
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
tl.store(out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/k2/ck2mamkqpmuzem4n3p4ij6fmfpy2bcbblg6sx6wwslgqwuqq5ifh.py
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%mul, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/lp/clprvnh5p6cmadxtwzizwydrpjlwxohxixbw4ntucp6srbu6gtis.py
# Topologically Sorted Source Nodes: [out_2, out_3], Original ATen: [aten.sigmoid, aten.mul]
# Source node to ATen node mapping:
# out_2 => sigmoid_1
# out_3 => mul_1
# Graph fragment:
# %sigmoid_1 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_1,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, %sigmoid_1), kwargs = {})
triton_poi_fused_mul_sigmoid_3 = async_compile.triton('triton_poi_fused_mul_sigmoid_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_sigmoid_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_sigmoid_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tl.sigmoid(tmp1)
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_mean_0.run(buf1, primals_1, 16, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 1, 1), (4, 1, 1, 1))
buf3 = buf2; del buf2 # reuse
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv2d, sigmoid, out_1], Original ATen: [aten.convolution, aten.sigmoid, aten.mul]
triton_poi_fused_convolution_mul_sigmoid_1.run(buf3, primals_3, buf4, 16, grid=grid(16), stream=stream0)
del primals_3
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 1, 1), (4, 1, 1, 1))
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf6, primals_5, 16, grid=grid(16), stream=stream0)
del primals_5
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2, out_3], Original ATen: [aten.sigmoid, aten.mul]
triton_poi_fused_mul_sigmoid_3.run(primals_1, buf6, buf7, 256, grid=grid(256), stream=stream0)
return (buf7, primals_1, primals_2, primals_4, buf1, buf3, buf4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
def swish(x):
return x * x.sigmoid()
class SE(nn.Module):
"""Squeeze-and-Excitation block with Swish."""
def __init__(self, in_planes, se_planes):
super(SE, self).__init__()
self.se1 = nn.Conv2d(in_planes, se_planes, kernel_size=1, bias=True)
self.se2 = nn.Conv2d(se_planes, in_planes, kernel_size=1, bias=True)
def forward(self, x):
out = F.adaptive_avg_pool2d(x, (1, 1))
out = swish(self.se1(out))
out = self.se2(out).sigmoid()
out = x * out
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_planes': 4, 'se_planes': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mean_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 16.0
tmp6 = tmp4 / tmp5
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp6, xmask)
@triton.jit
def triton_poi_fused_convolution_mul_sigmoid_1(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp4 = tmp2 * tmp3
tl.store(in_out_ptr0 + x2, tmp2, xmask)
tl.store(out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, xmask)
@triton.jit
def triton_poi_fused_mul_sigmoid_3(in_ptr0, in_ptr1, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 16
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tl.sigmoid(tmp1)
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + x2, tmp3, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf0
get_raw_stream(0)
triton_per_fused_mean_0[grid(16)](buf1, primals_1, 16, 16, XBLOCK=8,
num_warps=2, num_stages=1)
buf2 = extern_kernels.convolution(buf1, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 1, 1), (4, 1, 1, 1))
buf3 = buf2
del buf2
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
triton_poi_fused_convolution_mul_sigmoid_1[grid(16)](buf3,
primals_3, buf4, 16, XBLOCK=16, num_warps=1, num_stages=1)
del primals_3
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 1, 1), (4, 1, 1, 1))
buf6 = buf5
del buf5
triton_poi_fused_convolution_2[grid(16)](buf6, primals_5, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_5
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_mul_sigmoid_3[grid(256)](primals_1, buf6, buf7,
256, XBLOCK=256, num_warps=4, num_stages=1)
return buf7, primals_1, primals_2, primals_4, buf1, buf3, buf4, buf6
def swish(x):
return x * x.sigmoid()
class SENew(nn.Module):
"""Squeeze-and-Excitation block with Swish."""
def __init__(self, in_planes, se_planes):
super(SENew, self).__init__()
self.se1 = nn.Conv2d(in_planes, se_planes, kernel_size=1, bias=True)
self.se2 = nn.Conv2d(se_planes, in_planes, kernel_size=1, bias=True)
def forward(self, input_0):
primals_2 = self.se1.weight
primals_3 = self.se1.bias
primals_4 = self.se2.weight
primals_5 = self.se2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
liormagram/pytorch-cifar
|
SE
| false | 10,422 |
[
"MIT"
] | 0 |
2ed0fabe6cbd4a468c5c4d155fb76c5b9ad4a764
|
https://github.com/liormagram/pytorch-cifar/tree/2ed0fabe6cbd4a468c5c4d155fb76c5b9ad4a764
|
MultiHeadQKVAttention
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wb/cwbnzhah456ssqf5k4ob4llmezbi7o6givtx6ppibkklzj3kmilo.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + (16*y0)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wl/cwll6lomrruls6u4c3ua52p4g5jvoxwf6bvyj4dolvf23rk3zogz.py
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_3 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf0)
del primals_4
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf1)
del primals_6
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf2)
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(buf0, primals_5, buf3, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_5
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf1, primals_7, buf4, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_7
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [contiguous_2], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf2, primals_9, buf8, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_9
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [o], Original ATen: [aten.bmm]
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [linear_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf11)
del primals_11
return (reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), primals_10, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'d_k': 4, 'd_v': 4, 'n_heads': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + 16 * y0), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf0)
del primals_4
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf1)
del primals_6
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf2)
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(4, 16)](buf0, primals_5, buf3, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_5
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf0
triton_poi_fused_clone_0[grid(4, 16)](buf1, primals_7, buf4, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_7
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf5, buf6, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf7 = buf5
del buf5
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_0[grid(4, 16)](buf2, primals_9, buf8, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_9
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0)
del buf2
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1,
0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_3[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf11)
del primals_11
return reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0
), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_3, (16, 4), (4, 1), 0
), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0
), primals_10, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0)
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttentionNew(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, input_0, input_1, input_2):
primals_4 = self.q_projector.weight
primals_5 = self.q_projector.bias
primals_6 = self.k_projector.weight
primals_7 = self.k_projector.bias
primals_8 = self.v_projector.weight
primals_9 = self.v_projector.bias
primals_10 = self.o_projector.weight
primals_11 = self.o_projector.bias
primals_1 = input_0
primals_2 = input_1
primals_3 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
|
karayanni/torch-scae
|
MultiHeadQKVAttention
| false | 10,423 |
[
"Apache-2.0"
] | 0 |
e044662d8942d8d1923d13d071f375144cf4a1e8
|
https://github.com/karayanni/torch-scae/tree/e044662d8942d8d1923d13d071f375144cf4a1e8
|
AFMLayer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/oj/coji63cmjptmfiahnhfxrcymtijnwomdesxsksu5cd5o6hnjtmkc.py
# Topologically Sorted Source Nodes: [p, q, inner_product], Original ATen: [aten.cat, aten.mul]
# Source node to ATen node mapping:
# inner_product => mul
# p => cat
# q => cat_1
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select, %select, %select, %select_1, %select_1, %select_2], 1), kwargs = {})
# %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_1, %select_2, %select_3, %select_2, %select_3, %select_3], 1), kwargs = {})
# %mul : [num_users=3] = call_function[target=torch.ops.aten.mul.Tensor](args = (%cat, %cat_1), kwargs = {})
triton_poi_fused_cat_mul_0 = async_compile.triton('triton_poi_fused_cat_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_mul_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_mul_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4) % 24
x0 = xindex % 4
x2 = (xindex // 96)
x3 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (4*x1) + (16*x2)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr0 + (x0 + (4*((-4) + x1)) + (16*x2)), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr0 + (x0 + (4*((-8) + x1)) + (16*x2)), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr0 + (64 + x0 + (4*((-12) + x1)) + (16*x2)), tmp19 & xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tmp22 = tl.full([1], 20, tl.int64)
tmp23 = tmp0 < tmp22
tmp24 = tmp21 & tmp23
tmp25 = tl.load(in_ptr0 + (64 + x0 + (4*((-16) + x1)) + (16*x2)), tmp24 & xmask, other=0.0)
tmp26 = tmp0 >= tmp22
tmp27 = tl.full([1], 24, tl.int64)
tmp28 = tmp0 < tmp27
tmp29 = tl.load(in_ptr0 + (128 + x0 + (4*((-20) + x1)) + (16*x2)), tmp26 & xmask, other=0.0)
tmp30 = tl.where(tmp24, tmp25, tmp29)
tmp31 = tl.where(tmp19, tmp20, tmp30)
tmp32 = tl.where(tmp14, tmp15, tmp31)
tmp33 = tl.where(tmp9, tmp10, tmp32)
tmp34 = tl.where(tmp4, tmp5, tmp33)
tmp35 = tl.load(in_ptr0 + (64 + x0 + (4*x1) + (16*x2)), tmp4 & xmask, other=0.0)
tmp36 = tl.load(in_ptr0 + (128 + x0 + (4*((-4) + x1)) + (16*x2)), tmp9 & xmask, other=0.0)
tmp37 = tl.load(in_ptr0 + (192 + x0 + (4*((-8) + x1)) + (16*x2)), tmp14 & xmask, other=0.0)
tmp38 = tl.load(in_ptr0 + (128 + x0 + (4*((-12) + x1)) + (16*x2)), tmp19 & xmask, other=0.0)
tmp39 = tl.load(in_ptr0 + (192 + x0 + (4*((-16) + x1)) + (16*x2)), tmp24 & xmask, other=0.0)
tmp40 = tl.load(in_ptr0 + (192 + x0 + (4*((-20) + x1)) + (16*x2)), tmp26 & xmask, other=0.0)
tmp41 = tl.where(tmp24, tmp39, tmp40)
tmp42 = tl.where(tmp19, tmp38, tmp41)
tmp43 = tl.where(tmp14, tmp37, tmp42)
tmp44 = tl.where(tmp9, tmp36, tmp43)
tmp45 = tl.where(tmp4, tmp35, tmp44)
tmp46 = tmp34 * tmp45
tl.store(in_out_ptr0 + (x3), tmp46, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/j2/cj26ownu73m72kwjlseu3qfwtrz4f3ru464aa4zuhodtujlnjupm.py
# Topologically Sorted Source Nodes: [add, attention_temp], Original ATen: [aten.add, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# add => add
# attention_temp => relu
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_2, %primals_3), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_1(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/p4/cp4mdcdve4y73ad5mzhckzksofhes3a2n2zye5hynnmbc62ct27d.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax, div, exp, sub, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_per_fused__softmax_2 = async_compile.triton('triton_per_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 32],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__softmax_2(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (24*x0)), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float("-inf"))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + (24*x0)), tmp11, rmask & xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ui/cui4pbynpryqmgmjhsdzeompa6sltxsmy5ggopxiaqdlvyafsjpl.py
# Topologically Sorted Source Nodes: [mul_1, attention_output], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# attention_output => sum_2
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %mul), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_1, [1]), kwargs = {})
triton_per_fused_mul_sum_3 = async_compile.triton('triton_per_fused_mul_sum_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 32],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_sum_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_sum_3(in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r2 = rindex
x1 = (xindex // 4)
x0 = xindex % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + (r2 + (24*x1)), rmask & xmask, eviction_policy='evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (x0 + (4*r2) + (96*x1)), rmask & xmask, other=0.0)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(rmask & xmask, tmp3, 0)
tmp6 = tl.sum(tmp5, 1)[:, None]
tl.store(out_ptr0 + (x3), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 1), (1, 1))
assert_size_stride(primals_5, (4, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 24, 4), (96, 4, 1), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [p, q, inner_product], Original ATen: [aten.cat, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_mul_0.run(buf2, primals_1, 384, grid=grid(384), stream=stream0)
del primals_1
buf3 = empty_strided_cuda((96, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tensordot], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf2, (96, 4), (4, 1), 0), primals_2, out=buf3)
del primals_2
buf4 = reinterpret_tensor(buf3, (4, 24, 4), (96, 4, 1), 0); del buf3 # reuse
buf11 = empty_strided_cuda((4, 24, 4), (96, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [add, attention_temp], Original ATen: [aten.add, aten.relu, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_1.run(buf4, primals_3, buf11, 384, grid=grid(384), stream=stream0)
del primals_3
buf5 = empty_strided_cuda((96, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [tensordot_1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf4, (96, 4), (4, 1), 0), primals_4, out=buf5)
buf8 = empty_strided_cuda((4, 24, 1), (24, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_per_fused__softmax_2.run(buf5, buf8, 4, 24, grid=grid(4), stream=stream0)
del buf5
buf9 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, attention_output], Original ATen: [aten.mul, aten.sum]
triton_per_fused_mul_sum_3.run(buf8, buf2, buf9, 16, 24, grid=grid(16), stream=stream0)
buf10 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [afm_out], Original ATen: [aten.mm]
extern_kernels.mm(buf9, primals_5, out=buf10)
return (buf10, buf8, buf2, buf8, reinterpret_tensor(buf9, (4, 4), (1, 4), 0), reinterpret_tensor(primals_5, (1, 4), (1, 1), 0), reinterpret_tensor(buf4, (4, 96), (1, 4), 0), reinterpret_tensor(primals_4, (1, 4), (1, 1), 0), buf11, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 1), (1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 1), (1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import itertools
import torch
import torch.nn as nn
import torch.nn.functional as F
from sklearn.metrics import *
import torch.onnx
import torch as torch
class AFMLayer(nn.Module):
"""Attentonal Factorization Machine models pairwise (order-2) feature
interactions without linear term and bias.
Input shape
- A list of 3D tensor with shape: ``(batch_size,1,embedding_size)``.
Output shape
- 2D tensor with shape: ``(batch_size, 1)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **attention_factor** : Positive integer, dimensionality of the
attention network output space.
- **l2_reg_w** : float between 0 and 1. L2 regularizer strength
applied to attention network.
- **dropout_rate** : float between in [0,1). Fraction of the attention net output units to dropout.
- **seed** : A Python integer to use as random seed.
References
- [Attentional Factorization Machines : Learning the Weight of Feature
Interactions via Attention Networks](https://arxiv.org/pdf/1708.04617.pdf)
"""
def __init__(self, in_features, attention_factor=4, l2_reg_w=0,
dropout_rate=0, seed=1024, device='cpu'):
super(AFMLayer, self).__init__()
self.attention_factor = attention_factor
self.l2_reg_w = l2_reg_w
self.dropout_rate = dropout_rate
self.seed = seed
embedding_size = in_features
self.attention_W = nn.Parameter(torch.Tensor(embedding_size, self.
attention_factor))
self.attention_b = nn.Parameter(torch.Tensor(self.attention_factor))
self.projection_h = nn.Parameter(torch.Tensor(self.attention_factor, 1)
)
self.projection_p = nn.Parameter(torch.Tensor(embedding_size, 1))
for tensor in [self.attention_W, self.projection_h, self.projection_p]:
nn.init.xavier_normal_(tensor)
for tensor in [self.attention_b]:
nn.init.zeros_(tensor)
self.dropout = nn.Dropout(dropout_rate)
self
def forward(self, inputs):
embeds_vec_list = inputs
row = []
col = []
for r, c in itertools.combinations(embeds_vec_list, 2):
row.append(r)
col.append(c)
p = torch.cat(row, dim=1)
q = torch.cat(col, dim=1)
inner_product = p * q
bi_interaction = inner_product
attention_temp = F.relu(torch.tensordot(bi_interaction, self.
attention_W, dims=([-1], [0])) + self.attention_b)
self.normalized_att_score = F.softmax(torch.tensordot(
attention_temp, self.projection_h, dims=([-1], [0])), dim=1)
attention_output = torch.sum(self.normalized_att_score *
bi_interaction, dim=1)
attention_output = self.dropout(attention_output)
afm_out = torch.tensordot(attention_output, self.projection_p, dims
=([-1], [0]))
return afm_out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_features': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
from sklearn.metrics import *
import torch.onnx
import torch as torch
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_mul_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4 % 24
x0 = xindex % 4
x2 = xindex // 96
x3 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 4 * x1 + 16 * x2), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr0 + (x0 + 4 * (-4 + x1) + 16 * x2), tmp9 & xmask,
other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr0 + (x0 + 4 * (-8 + x1) + 16 * x2), tmp14 & xmask,
other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr0 + (64 + x0 + 4 * (-12 + x1) + 16 * x2), tmp19 &
xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tmp22 = tl.full([1], 20, tl.int64)
tmp23 = tmp0 < tmp22
tmp24 = tmp21 & tmp23
tmp25 = tl.load(in_ptr0 + (64 + x0 + 4 * (-16 + x1) + 16 * x2), tmp24 &
xmask, other=0.0)
tmp26 = tmp0 >= tmp22
tl.full([1], 24, tl.int64)
tmp29 = tl.load(in_ptr0 + (128 + x0 + 4 * (-20 + x1) + 16 * x2), tmp26 &
xmask, other=0.0)
tmp30 = tl.where(tmp24, tmp25, tmp29)
tmp31 = tl.where(tmp19, tmp20, tmp30)
tmp32 = tl.where(tmp14, tmp15, tmp31)
tmp33 = tl.where(tmp9, tmp10, tmp32)
tmp34 = tl.where(tmp4, tmp5, tmp33)
tmp35 = tl.load(in_ptr0 + (64 + x0 + 4 * x1 + 16 * x2), tmp4 & xmask,
other=0.0)
tmp36 = tl.load(in_ptr0 + (128 + x0 + 4 * (-4 + x1) + 16 * x2), tmp9 &
xmask, other=0.0)
tmp37 = tl.load(in_ptr0 + (192 + x0 + 4 * (-8 + x1) + 16 * x2), tmp14 &
xmask, other=0.0)
tmp38 = tl.load(in_ptr0 + (128 + x0 + 4 * (-12 + x1) + 16 * x2), tmp19 &
xmask, other=0.0)
tmp39 = tl.load(in_ptr0 + (192 + x0 + 4 * (-16 + x1) + 16 * x2), tmp24 &
xmask, other=0.0)
tmp40 = tl.load(in_ptr0 + (192 + x0 + 4 * (-20 + x1) + 16 * x2), tmp26 &
xmask, other=0.0)
tmp41 = tl.where(tmp24, tmp39, tmp40)
tmp42 = tl.where(tmp19, tmp38, tmp41)
tmp43 = tl.where(tmp14, tmp37, tmp42)
tmp44 = tl.where(tmp9, tmp36, tmp43)
tmp45 = tl.where(tmp4, tmp35, tmp44)
tmp46 = tmp34 * tmp45
tl.store(in_out_ptr0 + x3, tmp46, xmask)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_1(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_per_fused__softmax_2(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 4
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 24 * x0), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float('-inf'))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + 24 * x0), tmp11, rmask & xmask)
@triton.jit
def triton_per_fused_mul_sum_3(in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 16
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r2 = rindex
x1 = xindex // 4
x0 = xindex % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + (r2 + 24 * x1), rmask & xmask, eviction_policy
='evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (x0 + 4 * r2 + 96 * x1), rmask & xmask, other=0.0)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(rmask & xmask, tmp3, 0)
tmp6 = tl.sum(tmp5, 1)[:, None]
tl.store(out_ptr0 + x3, tmp6, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 1), (1, 1))
assert_size_stride(primals_5, (4, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 24, 4), (96, 4, 1), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_cat_mul_0[grid(384)](buf2, primals_1, 384, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_1
buf3 = empty_strided_cuda((96, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf2, (96, 4), (4, 1), 0),
primals_2, out=buf3)
del primals_2
buf4 = reinterpret_tensor(buf3, (4, 24, 4), (96, 4, 1), 0)
del buf3
buf11 = empty_strided_cuda((4, 24, 4), (96, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_1[grid(384)](buf4,
primals_3, buf11, 384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
buf5 = empty_strided_cuda((96, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf4, (96, 4), (4, 1), 0),
primals_4, out=buf5)
buf8 = empty_strided_cuda((4, 24, 1), (24, 1, 1), torch.float32)
triton_per_fused__softmax_2[grid(4)](buf5, buf8, 4, 24, XBLOCK=1,
num_warps=2, num_stages=1)
del buf5
buf9 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_per_fused_mul_sum_3[grid(16)](buf8, buf2, buf9, 16, 24,
XBLOCK=1, num_warps=2, num_stages=1)
buf10 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
extern_kernels.mm(buf9, primals_5, out=buf10)
return buf10, buf8, buf2, buf8, reinterpret_tensor(buf9, (4, 4), (1, 4), 0
), reinterpret_tensor(primals_5, (1, 4), (1, 1), 0
), reinterpret_tensor(buf4, (4, 96), (1, 4), 0), reinterpret_tensor(
primals_4, (1, 4), (1, 1), 0), buf11
class AFMLayerNew(nn.Module):
"""Attentonal Factorization Machine models pairwise (order-2) feature
interactions without linear term and bias.
Input shape
- A list of 3D tensor with shape: ``(batch_size,1,embedding_size)``.
Output shape
- 2D tensor with shape: ``(batch_size, 1)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **attention_factor** : Positive integer, dimensionality of the
attention network output space.
- **l2_reg_w** : float between 0 and 1. L2 regularizer strength
applied to attention network.
- **dropout_rate** : float between in [0,1). Fraction of the attention net output units to dropout.
- **seed** : A Python integer to use as random seed.
References
- [Attentional Factorization Machines : Learning the Weight of Feature
Interactions via Attention Networks](https://arxiv.org/pdf/1708.04617.pdf)
"""
def __init__(self, in_features, attention_factor=4, l2_reg_w=0,
dropout_rate=0, seed=1024, device='cpu'):
super(AFMLayerNew, self).__init__()
self.attention_factor = attention_factor
self.l2_reg_w = l2_reg_w
self.dropout_rate = dropout_rate
self.seed = seed
embedding_size = in_features
self.attention_W = nn.Parameter(torch.Tensor(embedding_size, self.
attention_factor))
self.attention_b = nn.Parameter(torch.Tensor(self.attention_factor))
self.projection_h = nn.Parameter(torch.Tensor(self.attention_factor, 1)
)
self.projection_p = nn.Parameter(torch.Tensor(embedding_size, 1))
for tensor in [self.attention_W, self.projection_h, self.projection_p]:
nn.init.xavier_normal_(tensor)
for tensor in [self.attention_b]:
nn.init.zeros_(tensor)
self.dropout = nn.Dropout(dropout_rate)
self
def forward(self, input_0):
primals_2 = self.attention_W
primals_3 = self.attention_b
primals_4 = self.projection_h
primals_5 = self.projection_p
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
dulvqingyunLT/DeepCTR-Torch
|
AFMLayer
| false | 10,424 |
[
"Apache-2.0"
] | 0 |
f40cf08f3469aa471f9ca69e44c5de51180341cc
|
https://github.com/dulvqingyunLT/DeepCTR-Torch/tree/f40cf08f3469aa471f9ca69e44c5de51180341cc
|
DRRN
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/77/c773z3ksgrnasomgfhaw574ix3f2o4yv2n3kziqh4wol2y4gnmrr.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=4] = call_function[target=torch.ops.aten.relu.default](args = (%primals_1,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + (x0), tmp2, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ej/cejmpqb52jrqlle33cesamyex5ioj4utjclnb7tlrznwezqa5gnx.py
# Topologically Sorted Source Nodes: [relu_1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# relu_1 => relu_1
# Graph fragment:
# %relu_1 : [num_users=27] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_relu_1 = async_compile.triton('triton_poi_fused_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2097152],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_1(in_out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2097152
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + (x0), tmp2, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/q5/cq5vgxowbn4agnifxmmz3bdax3cn5mjnmy57emu5bjdwqabakzok.py
# Topologically Sorted Source Nodes: [out_1, relu_3], Original ATen: [aten.add, aten.relu]
# Source node to ATen node mapping:
# out_1 => add
# relu_3 => relu_3
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convolution_2, %relu_1), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add,), kwargs = {})
triton_poi_fused_add_relu_2 = async_compile.triton('triton_poi_fused_add_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2097152],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2097152
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), None)
tmp1 = tl.load(in_ptr0 + (x0), None)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x0), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/qf/cqfjpp4227icr7ijrkgfudkoifwsutjs7zhnrdiifjqxsnu7xuij.py
# Topologically Sorted Source Nodes: [out_51], Original ATen: [aten.add]
# Source node to ATen node mapping:
# out_51 => add_25
# Graph fragment:
# %add_25 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convolution_51, %relu), kwargs = {})
# %copy_ : [num_users=0] = call_function[target=torch.ops.aten.copy_.default](args = (%primals_1, %relu), kwargs = {})
triton_poi_fused_add_3 = async_compile.triton('triton_poi_fused_add_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_3', 'mutated_arg_names': ['in_out_ptr0', 'out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_3(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), None)
tmp1 = tl.load(in_ptr0 + (x0), None)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x0), tmp2, None)
tl.store(out_ptr0 + (x0), tmp1, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_2, (128, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_3, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_4, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_5, (1, 128, 3, 3), (1152, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 64, 64), (4096, 4096, 64, 1), torch.float32)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(primals_1, buf0, 16384, grid=grid(16384), stream=stream0)
# Topologically Sorted Source Nodes: [inputs], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [relu_1], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf2, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [relu_2], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf4, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [out_1, relu_3], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf6, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_3], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(buf6, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [relu_4], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf8, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.convolution]
buf9 = extern_kernels.convolution(buf8, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf10 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [out_3, relu_5], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf10, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_5], Original ATen: [aten.convolution]
buf11 = extern_kernels.convolution(buf10, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf12 = buf11; del buf11 # reuse
# Topologically Sorted Source Nodes: [relu_6], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf12, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_4], Original ATen: [aten.convolution]
buf13 = extern_kernels.convolution(buf12, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf13, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf14 = buf13; del buf13 # reuse
# Topologically Sorted Source Nodes: [out_5, relu_7], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf14, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_7], Original ATen: [aten.convolution]
buf15 = extern_kernels.convolution(buf14, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf15, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf16 = buf15; del buf15 # reuse
# Topologically Sorted Source Nodes: [relu_8], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf16, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_6], Original ATen: [aten.convolution]
buf17 = extern_kernels.convolution(buf16, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf18 = buf17; del buf17 # reuse
# Topologically Sorted Source Nodes: [out_7, relu_9], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf18, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_9], Original ATen: [aten.convolution]
buf19 = extern_kernels.convolution(buf18, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf20 = buf19; del buf19 # reuse
# Topologically Sorted Source Nodes: [relu_10], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf20, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_8], Original ATen: [aten.convolution]
buf21 = extern_kernels.convolution(buf20, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf21, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf22 = buf21; del buf21 # reuse
# Topologically Sorted Source Nodes: [out_9, relu_11], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf22, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_11], Original ATen: [aten.convolution]
buf23 = extern_kernels.convolution(buf22, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf24 = buf23; del buf23 # reuse
# Topologically Sorted Source Nodes: [relu_12], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf24, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_10], Original ATen: [aten.convolution]
buf25 = extern_kernels.convolution(buf24, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf25, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf26 = buf25; del buf25 # reuse
# Topologically Sorted Source Nodes: [out_11, relu_13], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf26, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_13], Original ATen: [aten.convolution]
buf27 = extern_kernels.convolution(buf26, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf27, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf28 = buf27; del buf27 # reuse
# Topologically Sorted Source Nodes: [relu_14], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf28, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_12], Original ATen: [aten.convolution]
buf29 = extern_kernels.convolution(buf28, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf29, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf30 = buf29; del buf29 # reuse
# Topologically Sorted Source Nodes: [out_13, relu_15], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf30, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_15], Original ATen: [aten.convolution]
buf31 = extern_kernels.convolution(buf30, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf31, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf32 = buf31; del buf31 # reuse
# Topologically Sorted Source Nodes: [relu_16], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf32, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_14], Original ATen: [aten.convolution]
buf33 = extern_kernels.convolution(buf32, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf33, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf34 = buf33; del buf33 # reuse
# Topologically Sorted Source Nodes: [out_15, relu_17], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf34, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_17], Original ATen: [aten.convolution]
buf35 = extern_kernels.convolution(buf34, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf35, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf36 = buf35; del buf35 # reuse
# Topologically Sorted Source Nodes: [relu_18], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf36, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_16], Original ATen: [aten.convolution]
buf37 = extern_kernels.convolution(buf36, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf37, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf38 = buf37; del buf37 # reuse
# Topologically Sorted Source Nodes: [out_17, relu_19], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf38, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_19], Original ATen: [aten.convolution]
buf39 = extern_kernels.convolution(buf38, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf39, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf40 = buf39; del buf39 # reuse
# Topologically Sorted Source Nodes: [relu_20], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf40, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_18], Original ATen: [aten.convolution]
buf41 = extern_kernels.convolution(buf40, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf41, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf42 = buf41; del buf41 # reuse
# Topologically Sorted Source Nodes: [out_19, relu_21], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf42, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_21], Original ATen: [aten.convolution]
buf43 = extern_kernels.convolution(buf42, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf43, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf44 = buf43; del buf43 # reuse
# Topologically Sorted Source Nodes: [relu_22], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf44, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_20], Original ATen: [aten.convolution]
buf45 = extern_kernels.convolution(buf44, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf45, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf46 = buf45; del buf45 # reuse
# Topologically Sorted Source Nodes: [out_21, relu_23], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf46, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_23], Original ATen: [aten.convolution]
buf47 = extern_kernels.convolution(buf46, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf47, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf48 = buf47; del buf47 # reuse
# Topologically Sorted Source Nodes: [relu_24], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf48, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_22], Original ATen: [aten.convolution]
buf49 = extern_kernels.convolution(buf48, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf49, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf50 = buf49; del buf49 # reuse
# Topologically Sorted Source Nodes: [out_23, relu_25], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf50, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_25], Original ATen: [aten.convolution]
buf51 = extern_kernels.convolution(buf50, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf51, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf52 = buf51; del buf51 # reuse
# Topologically Sorted Source Nodes: [relu_26], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf52, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_24], Original ATen: [aten.convolution]
buf53 = extern_kernels.convolution(buf52, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf53, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf54 = buf53; del buf53 # reuse
# Topologically Sorted Source Nodes: [out_25, relu_27], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf54, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_27], Original ATen: [aten.convolution]
buf55 = extern_kernels.convolution(buf54, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf55, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf56 = buf55; del buf55 # reuse
# Topologically Sorted Source Nodes: [relu_28], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf56, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_26], Original ATen: [aten.convolution]
buf57 = extern_kernels.convolution(buf56, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf57, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf58 = buf57; del buf57 # reuse
# Topologically Sorted Source Nodes: [out_27, relu_29], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf58, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_29], Original ATen: [aten.convolution]
buf59 = extern_kernels.convolution(buf58, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf59, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf60 = buf59; del buf59 # reuse
# Topologically Sorted Source Nodes: [relu_30], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf60, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_28], Original ATen: [aten.convolution]
buf61 = extern_kernels.convolution(buf60, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf61, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf62 = buf61; del buf61 # reuse
# Topologically Sorted Source Nodes: [out_29, relu_31], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf62, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_31], Original ATen: [aten.convolution]
buf63 = extern_kernels.convolution(buf62, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf63, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf64 = buf63; del buf63 # reuse
# Topologically Sorted Source Nodes: [relu_32], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf64, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_30], Original ATen: [aten.convolution]
buf65 = extern_kernels.convolution(buf64, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf65, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf66 = buf65; del buf65 # reuse
# Topologically Sorted Source Nodes: [out_31, relu_33], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf66, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_33], Original ATen: [aten.convolution]
buf67 = extern_kernels.convolution(buf66, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf67, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf68 = buf67; del buf67 # reuse
# Topologically Sorted Source Nodes: [relu_34], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf68, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_32], Original ATen: [aten.convolution]
buf69 = extern_kernels.convolution(buf68, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf69, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf70 = buf69; del buf69 # reuse
# Topologically Sorted Source Nodes: [out_33, relu_35], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf70, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_35], Original ATen: [aten.convolution]
buf71 = extern_kernels.convolution(buf70, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf71, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf72 = buf71; del buf71 # reuse
# Topologically Sorted Source Nodes: [relu_36], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf72, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_34], Original ATen: [aten.convolution]
buf73 = extern_kernels.convolution(buf72, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf73, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf74 = buf73; del buf73 # reuse
# Topologically Sorted Source Nodes: [out_35, relu_37], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf74, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_37], Original ATen: [aten.convolution]
buf75 = extern_kernels.convolution(buf74, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf75, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf76 = buf75; del buf75 # reuse
# Topologically Sorted Source Nodes: [relu_38], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf76, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_36], Original ATen: [aten.convolution]
buf77 = extern_kernels.convolution(buf76, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf77, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf78 = buf77; del buf77 # reuse
# Topologically Sorted Source Nodes: [out_37, relu_39], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf78, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_39], Original ATen: [aten.convolution]
buf79 = extern_kernels.convolution(buf78, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf79, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf80 = buf79; del buf79 # reuse
# Topologically Sorted Source Nodes: [relu_40], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf80, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_38], Original ATen: [aten.convolution]
buf81 = extern_kernels.convolution(buf80, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf81, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf82 = buf81; del buf81 # reuse
# Topologically Sorted Source Nodes: [out_39, relu_41], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf82, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_41], Original ATen: [aten.convolution]
buf83 = extern_kernels.convolution(buf82, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf83, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf84 = buf83; del buf83 # reuse
# Topologically Sorted Source Nodes: [relu_42], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf84, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_40], Original ATen: [aten.convolution]
buf85 = extern_kernels.convolution(buf84, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf85, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf86 = buf85; del buf85 # reuse
# Topologically Sorted Source Nodes: [out_41, relu_43], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf86, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_43], Original ATen: [aten.convolution]
buf87 = extern_kernels.convolution(buf86, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf87, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf88 = buf87; del buf87 # reuse
# Topologically Sorted Source Nodes: [relu_44], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf88, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_42], Original ATen: [aten.convolution]
buf89 = extern_kernels.convolution(buf88, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf89, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf90 = buf89; del buf89 # reuse
# Topologically Sorted Source Nodes: [out_43, relu_45], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf90, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_45], Original ATen: [aten.convolution]
buf91 = extern_kernels.convolution(buf90, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf91, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf92 = buf91; del buf91 # reuse
# Topologically Sorted Source Nodes: [relu_46], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf92, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_44], Original ATen: [aten.convolution]
buf93 = extern_kernels.convolution(buf92, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf93, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf94 = buf93; del buf93 # reuse
# Topologically Sorted Source Nodes: [out_45, relu_47], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf94, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_47], Original ATen: [aten.convolution]
buf95 = extern_kernels.convolution(buf94, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf95, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf96 = buf95; del buf95 # reuse
# Topologically Sorted Source Nodes: [relu_48], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf96, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_46], Original ATen: [aten.convolution]
buf97 = extern_kernels.convolution(buf96, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf97, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf98 = buf97; del buf97 # reuse
# Topologically Sorted Source Nodes: [out_47, relu_49], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf98, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_49], Original ATen: [aten.convolution]
buf99 = extern_kernels.convolution(buf98, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf99, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf100 = buf99; del buf99 # reuse
# Topologically Sorted Source Nodes: [relu_50], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf100, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_48], Original ATen: [aten.convolution]
buf101 = extern_kernels.convolution(buf100, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf101, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf102 = buf101; del buf101 # reuse
# Topologically Sorted Source Nodes: [out_49, relu_51], Original ATen: [aten.add, aten.relu]
triton_poi_fused_add_relu_2.run(buf102, buf2, 2097152, grid=grid(2097152), stream=stream0)
# Topologically Sorted Source Nodes: [out_50], Original ATen: [aten.convolution]
buf103 = extern_kernels.convolution(buf102, primals_5, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf103, (4, 1, 64, 64), (4096, 4096, 64, 1))
buf104 = buf103; del buf103 # reuse
# Topologically Sorted Source Nodes: [out_51], Original ATen: [aten.add]
triton_poi_fused_add_3.run(buf104, buf0, primals_1, 16384, grid=grid(16384), stream=stream0)
del primals_1
return (buf104, primals_2, primals_3, primals_4, primals_5, buf0, buf2, buf4, buf6, buf8, buf10, buf12, buf14, buf16, buf18, buf20, buf22, buf24, buf26, buf28, buf30, buf32, buf34, buf36, buf38, buf40, buf42, buf44, buf46, buf48, buf50, buf52, buf54, buf56, buf58, buf60, buf62, buf64, buf66, buf68, buf70, buf72, buf74, buf76, buf78, buf80, buf82, buf84, buf86, buf88, buf90, buf92, buf94, buf96, buf98, buf100, buf102, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 1, 64, 64), (4096, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, 1, 3, 3), (9, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from math import sqrt
class DRRN(nn.Module):
def __init__(self):
super(DRRN, self).__init__()
self.input = nn.Conv2d(in_channels=1, out_channels=128, kernel_size
=3, stride=1, padding=1, bias=False)
self.conv1 = nn.Conv2d(in_channels=128, out_channels=128,
kernel_size=3, stride=1, padding=1, bias=False)
self.conv2 = nn.Conv2d(in_channels=128, out_channels=128,
kernel_size=3, stride=1, padding=1, bias=False)
self.output = nn.Conv2d(in_channels=128, out_channels=1,
kernel_size=3, stride=1, padding=1, bias=False)
self.relu = nn.ReLU(inplace=True)
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, sqrt(2.0 / n))
def forward(self, x):
residual = x
inputs = self.input(self.relu(x))
out = inputs
for _ in range(25):
out = self.conv2(self.relu(self.conv1(self.relu(out))))
out = torch.add(out, inputs)
out = self.output(self.relu(out))
out = torch.add(out, residual)
return out
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
from math import sqrt
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + x0, tmp2, None)
@triton.jit
def triton_poi_fused_relu_1(in_out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + x0, tmp2, None)
@triton.jit
def triton_poi_fused_add_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, None)
tmp1 = tl.load(in_ptr0 + x0, None)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x0, tmp4, None)
@triton.jit
def triton_poi_fused_add_3(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, None)
tmp1 = tl.load(in_ptr0 + x0, None)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x0, tmp2, None)
tl.store(out_ptr0 + x0, tmp1, None)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_2, (128, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_3, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_4, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_5, (1, 128, 3, 3), (1152, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 64, 64), (4096, 4096, 64, 1),
torch.float32)
get_raw_stream(0)
triton_poi_fused_relu_0[grid(16384)](primals_1, buf0, 16384, XBLOCK
=256, num_warps=4, num_stages=1)
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf2 = buf1
del buf1
triton_poi_fused_relu_1[grid(2097152)](buf2, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf3 = extern_kernels.convolution(buf2, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf4 = buf3
del buf3
triton_poi_fused_relu_1[grid(2097152)](buf4, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf6 = buf5
del buf5
triton_poi_fused_add_relu_2[grid(2097152)](buf6, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf7 = extern_kernels.convolution(buf6, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf8 = buf7
del buf7
triton_poi_fused_relu_1[grid(2097152)](buf8, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf9 = extern_kernels.convolution(buf8, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf10 = buf9
del buf9
triton_poi_fused_add_relu_2[grid(2097152)](buf10, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf11 = extern_kernels.convolution(buf10, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf12 = buf11
del buf11
triton_poi_fused_relu_1[grid(2097152)](buf12, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf13 = extern_kernels.convolution(buf12, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf13, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf14 = buf13
del buf13
triton_poi_fused_add_relu_2[grid(2097152)](buf14, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf15 = extern_kernels.convolution(buf14, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf15, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf16 = buf15
del buf15
triton_poi_fused_relu_1[grid(2097152)](buf16, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf17 = extern_kernels.convolution(buf16, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf18 = buf17
del buf17
triton_poi_fused_add_relu_2[grid(2097152)](buf18, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf19 = extern_kernels.convolution(buf18, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf20 = buf19
del buf19
triton_poi_fused_relu_1[grid(2097152)](buf20, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf21 = extern_kernels.convolution(buf20, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf21, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf22 = buf21
del buf21
triton_poi_fused_add_relu_2[grid(2097152)](buf22, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf23 = extern_kernels.convolution(buf22, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf24 = buf23
del buf23
triton_poi_fused_relu_1[grid(2097152)](buf24, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf25 = extern_kernels.convolution(buf24, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf25, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf26 = buf25
del buf25
triton_poi_fused_add_relu_2[grid(2097152)](buf26, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf27 = extern_kernels.convolution(buf26, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf27, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf28 = buf27
del buf27
triton_poi_fused_relu_1[grid(2097152)](buf28, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf29 = extern_kernels.convolution(buf28, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf29, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf30 = buf29
del buf29
triton_poi_fused_add_relu_2[grid(2097152)](buf30, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf31 = extern_kernels.convolution(buf30, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf31, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf32 = buf31
del buf31
triton_poi_fused_relu_1[grid(2097152)](buf32, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf33 = extern_kernels.convolution(buf32, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf33, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf34 = buf33
del buf33
triton_poi_fused_add_relu_2[grid(2097152)](buf34, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf35 = extern_kernels.convolution(buf34, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf35, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf36 = buf35
del buf35
triton_poi_fused_relu_1[grid(2097152)](buf36, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf37 = extern_kernels.convolution(buf36, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf37, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf38 = buf37
del buf37
triton_poi_fused_add_relu_2[grid(2097152)](buf38, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf39 = extern_kernels.convolution(buf38, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf39, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf40 = buf39
del buf39
triton_poi_fused_relu_1[grid(2097152)](buf40, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf41 = extern_kernels.convolution(buf40, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf41, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf42 = buf41
del buf41
triton_poi_fused_add_relu_2[grid(2097152)](buf42, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf43 = extern_kernels.convolution(buf42, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf43, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf44 = buf43
del buf43
triton_poi_fused_relu_1[grid(2097152)](buf44, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf45 = extern_kernels.convolution(buf44, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf45, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf46 = buf45
del buf45
triton_poi_fused_add_relu_2[grid(2097152)](buf46, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf47 = extern_kernels.convolution(buf46, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf47, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf48 = buf47
del buf47
triton_poi_fused_relu_1[grid(2097152)](buf48, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf49 = extern_kernels.convolution(buf48, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf49, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf50 = buf49
del buf49
triton_poi_fused_add_relu_2[grid(2097152)](buf50, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf51 = extern_kernels.convolution(buf50, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf51, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf52 = buf51
del buf51
triton_poi_fused_relu_1[grid(2097152)](buf52, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf53 = extern_kernels.convolution(buf52, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf53, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf54 = buf53
del buf53
triton_poi_fused_add_relu_2[grid(2097152)](buf54, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf55 = extern_kernels.convolution(buf54, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf55, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf56 = buf55
del buf55
triton_poi_fused_relu_1[grid(2097152)](buf56, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf57 = extern_kernels.convolution(buf56, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf57, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf58 = buf57
del buf57
triton_poi_fused_add_relu_2[grid(2097152)](buf58, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf59 = extern_kernels.convolution(buf58, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf59, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf60 = buf59
del buf59
triton_poi_fused_relu_1[grid(2097152)](buf60, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf61 = extern_kernels.convolution(buf60, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf61, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf62 = buf61
del buf61
triton_poi_fused_add_relu_2[grid(2097152)](buf62, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf63 = extern_kernels.convolution(buf62, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf63, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf64 = buf63
del buf63
triton_poi_fused_relu_1[grid(2097152)](buf64, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf65 = extern_kernels.convolution(buf64, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf65, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf66 = buf65
del buf65
triton_poi_fused_add_relu_2[grid(2097152)](buf66, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf67 = extern_kernels.convolution(buf66, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf67, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf68 = buf67
del buf67
triton_poi_fused_relu_1[grid(2097152)](buf68, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf69 = extern_kernels.convolution(buf68, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf69, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf70 = buf69
del buf69
triton_poi_fused_add_relu_2[grid(2097152)](buf70, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf71 = extern_kernels.convolution(buf70, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf71, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf72 = buf71
del buf71
triton_poi_fused_relu_1[grid(2097152)](buf72, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf73 = extern_kernels.convolution(buf72, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf73, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf74 = buf73
del buf73
triton_poi_fused_add_relu_2[grid(2097152)](buf74, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf75 = extern_kernels.convolution(buf74, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf75, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf76 = buf75
del buf75
triton_poi_fused_relu_1[grid(2097152)](buf76, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf77 = extern_kernels.convolution(buf76, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf77, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf78 = buf77
del buf77
triton_poi_fused_add_relu_2[grid(2097152)](buf78, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf79 = extern_kernels.convolution(buf78, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf79, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf80 = buf79
del buf79
triton_poi_fused_relu_1[grid(2097152)](buf80, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf81 = extern_kernels.convolution(buf80, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf81, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf82 = buf81
del buf81
triton_poi_fused_add_relu_2[grid(2097152)](buf82, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf83 = extern_kernels.convolution(buf82, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf83, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf84 = buf83
del buf83
triton_poi_fused_relu_1[grid(2097152)](buf84, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf85 = extern_kernels.convolution(buf84, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf85, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf86 = buf85
del buf85
triton_poi_fused_add_relu_2[grid(2097152)](buf86, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf87 = extern_kernels.convolution(buf86, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf87, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf88 = buf87
del buf87
triton_poi_fused_relu_1[grid(2097152)](buf88, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf89 = extern_kernels.convolution(buf88, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf89, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf90 = buf89
del buf89
triton_poi_fused_add_relu_2[grid(2097152)](buf90, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf91 = extern_kernels.convolution(buf90, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf91, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf92 = buf91
del buf91
triton_poi_fused_relu_1[grid(2097152)](buf92, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf93 = extern_kernels.convolution(buf92, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf93, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf94 = buf93
del buf93
triton_poi_fused_add_relu_2[grid(2097152)](buf94, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf95 = extern_kernels.convolution(buf94, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf95, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf96 = buf95
del buf95
triton_poi_fused_relu_1[grid(2097152)](buf96, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf97 = extern_kernels.convolution(buf96, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf97, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf98 = buf97
del buf97
triton_poi_fused_add_relu_2[grid(2097152)](buf98, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf99 = extern_kernels.convolution(buf98, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf99, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf100 = buf99
del buf99
triton_poi_fused_relu_1[grid(2097152)](buf100, 2097152, XBLOCK=512,
num_warps=8, num_stages=1)
buf101 = extern_kernels.convolution(buf100, primals_4, stride=(1, 1
), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf101, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf102 = buf101
del buf101
triton_poi_fused_add_relu_2[grid(2097152)](buf102, buf2, 2097152,
XBLOCK=512, num_warps=8, num_stages=1)
buf103 = extern_kernels.convolution(buf102, primals_5, stride=(1, 1
), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf103, (4, 1, 64, 64), (4096, 4096, 64, 1))
buf104 = buf103
del buf103
triton_poi_fused_add_3[grid(16384)](buf104, buf0, primals_1, 16384,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_1
return (buf104, primals_2, primals_3, primals_4, primals_5, buf0, buf2,
buf4, buf6, buf8, buf10, buf12, buf14, buf16, buf18, buf20, buf22,
buf24, buf26, buf28, buf30, buf32, buf34, buf36, buf38, buf40,
buf42, buf44, buf46, buf48, buf50, buf52, buf54, buf56, buf58,
buf60, buf62, buf64, buf66, buf68, buf70, buf72, buf74, buf76,
buf78, buf80, buf82, buf84, buf86, buf88, buf90, buf92, buf94,
buf96, buf98, buf100, buf102)
class DRRNNew(nn.Module):
def __init__(self):
super(DRRNNew, self).__init__()
self.input = nn.Conv2d(in_channels=1, out_channels=128, kernel_size
=3, stride=1, padding=1, bias=False)
self.conv1 = nn.Conv2d(in_channels=128, out_channels=128,
kernel_size=3, stride=1, padding=1, bias=False)
self.conv2 = nn.Conv2d(in_channels=128, out_channels=128,
kernel_size=3, stride=1, padding=1, bias=False)
self.output = nn.Conv2d(in_channels=128, out_channels=1,
kernel_size=3, stride=1, padding=1, bias=False)
self.relu = nn.ReLU(inplace=True)
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, sqrt(2.0 / n))
def forward(self, input_0):
primals_2 = self.input.weight
primals_3 = self.conv1.weight
primals_4 = self.conv2.weight
primals_5 = self.output.weight
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
loyo1990/DRRN-pytorch
|
DRRN
| false | 10,425 |
[
"MIT"
] | 0 |
63d7dfd4c6bcb4f7b668fc2f5b4e2031cbba6619
|
https://github.com/loyo1990/DRRN-pytorch/tree/63d7dfd4c6bcb4f7b668fc2f5b4e2031cbba6619
|
UpSampleX2
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/je/cjeyximapx6ybmapc54z7zc3bjcfcpkpctbbfywktcuco4gohaj7.py
# Topologically Sorted Source Nodes: [out, prelu], Original ATen: [aten.convolution, aten._prelu_kernel]
# Source node to ATen node mapping:
# out => convolution
# prelu => gt, mul, where
# Graph fragment:
# %convolution : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [2, 2], [1, 1], [1, 1], True, [0, 0], 1), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %convolution), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %convolution, %mul), kwargs = {})
triton_poi_fused__prelu_kernel_convolution_0 = async_compile.triton('triton_poi_fused__prelu_kernel_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__prelu_kernel_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__prelu_kernel_convolution_0(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 784
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 49) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (0))
tmp6 = tl.broadcast_to(tmp5, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp7 = tmp6 * tmp2
tmp8 = tl.where(tmp4, tmp2, tmp7)
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2, 2), padding=(1, 1), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 7, 7), (196, 49, 7, 1))
buf1 = buf0; del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 7, 7), (196, 49, 7, 1), torch.float32)
# Topologically Sorted Source Nodes: [out, prelu], Original ATen: [aten.convolution, aten._prelu_kernel]
stream0 = get_raw_stream(0)
triton_poi_fused__prelu_kernel_convolution_0.run(buf1, primals_2, primals_4, buf2, 784, grid=grid(784), stream=stream0)
del primals_2
return (buf2, primals_1, primals_3, primals_4, buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torchvision.transforms import *
class DeconvBlock(torch.nn.Module):
def __init__(self, input_size, output_size, kernel_size=4, stride=2,
padding=1, bias=True, activation='prelu', norm=None):
super(DeconvBlock, self).__init__()
self.deconv = torch.nn.ConvTranspose2d(input_size, output_size,
kernel_size, stride, padding, bias=bias)
self.norm = norm
if self.norm == 'batch':
self.bn = torch.nn.BatchNorm2d(output_size)
elif self.norm == 'instance':
self.bn = torch.nn.InstanceNorm2d(output_size)
self.activation = activation
if self.activation == 'relu':
self.act = torch.nn.ReLU(True)
elif self.activation == 'prelu':
self.act = torch.nn.PReLU()
elif self.activation == 'lrelu':
self.act = torch.nn.LeakyReLU(0.2, True)
elif self.activation == 'tanh':
self.act = torch.nn.Tanh()
elif self.activation == 'sigmoid':
self.act = torch.nn.Sigmoid()
def forward(self, x):
if self.norm is not None:
out = self.bn(self.deconv(x))
else:
out = self.deconv(x)
if self.activation is not None:
return self.act(out)
else:
return out
class UpSampleX2(torch.nn.Module):
def __init__(self, num_filter, kernel_size=3, stride=2, padding=1, bias
=True, activation='prelu', norm=None):
super(UpSampleX2, self).__init__()
self.down_conv = DeconvBlock(num_filter, num_filter, kernel_size,
stride, padding, activation, norm=None)
def forward(self, x):
return self.down_conv(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_filter': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torchvision.transforms import *
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__prelu_kernel_convolution_0(in_out_ptr0, in_ptr0,
in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 784
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 49 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + 0)
tmp6 = tl.broadcast_to(tmp5, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp7 = tmp6 * tmp2
tmp8 = tl.where(tmp4, tmp2, tmp7)
tl.store(in_out_ptr0 + x3, tmp2, xmask)
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2,
2), padding=(1, 1), dilation=(1, 1), transposed=True,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 7, 7), (196, 49, 7, 1))
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 4, 7, 7), (196, 49, 7, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__prelu_kernel_convolution_0[grid(784)](buf1,
primals_2, primals_4, buf2, 784, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_2
return buf2, primals_1, primals_3, primals_4, buf1
class DeconvBlock(torch.nn.Module):
def __init__(self, input_size, output_size, kernel_size=4, stride=2,
padding=1, bias=True, activation='prelu', norm=None):
super(DeconvBlock, self).__init__()
self.deconv = torch.nn.ConvTranspose2d(input_size, output_size,
kernel_size, stride, padding, bias=bias)
self.norm = norm
if self.norm == 'batch':
self.bn = torch.nn.BatchNorm2d(output_size)
elif self.norm == 'instance':
self.bn = torch.nn.InstanceNorm2d(output_size)
self.activation = activation
if self.activation == 'relu':
self.act = torch.nn.ReLU(True)
elif self.activation == 'prelu':
self.act = torch.nn.PReLU()
elif self.activation == 'lrelu':
self.act = torch.nn.LeakyReLU(0.2, True)
elif self.activation == 'tanh':
self.act = torch.nn.Tanh()
elif self.activation == 'sigmoid':
self.act = torch.nn.Sigmoid()
def forward(self, x):
if self.norm is not None:
out = self.bn(self.deconv(x))
else:
out = self.deconv(x)
if self.activation is not None:
return self.act(out)
else:
return out
class UpSampleX2New(torch.nn.Module):
def __init__(self, num_filter, kernel_size=3, stride=2, padding=1, bias
=True, activation='prelu', norm=None):
super(UpSampleX2New, self).__init__()
self.down_conv = DeconvBlock(num_filter, num_filter, kernel_size,
stride, padding, activation, norm=None)
def forward(self, input_0):
primals_1 = self.down_conv.deconv.weight
primals_2 = self.down_conv.deconv.bias
primals_4 = self.down_conv.act.weight
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
|
lizatish/My_CNN
|
UpSampleX2
| false | 10,426 |
[
"MIT"
] | 0 |
b13818bcce2f8a3697d20e34157e3dce53f953ee
|
https://github.com/lizatish/My_CNN/tree/b13818bcce2f8a3697d20e34157e3dce53f953ee
|
InteractingLayer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ak/caklf2pyfxcrlk24jikpqfulearftlggift3k32sdo56rqpygtef.py
# Topologically Sorted Source Nodes: [querys_1], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# querys_1 => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem, %getitem_1],), kwargs = {})
triton_poi_fused_stack_0 = async_compile.triton('triton_poi_fused_stack_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_stack_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 8)
x0 = xindex % 2
x1 = (xindex // 2) % 4
x3 = xindex
tmp0 = x2
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (4*x1) + (16*x2)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr0 + (2 + x0 + (4*x1) + (16*((-4) + x2))), tmp6 & xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + (x3), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/d7/cd7xa5d4yg5y7exr6s4sr25rd6okj4v7452l7cyhxnqr3mcd4qhj.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_15, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_15, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/kx/ckxzcazhsdasvh5sdcvshdrxriufwxfrn25tt7nuks5deb2u6ei5.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ee/ceebt7dp2dnq4qecrxhaoiakzivwsb6pug54s5t7st6c2qpbsii7.py
# Topologically Sorted Source Nodes: [result_4], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# result_4 => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%squeeze_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_3 = async_compile.triton('triton_poi_fused_relu_threshold_backward_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_3', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_3(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp11 = tl.load(in_out_ptr0 + (x2), xmask)
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((2*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 4, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr0 + (32 + (2*x1) + ((-2) + x0)), tmp6 & xmask, eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp12 = tmp10 + tmp11
tmp13 = tl.full([1], 0, tl.int32)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp15 = 0.0
tmp16 = tmp14 <= tmp15
tl.store(in_out_ptr0 + (x2), tmp14, xmask)
tl.store(out_ptr0 + (x2), tmp16, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [querys], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), primals_2, out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [keys], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [values], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((8, 4, 2), (8, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [querys_1], Original ATen: [aten.stack]
stream0 = get_raw_stream(0)
triton_poi_fused_stack_0.run(buf0, buf3, 64, grid=grid(64), stream=stream0)
buf4 = reinterpret_tensor(buf0, (8, 4, 2), (8, 2, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [keys_1], Original ATen: [aten.stack]
triton_poi_fused_stack_0.run(buf1, buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((8, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [inner_product], Original ATen: [aten.bmm]
extern_kernels.bmm(buf3, reinterpret_tensor(buf4, (8, 2, 4), (8, 1, 2), 0), out=buf5)
buf6 = empty_strided_cuda((2, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf5, buf6, 128, grid=grid(128), stream=stream0)
buf7 = reinterpret_tensor(buf5, (2, 4, 4, 4), (64, 16, 4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 128, grid=grid(128), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf1, (8, 4, 2), (8, 2, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [values_1], Original ATen: [aten.stack]
triton_poi_fused_stack_0.run(buf2, buf8, 64, grid=grid(64), stream=stream0)
buf9 = reinterpret_tensor(buf2, (8, 4, 2), (8, 2, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf7, (8, 4, 4), (16, 4, 1), 0), buf8, out=buf9)
buf10 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tensordot_3], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), primals_5, out=buf10)
del primals_5
buf11 = reinterpret_tensor(buf10, (4, 4, 4), (16, 4, 1), 0); del buf10 # reuse
buf12 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [result_4], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_3.run(buf11, buf9, buf12, 64, grid=grid(64), stream=stream0)
del buf9
return (buf11, buf7, buf7, buf12, reinterpret_tensor(primals_1, (4, 16), (1, 4), 0), reinterpret_tensor(buf8, (8, 2, 4), (8, 1, 2), 0), reinterpret_tensor(buf3, (8, 2, 4), (8, 1, 2), 0), buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from sklearn.metrics import *
import torch.onnx
import torch as torch
class InteractingLayer(nn.Module):
"""A Layer used in AutoInt that model the correlations between different feature fields by multi-head self-attention mechanism.
Input shape
- A 3D tensor with shape: ``(batch_size,field_size,embedding_size)``.
Output shape
- 3D tensor with shape:``(batch_size,field_size,embedding_size)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **head_num**: int.The head number in multi-head self-attention network.
- **use_res**: bool.Whether or not use standard residual connections before output.
- **seed**: A Python integer to use as random seed.
References
- [Song W, Shi C, Xiao Z, et al. AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks[J]. arXiv preprint arXiv:1810.11921, 2018.](https://arxiv.org/abs/1810.11921)
"""
def __init__(self, embedding_size, head_num=2, use_res=True, scaling=
False, seed=1024, device='cpu'):
super(InteractingLayer, self).__init__()
if head_num <= 0:
raise ValueError('head_num must be a int > 0')
if embedding_size % head_num != 0:
raise ValueError(
'embedding_size is not an integer multiple of head_num!')
self.att_embedding_size = embedding_size // head_num
self.head_num = head_num
self.use_res = use_res
self.scaling = scaling
self.seed = seed
self.W_Query = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
self.W_key = nn.Parameter(torch.Tensor(embedding_size, embedding_size))
self.W_Value = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
if self.use_res:
self.W_Res = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
for tensor in self.parameters():
nn.init.normal_(tensor, mean=0.0, std=0.05)
self
def forward(self, inputs):
if len(inputs.shape) != 3:
raise ValueError(
'Unexpected inputs dimensions %d, expect to be 3 dimensions' %
len(inputs.shape))
querys = torch.tensordot(inputs, self.W_Query, dims=([-1], [0]))
keys = torch.tensordot(inputs, self.W_key, dims=([-1], [0]))
values = torch.tensordot(inputs, self.W_Value, dims=([-1], [0]))
querys = torch.stack(torch.split(querys, self.att_embedding_size,
dim=2))
keys = torch.stack(torch.split(keys, self.att_embedding_size, dim=2))
values = torch.stack(torch.split(values, self.att_embedding_size,
dim=2))
inner_product = torch.einsum('bnik,bnjk->bnij', querys, keys)
if self.scaling:
inner_product /= self.att_embedding_size ** 0.5
self.normalized_att_scores = F.softmax(inner_product, dim=-1)
result = torch.matmul(self.normalized_att_scores, values)
result = torch.cat(torch.split(result, 1), dim=-1)
result = torch.squeeze(result, dim=0)
if self.use_res:
result += torch.tensordot(inputs, self.W_Res, dims=([-1], [0]))
result = F.relu(result)
return result
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'embedding_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
from sklearn.metrics import *
import torch.onnx
import torch as torch
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_stack_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 8
x0 = xindex % 2
x1 = xindex // 2 % 4
x3 = xindex
tmp0 = x2
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 4 * x1 + 16 * x2), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp9 = tl.load(in_ptr0 + (2 + x0 + 4 * x1 + 16 * (-4 + x2)), tmp6 &
xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + x3, tmp10, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_3(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp11 = tl.load(in_out_ptr0 + x2, xmask)
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (2 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 4, tl.int64)
tmp9 = tl.load(in_ptr0 + (32 + 2 * x1 + (-2 + x0)), tmp6 & xmask,
eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp12 = tmp10 + tmp11
tmp13 = tl.full([1], 0, tl.int32)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp15 = 0.0
tmp16 = tmp14 <= tmp15
tl.store(in_out_ptr0 + x2, tmp14, xmask)
tl.store(out_ptr0 + x2, tmp16, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
primals_2, out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((8, 4, 2), (8, 2, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_stack_0[grid(64)](buf0, buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf4 = reinterpret_tensor(buf0, (8, 4, 2), (8, 2, 1), 0)
del buf0
triton_poi_fused_stack_0[grid(64)](buf1, buf4, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((8, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(buf3, reinterpret_tensor(buf4, (8, 2, 4), (8, 1,
2), 0), out=buf5)
buf6 = empty_strided_cuda((2, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(128)](buf5, buf6, 128, XBLOCK=128,
num_warps=4, num_stages=1)
buf7 = reinterpret_tensor(buf5, (2, 4, 4, 4), (64, 16, 4, 1), 0)
del buf5
triton_poi_fused__softmax_2[grid(128)](buf6, buf7, 128, XBLOCK=128,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf1, (8, 4, 2), (8, 2, 1), 0)
del buf1
triton_poi_fused_stack_0[grid(64)](buf2, buf8, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf9 = reinterpret_tensor(buf2, (8, 4, 2), (8, 2, 1), 0)
del buf2
extern_kernels.bmm(reinterpret_tensor(buf7, (8, 4, 4), (16, 4, 1),
0), buf8, out=buf9)
buf10 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
primals_5, out=buf10)
del primals_5
buf11 = reinterpret_tensor(buf10, (4, 4, 4), (16, 4, 1), 0)
del buf10
buf12 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_3[grid(64)](buf11, buf9,
buf12, 64, XBLOCK=64, num_warps=1, num_stages=1)
del buf9
return buf11, buf7, buf7, buf12, reinterpret_tensor(primals_1, (4, 16),
(1, 4), 0), reinterpret_tensor(buf8, (8, 2, 4), (8, 1, 2), 0
), reinterpret_tensor(buf3, (8, 2, 4), (8, 1, 2), 0), buf4
class InteractingLayerNew(nn.Module):
"""A Layer used in AutoInt that model the correlations between different feature fields by multi-head self-attention mechanism.
Input shape
- A 3D tensor with shape: ``(batch_size,field_size,embedding_size)``.
Output shape
- 3D tensor with shape:``(batch_size,field_size,embedding_size)``.
Arguments
- **in_features** : Positive integer, dimensionality of input features.
- **head_num**: int.The head number in multi-head self-attention network.
- **use_res**: bool.Whether or not use standard residual connections before output.
- **seed**: A Python integer to use as random seed.
References
- [Song W, Shi C, Xiao Z, et al. AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks[J]. arXiv preprint arXiv:1810.11921, 2018.](https://arxiv.org/abs/1810.11921)
"""
def __init__(self, embedding_size, head_num=2, use_res=True, scaling=
False, seed=1024, device='cpu'):
super(InteractingLayerNew, self).__init__()
if head_num <= 0:
raise ValueError('head_num must be a int > 0')
if embedding_size % head_num != 0:
raise ValueError(
'embedding_size is not an integer multiple of head_num!')
self.att_embedding_size = embedding_size // head_num
self.head_num = head_num
self.use_res = use_res
self.scaling = scaling
self.seed = seed
self.W_Query = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
self.W_key = nn.Parameter(torch.Tensor(embedding_size, embedding_size))
self.W_Value = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
if self.use_res:
self.W_Res = nn.Parameter(torch.Tensor(embedding_size,
embedding_size))
for tensor in self.parameters():
nn.init.normal_(tensor, mean=0.0, std=0.05)
self
def forward(self, input_0):
primals_2 = self.W_Query
primals_3 = self.W_key
primals_4 = self.W_Value
primals_5 = self.W_Res
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
dulvqingyunLT/DeepCTR-Torch
|
InteractingLayer
| false | 10,427 |
[
"Apache-2.0"
] | 0 |
f40cf08f3469aa471f9ca69e44c5de51180341cc
|
https://github.com/dulvqingyunLT/DeepCTR-Torch/tree/f40cf08f3469aa471f9ca69e44c5de51180341cc
|
CriticMlp
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/zb/czbrdc6746xv7kfxrqkzgbhm74ijdfuyfd3sz3llzzwzm6wzxmfi.py
# Topologically Sorted Source Nodes: [global_obs_1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# global_obs_1 => relu
# Graph fragment:
# %add_tensor_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_4, %primals_3), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_4,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xc/cxcbskmvfetvwy5kto3hmjtnvu347mijo2mhvgsdlafetsnubgud.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# x => cat
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%view_2, %relu_2], 1), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*(x1 // 4)) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr1 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tmp13 = tl.full([1], 8, tl.int64)
tmp14 = tmp0 < tmp13
tmp15 = tl.load(in_ptr2 + ((4*x1) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp16 = tl.load(in_ptr3 + ((-4) + x0), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp17 = tmp15 + tmp16
tmp18 = triton_helpers.maximum(tmp8, tmp17)
tmp19 = tl.full(tmp18.shape, 0.0, tmp18.dtype)
tmp20 = tl.where(tmp12, tmp18, tmp19)
tmp21 = tl.where(tmp4, tmp11, tmp20)
tl.store(out_ptr0 + (x2), tmp21, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/se/csemwzmaiwyhjjvzkbk56prr5otscb72tpb7v2t622v5ck5zy3s4.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_1 => relu_3
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_9), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_2 = async_compile.triton('triton_poi_fused_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6y/c6yl4uv4syttx5p2rvialnpvw2b2afygk6u4wtiv3hxyrze6hevt.py
# Topologically Sorted Source Nodes: [local_obs_1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# local_obs_1 => relu_2
# Graph fragment:
# %add_tensor_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_2, %primals_7), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_2,), kwargs = {})
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_2, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_3 = async_compile.triton('triton_poi_fused_relu_threshold_backward_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ac/cac6aispvtf6fyinzk7ei7ozijq6adhgepxz7l4mnfner74j5yma.py
# Topologically Sorted Source Nodes: [global_obs_2], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# global_obs_2 => relu_1
# Graph fragment:
# %add_tensor_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_3, %primals_5), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_3,), kwargs = {})
# %le_3 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_4 = async_compile.triton('triton_poi_fused_relu_threshold_backward_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_4(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 16), (16, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 8), (8, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
assert_size_stride(primals_12, (4, 4), (4, 1))
assert_size_stride(primals_13, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 16), (16, 1), 0), reinterpret_tensor(primals_2, (16, 4), (1, 16), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [global_obs_1], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf3)
del primals_6
buf4 = empty_strided_cuda((64, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf2, primals_5, buf3, primals_7, buf4, 512, grid=grid(512), stream=stream0)
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf4, reinterpret_tensor(primals_8, (8, 4), (1, 8), 0), out=buf5)
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf6, primals_9, 256, grid=grid(256), stream=stream0)
del primals_9
buf7 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf6, reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), out=buf7)
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf8, primals_11, 256, grid=grid(256), stream=stream0)
del primals_11
buf9 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [q], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_13, buf8, reinterpret_tensor(primals_12, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf9)
del primals_13
buf10 = empty_strided_cuda((64, 4), (4, 1), torch.bool)
# Topologically Sorted Source Nodes: [local_obs_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_3.run(buf3, primals_7, buf10, 256, grid=grid(256), stream=stream0)
del buf3
del primals_7
buf11 = empty_strided_cuda((16, 4), (4, 1), torch.bool)
# Topologically Sorted Source Nodes: [global_obs_2], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_4.run(buf2, primals_5, buf11, 64, grid=grid(64), stream=stream0)
del buf2
del primals_5
return (reinterpret_tensor(buf9, (16, 4, 4), (16, 4, 1), 0), primals_1, buf1, buf4, buf6, buf8, primals_12, primals_10, primals_8, buf10, buf11, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 16), (16, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
def init_weights(layer, gain):
for p in layer.parameters():
if len(p.data.shape) >= 2:
nn.init.orthogonal_(p, gain=gain)
else:
p.data.zero_()
def all_init_weights(m, gain=2 ** 0.5):
init_weights(m, gain)
class CriticMlp(nn.Module):
def __init__(self, obs_size, n_agent, n_action, global_encode_size,
local_encode_size, fc1_size, fc2_size):
super(CriticMlp, self).__init__()
self.obs_size = obs_size
self.n_agent = n_agent
self.n_action = n_action
self.global_encode_fc1 = nn.Linear(obs_size * n_agent,
global_encode_size)
self.global_encode_fc2 = nn.Linear(global_encode_size,
global_encode_size)
self.local_encode_fc = nn.Linear(obs_size, local_encode_size)
self.fc1 = nn.Linear(global_encode_size + local_encode_size, fc1_size)
self.fc2 = nn.Linear(fc1_size, fc2_size)
self.fc3 = nn.Linear(fc2_size, n_action)
self.apply(all_init_weights)
init_weights(self.fc3, gain=1)
def forward(self, obs_j):
global_obs = obs_j.view(-1, self.obs_size * self.n_agent)
global_obs = F.relu(self.global_encode_fc1(global_obs))
global_obs = F.relu(self.global_encode_fc2(global_obs))
local_obs = obs_j.view(-1, self.obs_size)
local_obs = F.relu(self.local_encode_fc(local_obs))
global_obs = global_obs.repeat_interleave(self.n_agent, dim=0)
x = torch.cat((global_obs, local_obs), dim=1)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
q = self.fc3(x)
q = q.view(-1, self.n_agent, self.n_action)
return q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'obs_size': 4, 'n_agent': 4, 'n_action': 4,
'global_encode_size': 4, 'local_encode_size': 4, 'fc1_size': 4,
'fc2_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * (x1 // 4) + x0), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr1 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp15 = tl.load(in_ptr2 + (4 * x1 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp16 = tl.load(in_ptr3 + (-4 + x0), tmp12 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp17 = tmp15 + tmp16
tmp18 = triton_helpers.maximum(tmp8, tmp17)
tmp19 = tl.full(tmp18.shape, 0.0, tmp18.dtype)
tmp20 = tl.where(tmp12, tmp18, tmp19)
tmp21 = tl.where(tmp4, tmp11, tmp20)
tl.store(out_ptr0 + x2, tmp21, xmask)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_3(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_4(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + x2, tmp6, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 16), (16, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 8), (8, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
assert_size_stride(primals_12, (4, 4), (4, 1))
assert_size_stride(primals_13, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 16), (16, 1),
0), reinterpret_tensor(primals_2, (16, 4), (1, 16), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(64)](buf1, primals_3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (4, 4), (1, 4
), 0), out=buf2)
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf3)
del primals_6
buf4 = empty_strided_cuda((64, 8), (8, 1), torch.float32)
triton_poi_fused_cat_1[grid(512)](buf2, primals_5, buf3, primals_7,
buf4, 512, XBLOCK=128, num_warps=4, num_stages=1)
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(buf4, reinterpret_tensor(primals_8, (8, 4), (1, 8
), 0), out=buf5)
buf6 = buf5
del buf5
triton_poi_fused_relu_2[grid(256)](buf6, primals_9, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_9
buf7 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(buf6, reinterpret_tensor(primals_10, (4, 4), (1,
4), 0), out=buf7)
buf8 = buf7
del buf7
triton_poi_fused_relu_2[grid(256)](buf8, primals_11, 256, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_11
buf9 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_13, buf8, reinterpret_tensor(
primals_12, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf9)
del primals_13
buf10 = empty_strided_cuda((64, 4), (4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_3[grid(256)](buf3,
primals_7, buf10, 256, XBLOCK=256, num_warps=4, num_stages=1)
del buf3
del primals_7
buf11 = empty_strided_cuda((16, 4), (4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_4[grid(64)](buf2,
primals_5, buf11, 64, XBLOCK=64, num_warps=1, num_stages=1)
del buf2
del primals_5
return (reinterpret_tensor(buf9, (16, 4, 4), (16, 4, 1), 0), primals_1,
buf1, buf4, buf6, buf8, primals_12, primals_10, primals_8, buf10,
buf11, primals_4)
def init_weights(layer, gain):
for p in layer.parameters():
if len(p.data.shape) >= 2:
nn.init.orthogonal_(p, gain=gain)
else:
p.data.zero_()
def all_init_weights(m, gain=2 ** 0.5):
init_weights(m, gain)
class CriticMlpNew(nn.Module):
def __init__(self, obs_size, n_agent, n_action, global_encode_size,
local_encode_size, fc1_size, fc2_size):
super(CriticMlpNew, self).__init__()
self.obs_size = obs_size
self.n_agent = n_agent
self.n_action = n_action
self.global_encode_fc1 = nn.Linear(obs_size * n_agent,
global_encode_size)
self.global_encode_fc2 = nn.Linear(global_encode_size,
global_encode_size)
self.local_encode_fc = nn.Linear(obs_size, local_encode_size)
self.fc1 = nn.Linear(global_encode_size + local_encode_size, fc1_size)
self.fc2 = nn.Linear(fc1_size, fc2_size)
self.fc3 = nn.Linear(fc2_size, n_action)
self.apply(all_init_weights)
init_weights(self.fc3, gain=1)
def forward(self, input_0):
primals_2 = self.global_encode_fc1.weight
primals_3 = self.global_encode_fc1.bias
primals_4 = self.global_encode_fc2.weight
primals_5 = self.global_encode_fc2.bias
primals_6 = self.local_encode_fc.weight
primals_7 = self.local_encode_fc.bias
primals_8 = self.fc1.weight
primals_9 = self.fc1.bias
primals_10 = self.fc2.weight
primals_11 = self.fc2.bias
primals_12 = self.fc3.weight
primals_13 = self.fc3.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13])
return output[0]
|
heavenlysf/thesis
|
CriticMlp
| false | 10,428 |
[
"MIT"
] | 0 |
646553c45860f337c91a48ab7f666a174784472f
|
https://github.com/heavenlysf/thesis/tree/646553c45860f337c91a48ab7f666a174784472f
|
LayerNorm
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/nu/cnuyiboekdklervsd4zozxy6hj5ypmcigwk6x427x6ylotbbcr5k.py
# Topologically Sorted Source Nodes: [mean, std, sub, add, x, mul, x_1], Original ATen: [aten.mean, aten.std, aten.sub, aten.add, aten.div, aten.mul]
# Source node to ATen node mapping:
# add => add
# mean => mean
# mul => mul
# std => var
# sub => sub
# x => div
# x_1 => add_1
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%view, [1]), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%view, [1]), kwargs = {correction: 1.0})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %view_1), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, 1e-08), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %add), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %view_4), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %view_5), kwargs = {})
triton_per_fused_add_div_mean_mul_std_sub_0 = async_compile.triton('triton_per_fused_add_div_mean_mul_std_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_mean_mul_std_sub_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_mean_mul_std_sub_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp28 = tl.load(in_ptr1 + (r3), None, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + (r3), None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp6 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp1 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp4 / tmp19
tmp21 = 63.0
tmp22 = tmp18 / tmp21
tmp23 = libdevice.sqrt(tmp22)
tmp24 = 1e-08
tmp25 = tmp23 + tmp24
tmp26 = tmp0 - tmp20
tmp27 = tmp26 / tmp25
tmp29 = tmp27 * tmp28
tmp31 = tmp29 + tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x0), tmp25, xmask)
tl.store(out_ptr0 + (r1 + (64*x0)), tmp31, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf3 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf1 = buf0; del buf0 # reuse
buf5 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf3 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, std, sub, add, x, mul, x_1], Original ATen: [aten.mean, aten.std, aten.sub, aten.add, aten.div, aten.mul]
stream0 = get_raw_stream(0)
triton_per_fused_add_div_mean_mul_std_sub_0.run(buf1, buf5, primals_1, primals_2, primals_3, buf6, 4, 64, grid=grid(4), stream=stream0)
del primals_2
del primals_3
return (buf6, primals_1, reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0), buf5, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import Parameter
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_features': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_add_div_mean_mul_std_sub_0(in_out_ptr0, in_out_ptr1,
in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = rindex // 16
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp28 = tl.load(in_ptr1 + r3, None, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + r3, None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp6 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp1 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp4 / tmp19
tmp21 = 63.0
tmp22 = tmp18 / tmp21
tmp23 = libdevice.sqrt(tmp22)
tmp24 = 1e-08
tmp25 = tmp23 + tmp24
tmp26 = tmp0 - tmp20
tmp27 = tmp26 / tmp25
tmp29 = tmp27 * tmp28
tmp31 = tmp29 + tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x0, tmp25, xmask)
tl.store(out_ptr0 + (r1 + 64 * x0), tmp31, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4,), (1,), torch.float32)
buf3 = empty_strided_cuda((4,), (1,), torch.float32)
buf1 = buf0
del buf0
buf5 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf3
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_add_div_mean_mul_std_sub_0[grid(4)](buf1, buf5,
primals_1, primals_2, primals_3, buf6, 4, 64, XBLOCK=1,
num_warps=2, num_stages=1)
del primals_2
del primals_3
return buf6, primals_1, reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1,
1), 0), buf5
class LayerNormNew(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNormNew, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, input_0):
primals_2 = self.gamma
primals_3 = self.beta
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
kangzhiq/DeepFillv2_Pytorch
|
LayerNorm
| false | 10,429 |
[
"MIT"
] | 0 |
9c7ed61b25bb995713f89108b712490737abe1b1
|
https://github.com/kangzhiq/DeepFillv2_Pytorch/tree/9c7ed61b25bb995713f89108b712490737abe1b1
|
SAB
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wb/cwbnzhah456ssqf5k4ob4llmezbi7o6givtx6ppibkklzj3kmilo.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + (16*y0)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wl/cwll6lomrruls6u4c3ua52p4g5jvoxwf6bvyj4dolvf23rk3zogz.py
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_3 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/5i/c5ipggcwfmiy6xwyorkqi45ymmmjaizazawvwrlqypsmysj65d6x.py
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# h_1 => add
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_21, %primals_1), kwargs = {})
triton_poi_fused_add_4 = async_compile.triton('triton_poi_fused_add_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_4(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3j/c3jlveurtxkoy4v4xyl2fgccp7ieaeliub6jaslea3rhtsnp7cxd.py
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_2 => add_1
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_23,), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %relu), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_5 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_5(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
del primals_6
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(buf0, primals_3, buf3, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_3
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf1, primals_5, buf4, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [contiguous_2], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf2, primals_7, buf8, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_7
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [o], Original ATen: [aten.bmm]
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf11)
buf12 = reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0); del buf11 # reuse
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
triton_poi_fused_add_4.run(buf12, primals_9, primals_1, 64, grid=grid(64), stream=stream0)
del primals_9
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf12, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), out=buf13)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_5.run(buf12, buf13, primals_11, buf14, buf15, 64, grid=grid(64), stream=stream0)
del buf13
del primals_11
return (buf14, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(buf12, (16, 4), (4, 1), 0), buf15, primals_10, primals_8, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class SAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.mab = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
def forward(self, x, presence=None):
return self.mab(x, x, presence)
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'d': 4, 'n_heads': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + 16 * y0), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_4(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_5(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
del primals_6
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(4, 16)](buf0, primals_3, buf3, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_3
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf0
triton_poi_fused_clone_0[grid(4, 16)](buf1, primals_5, buf4, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf5, buf6, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf7 = buf5
del buf5
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_0[grid(4, 16)](buf2, primals_7, buf8, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_7
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0)
del buf2
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1,
0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_3[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.mm(reinterpret_tensor(buf10, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf11)
buf12 = reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0)
del buf11
triton_poi_fused_add_4[grid(64)](buf12, primals_9, primals_1, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_9
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf12, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), out=buf13)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_5[grid(64)](buf12,
buf13, primals_11, buf14, buf15, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf13
del primals_11
return buf14, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0
), reinterpret_tensor(buf12, (16, 4), (4, 1), 0
), buf15, primals_10, primals_8, reinterpret_tensor(buf8, (16, 1, 4
), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0)
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class SABNew(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.mab = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
def forward(self, input_0):
primals_2 = self.mab.mqkv.q_projector.weight
primals_3 = self.mab.mqkv.q_projector.bias
primals_4 = self.mab.mqkv.k_projector.weight
primals_5 = self.mab.mqkv.k_projector.bias
primals_6 = self.mab.mqkv.v_projector.weight
primals_7 = self.mab.mqkv.v_projector.bias
primals_8 = self.mab.mqkv.o_projector.weight
primals_9 = self.mab.mqkv.o_projector.bias
primals_10 = self.mab.fc.weight
primals_11 = self.mab.fc.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
|
karayanni/torch-scae
|
SAB
| false | 10,430 |
[
"Apache-2.0"
] | 0 |
e044662d8942d8d1923d13d071f375144cf4a1e8
|
https://github.com/karayanni/torch-scae/tree/e044662d8942d8d1923d13d071f375144cf4a1e8
|
MAB
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wb/cwbnzhah456ssqf5k4ob4llmezbi7o6givtx6ppibkklzj3kmilo.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + (16*y0)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wl/cwll6lomrruls6u4c3ua52p4g5jvoxwf6bvyj4dolvf23rk3zogz.py
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_3 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/5i/c5ipggcwfmiy6xwyorkqi45ymmmjaizazawvwrlqypsmysj65d6x.py
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# h_1 => add
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_21, %primals_1), kwargs = {})
triton_poi_fused_add_4 = async_compile.triton('triton_poi_fused_add_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_4(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3j/c3jlveurtxkoy4v4xyl2fgccp7ieaeliub6jaslea3rhtsnp7cxd.py
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_2 => add_1
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_23,), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %relu), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_5 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_5(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf0)
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf1)
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf2)
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(buf0, primals_4, buf3, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_4
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf1, primals_6, buf4, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_6
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [contiguous_2], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf2, primals_8, buf8, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_8
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [o], Original ATen: [aten.bmm]
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf11)
buf12 = reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0); del buf11 # reuse
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
triton_poi_fused_add_4.run(buf12, primals_10, primals_1, 64, grid=grid(64), stream=stream0)
del primals_10
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf12, (16, 4), (4, 1), 0), reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf13)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_5.run(buf12, buf13, primals_12, buf14, buf15, 64, grid=grid(64), stream=stream0)
del buf13
del primals_12
return (buf14, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(buf12, (16, 4), (4, 1), 0), buf15, primals_11, primals_9, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'d': 4, 'n_heads': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + 16 * y0), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_4(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_5(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12
) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf0)
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf1)
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf2)
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(4, 16)](buf0, primals_4, buf3, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_4
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf0
triton_poi_fused_clone_0[grid(4, 16)](buf1, primals_6, buf4, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_6
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf5, buf6, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf7 = buf5
del buf5
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_0[grid(4, 16)](buf2, primals_8, buf8, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_8
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0)
del buf2
extern_kernels.bmm(buf7, reinterpret_tensor(buf8, (16, 4, 1), (4, 1,
0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_3[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.mm(reinterpret_tensor(buf10, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf11)
buf12 = reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0)
del buf11
triton_poi_fused_add_4[grid(64)](buf12, primals_10, primals_1, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_10
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf12, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf13)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_5[grid(64)](buf12,
buf13, primals_12, buf14, buf15, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf13
del primals_12
return buf14, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0
), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0
), reinterpret_tensor(buf12, (16, 4), (4, 1), 0
), buf15, primals_11, primals_9, reinterpret_tensor(buf8, (16, 1, 4
), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0)
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MABNew(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, input_0, input_1):
primals_3 = self.mqkv.q_projector.weight
primals_4 = self.mqkv.q_projector.bias
primals_5 = self.mqkv.k_projector.weight
primals_6 = self.mqkv.k_projector.bias
primals_7 = self.mqkv.v_projector.weight
primals_8 = self.mqkv.v_projector.bias
primals_9 = self.mqkv.o_projector.weight
primals_10 = self.mqkv.o_projector.bias
primals_11 = self.fc.weight
primals_12 = self.fc.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12])
return output[0]
|
karayanni/torch-scae
|
MAB
| false | 10,431 |
[
"Apache-2.0"
] | 0 |
e044662d8942d8d1923d13d071f375144cf4a1e8
|
https://github.com/karayanni/torch-scae/tree/e044662d8942d8d1923d13d071f375144cf4a1e8
|
DiscriminatorHingeLoss
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/5m/c5mz5opmf2emkyscphowbwckdgbzlmt5fg36tj7scmuhse5cwbyg.py
# Topologically Sorted Source Nodes: [zeros_like, sub, minimum, mean, real_loss, zeros_like_1, sub_1, minimum_1, mean_1, fake_loss, add], Original ATen: [aten.zeros_like, aten.sub, aten.minimum, aten.mean, aten.neg, aten.rsub, aten.add]
# Source node to ATen node mapping:
# add => add
# fake_loss => neg_1
# mean => mean
# mean_1 => mean_1
# minimum => minimum
# minimum_1 => minimum_1
# real_loss => neg
# sub => sub
# sub_1 => sub_1
# zeros_like => full_default
# zeros_like_1 => full_default_1
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, 1), kwargs = {})
# %minimum : [num_users=1] = call_function[target=torch.ops.aten.minimum.default](args = (%full_default, %sub), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%minimum,), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%mean,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (-1, %arg1_1), kwargs = {})
# %minimum_1 : [num_users=1] = call_function[target=torch.ops.aten.minimum.default](args = (%full_default_1, %sub_1), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%minimum_1,), kwargs = {})
# %neg_1 : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%mean_1,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%neg, %neg_1), kwargs = {})
triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0 = async_compile.triton('triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp8 = tl.load(in_ptr1 + (r0), None)
tmp1 = 1.0
tmp2 = tmp0 - tmp1
tmp3 = 0.0
tmp4 = triton_helpers.minimum(tmp3, tmp2)
tmp5 = tl.broadcast_to(tmp4, [RBLOCK])
tmp7 = triton_helpers.promote_to_tensor(tl.sum(tmp5, 0))
tmp9 = -1.0
tmp10 = tmp9 - tmp8
tmp11 = triton_helpers.minimum(tmp3, tmp10)
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 256.0
tmp16 = tmp7 / tmp15
tmp17 = -tmp16
tmp18 = tmp14 / tmp15
tmp19 = -tmp18
tmp20 = tmp17 + tmp19
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp20, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [zeros_like, sub, minimum, mean, real_loss, zeros_like_1, sub_1, minimum_1, mean_1, fake_loss, add], Original ATen: [aten.zeros_like, aten.sub, aten.minimum, aten.mean, aten.neg, aten.rsub, aten.add]
stream0 = get_raw_stream(0)
triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0.run(buf2, arg0_1, arg1_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class DiscriminatorHingeLoss(nn.Module):
def __init__(self, reduction='mean'):
super(DiscriminatorHingeLoss, self).__init__()
if reduction not in ['mean', 'sum']:
raise ValueError(
'Valid values for the reduction param are `mean`, `sum`')
self.reduction = reduction
def forward(self, fake_out, real_out):
real_loss = -torch.minimum(torch.zeros_like(real_out), real_out - 1
).mean()
fake_loss = -torch.minimum(torch.zeros_like(fake_out), -1 - fake_out
).mean()
return real_loss + fake_loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0(in_out_ptr0,
in_ptr0, in_ptr1, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp8 = tl.load(in_ptr1 + r0, None)
tmp1 = 1.0
tmp2 = tmp0 - tmp1
tmp3 = 0.0
tmp4 = triton_helpers.minimum(tmp3, tmp2)
tmp5 = tl.broadcast_to(tmp4, [RBLOCK])
tmp7 = triton_helpers.promote_to_tensor(tl.sum(tmp5, 0))
tmp9 = -1.0
tmp10 = tmp9 - tmp8
tmp11 = triton_helpers.minimum(tmp3, tmp10)
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 256.0
tmp16 = tmp7 / tmp15
tmp17 = -tmp16
tmp18 = tmp14 / tmp15
tmp19 = -tmp18
tmp20 = tmp17 + tmp19
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp20, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_add_mean_minimum_neg_rsub_sub_zeros_like_0[grid(1)](
buf2, arg0_1, arg1_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf2,
class DiscriminatorHingeLossNew(nn.Module):
def __init__(self, reduction='mean'):
super(DiscriminatorHingeLossNew, self).__init__()
if reduction not in ['mean', 'sum']:
raise ValueError(
'Valid values for the reduction param are `mean`, `sum`')
self.reduction = reduction
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
kpandey008/SAGAN
|
DiscriminatorHingeLoss
| false | 10,432 |
[
"MIT"
] | 0 |
8e673d2ccabeb0450faf30dcb347b9ff2d710ae2
|
https://github.com/kpandey008/SAGAN/tree/8e673d2ccabeb0450faf30dcb347b9ff2d710ae2
|
TransposeConv2dLayer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/oj/cojl5mb3pzv5jbmfzjkbac5hekbmpvb72kof6ouyyasitrogdd6n.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._unsafe_index]
# Source node to ATen node mapping:
# x => _unsafe_index
# Graph fragment:
# %_unsafe_index : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %unsqueeze, %convert_element_type_1]), kwargs = {})
triton_poi_fused__unsafe_index_0 = async_compile.triton('triton_poi_fused__unsafe_index_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_index_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_index_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 8) % 8
x0 = xindex % 8
x2 = (xindex // 64)
x4 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = x0
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + (4*tmp4) + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x4), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/uo/cuoiyqgsyrfp53lkw4hij4ulyfkzax64rqr6gxumyfhn6ponmpoc.py
# Topologically Sorted Source Nodes: [x_2, x_3], Original ATen: [aten.convolution, aten.leaky_relu, aten.leaky_relu_backward]
# Source node to ATen node mapping:
# x_2 => convolution
# x_3 => gt, mul_4, where
# Graph fragment:
# %convolution : [num_users=3] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 0.2), kwargs = {})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt, %convolution, %mul_4), kwargs = {})
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%where, 0), kwargs = {})
triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1 = async_compile.triton('triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 25) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = tmp7 > tmp3
tl.store(in_out_ptr0 + (x3), tmp7, xmask)
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._unsafe_index]
stream0 = get_raw_stream(0)
triton_poi_fused__unsafe_index_0.run(primals_1, buf0, 1024, grid=grid(1024), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 5, 5), (100, 25, 5, 1))
buf2 = buf1; del buf1 # reuse
buf3 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_2, x_3], Original ATen: [aten.convolution, aten.leaky_relu, aten.leaky_relu_backward]
triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1.run(buf2, primals_3, buf3, 400, grid=grid(400), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import functional as F
from torch.nn import Parameter
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class Conv2dLayer(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='elu', norm=
'none', sn=False):
super(Conv2dLayer, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU(inplace=True)
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
def forward(self, x):
x = self.pad(x)
x = self.conv2d(x)
if self.norm:
x = self.norm(x)
if self.activation:
x = self.activation(x)
return x
class TransposeConv2dLayer(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='lrelu', norm=
'none', sn=False, scale_factor=2):
super(TransposeConv2dLayer, self).__init__()
self.scale_factor = scale_factor
self.conv2d = Conv2dLayer(in_channels, out_channels, kernel_size,
stride, padding, dilation, pad_type, activation, norm, sn)
def forward(self, x):
x = F.interpolate(x, scale_factor=self.scale_factor, mode='nearest')
x = self.conv2d(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__unsafe_index_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 8 % 8
x0 = xindex % 8
x2 = xindex // 64
x4 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = x0
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + 4 * tmp4 + 16 * x2), xmask,
eviction_policy='evict_last')
tl.store(out_ptr0 + x4, tmp9, xmask)
@triton.jit
def triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1(in_out_ptr0,
in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 25 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = tmp7 > tmp3
tl.store(in_out_ptr0 + x3, tmp7, xmask)
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__unsafe_index_0[grid(1024)](primals_1, buf0, 1024,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 5, 5), (100, 25, 5, 1))
buf2 = buf1
del buf1
buf3 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.bool)
triton_poi_fused_convolution_leaky_relu_leaky_relu_backward_1[grid(400)
](buf2, primals_3, buf3, 400, XBLOCK=256, num_warps=4, num_stages=1
)
del primals_3
return buf2, primals_2, buf0, buf3
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class Conv2dLayer(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='elu', norm=
'none', sn=False):
super(Conv2dLayer, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU(inplace=True)
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
def forward(self, x):
x = self.pad(x)
x = self.conv2d(x)
if self.norm:
x = self.norm(x)
if self.activation:
x = self.activation(x)
return x
class TransposeConv2dLayerNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='lrelu', norm=
'none', sn=False, scale_factor=2):
super(TransposeConv2dLayerNew, self).__init__()
self.scale_factor = scale_factor
self.conv2d = Conv2dLayer(in_channels, out_channels, kernel_size,
stride, padding, dilation, pad_type, activation, norm, sn)
def forward(self, input_0):
primals_1 = self.conv2d.conv2d.weight
primals_3 = self.conv2d.conv2d.bias
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
kangzhiq/DeepFillv2_Pytorch
|
TransposeConv2dLayer
| false | 10,433 |
[
"MIT"
] | 0 |
9c7ed61b25bb995713f89108b712490737abe1b1
|
https://github.com/kangzhiq/DeepFillv2_Pytorch/tree/9c7ed61b25bb995713f89108b712490737abe1b1
|
Net
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/64/c6433gmrrnjxfqy257tszqraltgza5i6uxb7jrygvwczu6poqbwh.py
# Topologically Sorted Source Nodes: [conv2d, relu], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d => convolution
# relu => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_0 = async_compile.triton('triton_poi_fused_convolution_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 156800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 784) % 50
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sl/csldohruv4ml6lfysawazvg2w6uwqz6rik7zvmyl6nevkhfstrv4.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# x => getitem, getitem_1
# Graph fragment:
# %getitem : [num_users=2] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 0), kwargs = {})
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 1), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_1 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i8', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 39200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 14
x3 = (xindex // 14)
x2 = (xindex // 9800)
x4 = xindex % 9800
tmp0 = tl.load(in_ptr0 + ((2*x0) + (56*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (56*x3)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (28 + (2*x0) + (56*x3)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (29 + (2*x0) + (56*x3)), xmask, eviction_policy='evict_last')
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + (x4 + (9824*x2)), tmp6, xmask)
tl.store(out_ptr1 + (x4 + (9856*x2)), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xe/cxelxvpw3asckozc53rh36773aohp5hqpbp2nos5ymcdqhxvo4bl.py
# Topologically Sorted Source Nodes: [conv2d_1, relu_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# relu_1 => relu_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%getitem, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_1,), kwargs = {})
triton_poi_fused_convolution_relu_2 = async_compile.triton('triton_poi_fused_convolution_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 100) % 16
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tn/ctnw4tbgfy47ppke77vu7rtiz7dl5o3ahickx4p64n7c5rmrrix6.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# x_1 => _low_memory_max_pool2d_with_offsets_1, getitem_3
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets_1 : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%relu_1, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_3 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_1, 1), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_3 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_3(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 1600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 5
x1 = (xindex // 5)
x2 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (20*x1)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (20*x1)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (10 + (2*x0) + (20*x1)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (11 + (2*x0) + (20*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tl.store(out_ptr0 + (x2), tmp15, xmask)
tl.store(out_ptr1 + (x2), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jn/cjnqv3sgcv5x2iz7ij5zdad6ofabcnonrlksgsxu2ob7n274gz6b.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_3 => relu_2
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_7), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_4 = async_compile.triton('triton_poi_fused_relu_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 480
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 120
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6m/c6m6u2ctjb4r4ra3sizrwezzkzegfp2ombflmfg3dwjfci2pen7h.py
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_4 => relu_3
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_9), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_5 = async_compile.triton('triton_poi_fused_relu_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_5(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 336
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 84
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (50, 3, 5, 5), (75, 25, 5, 1))
assert_size_stride(primals_2, (50, ), (1, ))
assert_size_stride(primals_3, (4, 3, 32, 32), (3072, 1024, 32, 1))
assert_size_stride(primals_4, (16, 50, 5, 5), (1250, 25, 5, 1))
assert_size_stride(primals_5, (16, ), (1, ))
assert_size_stride(primals_6, (120, 400), (400, 1))
assert_size_stride(primals_7, (120, ), (1, ))
assert_size_stride(primals_8, (84, 120), (120, 1))
assert_size_stride(primals_9, (84, ), (1, ))
assert_size_stride(primals_10, (10, 84), (84, 1))
assert_size_stride(primals_11, (10, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 50, 28, 28), (39200, 784, 28, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d, relu], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 156800, grid=grid(156800), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 50, 14, 14), (9824, 196, 14, 1), torch.float32)
buf3 = empty_strided_cuda((4, 50, 14, 14), (9856, 196, 14, 1), torch.int8)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.max_pool2d_with_indices]
triton_poi_fused_max_pool2d_with_indices_1.run(buf1, buf2, buf3, 39200, grid=grid(39200), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 16, 10, 10), (1600, 100, 10, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [conv2d_1, relu_1], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_2.run(buf5, primals_5, 6400, grid=grid(6400), stream=stream0)
del primals_5
buf6 = empty_strided_cuda((4, 16, 5, 5), (400, 25, 5, 1), torch.int8)
buf7 = empty_strided_cuda((4, 16, 5, 5), (400, 25, 5, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.max_pool2d_with_indices]
triton_poi_fused_max_pool2d_with_indices_3.run(buf5, buf6, buf7, 1600, grid=grid(1600), stream=stream0)
buf8 = empty_strided_cuda((4, 120), (120, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf7, (4, 400), (400, 1), 0), reinterpret_tensor(primals_6, (400, 120), (1, 400), 0), out=buf8)
buf9 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
triton_poi_fused_relu_4.run(buf9, primals_7, 480, grid=grid(480), stream=stream0)
del primals_7
buf10 = empty_strided_cuda((4, 84), (84, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf9, reinterpret_tensor(primals_8, (120, 84), (1, 120), 0), out=buf10)
buf11 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.relu]
triton_poi_fused_relu_5.run(buf11, primals_9, 336, grid=grid(336), stream=stream0)
del primals_9
buf12 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, buf11, reinterpret_tensor(primals_10, (84, 10), (1, 84), 0), alpha=1, beta=1, out=buf12)
del primals_11
return (buf12, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5, buf6, reinterpret_tensor(buf7, (4, 400), (400, 1), 0), buf9, buf11, primals_10, primals_8, primals_6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((50, 3, 5, 5), (75, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((50, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 3, 32, 32), (3072, 1024, 32, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((16, 50, 5, 5), (1250, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((120, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((120, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((84, 120), (120, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((84, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((10, 84), (84, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((10, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(3, 50, 5)
self.pool = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(50, 16, 5)
self.fc1 = nn.Linear(16 * 5 * 5, 120)
self.fc2 = nn.Linear(120, 84)
self.fc3 = nn.Linear(84, 10)
def forward(self, x):
x = self.pool(F.relu(self.conv1(x)))
x = self.pool(F.relu(self.conv2(x)))
x = x.view(-1, 16 * 5 * 5)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
def get_inputs():
return [torch.rand([4, 3, 32, 32])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 156800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 784 % 50
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_1(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 39200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 14
x3 = xindex // 14
x2 = xindex // 9800
x4 = xindex % 9800
tmp0 = tl.load(in_ptr0 + (2 * x0 + 56 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 56 * x3), xmask, eviction_policy
='evict_last')
tmp3 = tl.load(in_ptr0 + (28 + 2 * x0 + 56 * x3), xmask,
eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (29 + 2 * x0 + 56 * x3), xmask,
eviction_policy='evict_last')
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + (x4 + 9824 * x2), tmp6, xmask)
tl.store(out_ptr1 + (x4 + 9856 * x2), tmp16, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 100 % 16
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_3(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 1600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 5
x1 = xindex // 5
x2 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 20 * x1), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 20 * x1), xmask, eviction_policy
='evict_last')
tmp7 = tl.load(in_ptr0 + (10 + 2 * x0 + 20 * x1), xmask,
eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (11 + 2 * x0 + 20 * x1), xmask,
eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tl.store(out_ptr0 + x2, tmp15, xmask)
tl.store(out_ptr1 + x2, tmp16, xmask)
@triton.jit
def triton_poi_fused_relu_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 480
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 120
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_relu_5(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 336
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 84
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (50, 3, 5, 5), (75, 25, 5, 1))
assert_size_stride(primals_2, (50,), (1,))
assert_size_stride(primals_3, (4, 3, 32, 32), (3072, 1024, 32, 1))
assert_size_stride(primals_4, (16, 50, 5, 5), (1250, 25, 5, 1))
assert_size_stride(primals_5, (16,), (1,))
assert_size_stride(primals_6, (120, 400), (400, 1))
assert_size_stride(primals_7, (120,), (1,))
assert_size_stride(primals_8, (84, 120), (120, 1))
assert_size_stride(primals_9, (84,), (1,))
assert_size_stride(primals_10, (10, 84), (84, 1))
assert_size_stride(primals_11, (10,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 50, 28, 28), (39200, 784, 28, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(156800)](buf1, primals_2,
156800, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 50, 14, 14), (9824, 196, 14, 1),
torch.float32)
buf3 = empty_strided_cuda((4, 50, 14, 14), (9856, 196, 14, 1),
torch.int8)
triton_poi_fused_max_pool2d_with_indices_1[grid(39200)](buf1, buf2,
buf3, 39200, XBLOCK=512, num_warps=4, num_stages=1)
buf4 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 16, 10, 10), (1600, 100, 10, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_relu_2[grid(6400)](buf5, primals_5,
6400, XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf6 = empty_strided_cuda((4, 16, 5, 5), (400, 25, 5, 1), torch.int8)
buf7 = empty_strided_cuda((4, 16, 5, 5), (400, 25, 5, 1), torch.float32
)
triton_poi_fused_max_pool2d_with_indices_3[grid(1600)](buf5, buf6,
buf7, 1600, XBLOCK=128, num_warps=4, num_stages=1)
buf8 = empty_strided_cuda((4, 120), (120, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf7, (4, 400), (400, 1), 0),
reinterpret_tensor(primals_6, (400, 120), (1, 400), 0), out=buf8)
buf9 = buf8
del buf8
triton_poi_fused_relu_4[grid(480)](buf9, primals_7, 480, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_7
buf10 = empty_strided_cuda((4, 84), (84, 1), torch.float32)
extern_kernels.mm(buf9, reinterpret_tensor(primals_8, (120, 84), (1,
120), 0), out=buf10)
buf11 = buf10
del buf10
triton_poi_fused_relu_5[grid(336)](buf11, primals_9, 336, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_9
buf12 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
extern_kernels.addmm(primals_11, buf11, reinterpret_tensor(
primals_10, (84, 10), (1, 84), 0), alpha=1, beta=1, out=buf12)
del primals_11
return (buf12, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5,
buf6, reinterpret_tensor(buf7, (4, 400), (400, 1), 0), buf9, buf11,
primals_10, primals_8, primals_6)
class NetNew(nn.Module):
def __init__(self):
super(NetNew, self).__init__()
self.conv1 = nn.Conv2d(3, 50, 5)
self.pool = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(50, 16, 5)
self.fc1 = nn.Linear(16 * 5 * 5, 120)
self.fc2 = nn.Linear(120, 84)
self.fc3 = nn.Linear(84, 10)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_6 = self.fc1.weight
primals_7 = self.fc1.bias
primals_8 = self.fc2.weight
primals_9 = self.fc2.bias
primals_10 = self.fc3.weight
primals_11 = self.fc3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
|
lykasbongbongbong/Pytorch
|
Net
| false | 10,434 |
[
"MIT"
] | 0 |
f01d89fb51ac939f5a110f5ab6190c11917e66fc
|
https://github.com/lykasbongbongbong/Pytorch/tree/f01d89fb51ac939f5a110f5ab6190c11917e66fc
|
Max_AvgPool
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/jg/cjgfftmpiknb3zm55q2motbdtnmpjvcqr5oxilzriywq5mqodg3b.py
# Topologically Sorted Source Nodes: [max_pool2d, avg_pool2d, x], Original ATen: [aten.max_pool2d_with_indices, aten.avg_pool2d, aten.add]
# Source node to ATen node mapping:
# avg_pool2d => avg_pool2d
# max_pool2d => _low_memory_max_pool2d_with_offsets
# x => add
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets : [num_users=1] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%arg0_1, [3, 3], [2, 2], [1, 1], [1, 1], False), kwargs = {})
# %avg_pool2d : [num_users=1] = call_function[target=torch.ops.aten.avg_pool2d.default](args = (%arg0_1, [3, 3], [2, 2], [1, 1]), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, %avg_pool2d), kwargs = {})
triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0 = async_compile.triton('triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 18, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 2) % 2
x0 = xindex % 2
x4 = (xindex // 2)
x3 = xindex
tmp0 = (-1) + (2*x1)
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp2 & tmp4
tmp6 = (-1) + (2*x0)
tmp7 = tmp6 >= tmp1
tmp8 = tmp6 < tmp3
tmp9 = tmp7 & tmp8
tmp10 = tmp5 & tmp9
tmp11 = tl.load(in_ptr0 + ((-5) + (2*x0) + (8*x4)), tmp10 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp12 = 2*x0
tmp13 = tmp12 >= tmp1
tmp14 = tmp12 < tmp3
tmp15 = tmp13 & tmp14
tmp16 = tmp5 & tmp15
tmp17 = tl.load(in_ptr0 + ((-4) + (2*x0) + (8*x4)), tmp16 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp18 = triton_helpers.maximum(tmp17, tmp11)
tmp19 = 1 + (2*x0)
tmp20 = tmp19 >= tmp1
tmp21 = tmp19 < tmp3
tmp22 = tmp20 & tmp21
tmp23 = tmp5 & tmp22
tmp24 = tl.load(in_ptr0 + ((-3) + (2*x0) + (8*x4)), tmp23 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp25 = triton_helpers.maximum(tmp24, tmp18)
tmp26 = 2*x1
tmp27 = tmp26 >= tmp1
tmp28 = tmp26 < tmp3
tmp29 = tmp27 & tmp28
tmp30 = tmp29 & tmp9
tmp31 = tl.load(in_ptr0 + ((-1) + (2*x0) + (8*x4)), tmp30 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp32 = triton_helpers.maximum(tmp31, tmp25)
tmp33 = tmp29 & tmp15
tmp34 = tl.load(in_ptr0 + ((2*x0) + (8*x4)), tmp33 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp35 = triton_helpers.maximum(tmp34, tmp32)
tmp36 = tmp29 & tmp22
tmp37 = tl.load(in_ptr0 + (1 + (2*x0) + (8*x4)), tmp36 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp38 = triton_helpers.maximum(tmp37, tmp35)
tmp39 = 1 + (2*x1)
tmp40 = tmp39 >= tmp1
tmp41 = tmp39 < tmp3
tmp42 = tmp40 & tmp41
tmp43 = tmp42 & tmp9
tmp44 = tl.load(in_ptr0 + (3 + (2*x0) + (8*x4)), tmp43 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp45 = triton_helpers.maximum(tmp44, tmp38)
tmp46 = tmp42 & tmp15
tmp47 = tl.load(in_ptr0 + (4 + (2*x0) + (8*x4)), tmp46 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp48 = triton_helpers.maximum(tmp47, tmp45)
tmp49 = tmp42 & tmp22
tmp50 = tl.load(in_ptr0 + (5 + (2*x0) + (8*x4)), tmp49 & xmask, eviction_policy='evict_last', other=float("-inf"))
tmp51 = triton_helpers.maximum(tmp50, tmp48)
tmp52 = tl.load(in_ptr0 + ((-5) + (2*x0) + (8*x4)), tmp10 & xmask, eviction_policy='evict_last', other=0.0)
tmp53 = tl.load(in_ptr0 + ((-4) + (2*x0) + (8*x4)), tmp16 & xmask, eviction_policy='evict_last', other=0.0)
tmp54 = tmp53 + tmp52
tmp55 = tl.load(in_ptr0 + ((-3) + (2*x0) + (8*x4)), tmp23 & xmask, eviction_policy='evict_last', other=0.0)
tmp56 = tmp55 + tmp54
tmp57 = tl.load(in_ptr0 + ((-1) + (2*x0) + (8*x4)), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp58 = tmp57 + tmp56
tmp59 = tl.load(in_ptr0 + ((2*x0) + (8*x4)), tmp33 & xmask, eviction_policy='evict_last', other=0.0)
tmp60 = tmp59 + tmp58
tmp61 = tl.load(in_ptr0 + (1 + (2*x0) + (8*x4)), tmp36 & xmask, eviction_policy='evict_last', other=0.0)
tmp62 = tmp61 + tmp60
tmp63 = tl.load(in_ptr0 + (3 + (2*x0) + (8*x4)), tmp43 & xmask, eviction_policy='evict_last', other=0.0)
tmp64 = tmp63 + tmp62
tmp65 = tl.load(in_ptr0 + (4 + (2*x0) + (8*x4)), tmp46 & xmask, eviction_policy='evict_last', other=0.0)
tmp66 = tmp65 + tmp64
tmp67 = tl.load(in_ptr0 + (5 + (2*x0) + (8*x4)), tmp49 & xmask, eviction_policy='evict_last', other=0.0)
tmp68 = tmp67 + tmp66
tmp69 = 1 + ((-2)*x0) + ((-2)*x1) + (((5) * ((5) <= (2 + (2*x0))) + (2 + (2*x0)) * ((2 + (2*x0)) < (5)))*((5) * ((5) <= (2 + (2*x1))) + (2 + (2*x1)) * ((2 + (2*x1)) < (5)))) + ((-2)*x0*((5) * ((5) <= (2 + (2*x1))) + (2 + (2*x1)) * ((2 + (2*x1)) < (5)))) + ((-2)*x1*((5) * ((5) <= (2 + (2*x0))) + (2 + (2*x0)) * ((2 + (2*x0)) < (5)))) + (4*x0*x1) + ((5) * ((5) <= (2 + (2*x0))) + (2 + (2*x0)) * ((2 + (2*x0)) < (5))) + ((5) * ((5) <= (2 + (2*x1))) + (2 + (2*x1)) * ((2 + (2*x1)) < (5)))
tmp70 = tmp68 / tmp69
tmp71 = tmp51 + tmp70
tl.store(in_out_ptr0 + (x3), tmp71, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [max_pool2d, avg_pool2d, x], Original ATen: [aten.max_pool2d_with_indices, aten.avg_pool2d, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0.run(buf2, arg0_1, 64, grid=grid(64), stream=stream0)
del arg0_1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from itertools import product as product
class Max_AvgPool(nn.Module):
def __init__(self, kernel_size=(3, 3), stride=2, padding=1, dim=128):
super(Max_AvgPool, self).__init__()
self.Maxpool = nn.MaxPool2d(kernel_size=kernel_size, stride=stride,
padding=padding)
self.Avgpool = nn.AvgPool2d(kernel_size=kernel_size, stride=stride,
padding=padding)
def forward(self, x):
x = self.Maxpool(x) + self.Avgpool(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0(in_out_ptr0,
in_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 2 % 2
x0 = xindex % 2
x4 = xindex // 2
x3 = xindex
tmp0 = -1 + 2 * x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp2 & tmp4
tmp6 = -1 + 2 * x0
tmp7 = tmp6 >= tmp1
tmp8 = tmp6 < tmp3
tmp9 = tmp7 & tmp8
tmp10 = tmp5 & tmp9
tmp11 = tl.load(in_ptr0 + (-5 + 2 * x0 + 8 * x4), tmp10 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp12 = 2 * x0
tmp13 = tmp12 >= tmp1
tmp14 = tmp12 < tmp3
tmp15 = tmp13 & tmp14
tmp16 = tmp5 & tmp15
tmp17 = tl.load(in_ptr0 + (-4 + 2 * x0 + 8 * x4), tmp16 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp18 = triton_helpers.maximum(tmp17, tmp11)
tmp19 = 1 + 2 * x0
tmp20 = tmp19 >= tmp1
tmp21 = tmp19 < tmp3
tmp22 = tmp20 & tmp21
tmp23 = tmp5 & tmp22
tmp24 = tl.load(in_ptr0 + (-3 + 2 * x0 + 8 * x4), tmp23 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp25 = triton_helpers.maximum(tmp24, tmp18)
tmp26 = 2 * x1
tmp27 = tmp26 >= tmp1
tmp28 = tmp26 < tmp3
tmp29 = tmp27 & tmp28
tmp30 = tmp29 & tmp9
tmp31 = tl.load(in_ptr0 + (-1 + 2 * x0 + 8 * x4), tmp30 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp32 = triton_helpers.maximum(tmp31, tmp25)
tmp33 = tmp29 & tmp15
tmp34 = tl.load(in_ptr0 + (2 * x0 + 8 * x4), tmp33 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp35 = triton_helpers.maximum(tmp34, tmp32)
tmp36 = tmp29 & tmp22
tmp37 = tl.load(in_ptr0 + (1 + 2 * x0 + 8 * x4), tmp36 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp38 = triton_helpers.maximum(tmp37, tmp35)
tmp39 = 1 + 2 * x1
tmp40 = tmp39 >= tmp1
tmp41 = tmp39 < tmp3
tmp42 = tmp40 & tmp41
tmp43 = tmp42 & tmp9
tmp44 = tl.load(in_ptr0 + (3 + 2 * x0 + 8 * x4), tmp43 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp45 = triton_helpers.maximum(tmp44, tmp38)
tmp46 = tmp42 & tmp15
tmp47 = tl.load(in_ptr0 + (4 + 2 * x0 + 8 * x4), tmp46 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp48 = triton_helpers.maximum(tmp47, tmp45)
tmp49 = tmp42 & tmp22
tmp50 = tl.load(in_ptr0 + (5 + 2 * x0 + 8 * x4), tmp49 & xmask,
eviction_policy='evict_last', other=float('-inf'))
tmp51 = triton_helpers.maximum(tmp50, tmp48)
tmp52 = tl.load(in_ptr0 + (-5 + 2 * x0 + 8 * x4), tmp10 & xmask,
eviction_policy='evict_last', other=0.0)
tmp53 = tl.load(in_ptr0 + (-4 + 2 * x0 + 8 * x4), tmp16 & xmask,
eviction_policy='evict_last', other=0.0)
tmp54 = tmp53 + tmp52
tmp55 = tl.load(in_ptr0 + (-3 + 2 * x0 + 8 * x4), tmp23 & xmask,
eviction_policy='evict_last', other=0.0)
tmp56 = tmp55 + tmp54
tmp57 = tl.load(in_ptr0 + (-1 + 2 * x0 + 8 * x4), tmp30 & xmask,
eviction_policy='evict_last', other=0.0)
tmp58 = tmp57 + tmp56
tmp59 = tl.load(in_ptr0 + (2 * x0 + 8 * x4), tmp33 & xmask,
eviction_policy='evict_last', other=0.0)
tmp60 = tmp59 + tmp58
tmp61 = tl.load(in_ptr0 + (1 + 2 * x0 + 8 * x4), tmp36 & xmask,
eviction_policy='evict_last', other=0.0)
tmp62 = tmp61 + tmp60
tmp63 = tl.load(in_ptr0 + (3 + 2 * x0 + 8 * x4), tmp43 & xmask,
eviction_policy='evict_last', other=0.0)
tmp64 = tmp63 + tmp62
tmp65 = tl.load(in_ptr0 + (4 + 2 * x0 + 8 * x4), tmp46 & xmask,
eviction_policy='evict_last', other=0.0)
tmp66 = tmp65 + tmp64
tmp67 = tl.load(in_ptr0 + (5 + 2 * x0 + 8 * x4), tmp49 & xmask,
eviction_policy='evict_last', other=0.0)
tmp68 = tmp67 + tmp66
tmp69 = 1 + -2 * x0 + -2 * x1 + (5 * (5 <= 2 + 2 * x0) + (2 + 2 * x0) *
(2 + 2 * x0 < 5)) * (5 * (5 <= 2 + 2 * x1) + (2 + 2 * x1) * (2 + 2 *
x1 < 5)) + -2 * x0 * (5 * (5 <= 2 + 2 * x1) + (2 + 2 * x1) * (2 + 2 *
x1 < 5)) + -2 * x1 * (5 * (5 <= 2 + 2 * x0) + (2 + 2 * x0) * (2 + 2 *
x0 < 5)) + 4 * x0 * x1 + (5 * (5 <= 2 + 2 * x0) + (2 + 2 * x0) * (2 +
2 * x0 < 5)) + (5 * (5 <= 2 + 2 * x1) + (2 + 2 * x1) * (2 + 2 * x1 < 5)
)
tmp70 = tmp68 / tmp69
tmp71 = tmp51 + tmp70
tl.store(in_out_ptr0 + x3, tmp71, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_add_avg_pool2d_max_pool2d_with_indices_0[grid(64)](
buf2, arg0_1, 64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
return buf2,
class Max_AvgPoolNew(nn.Module):
def __init__(self, kernel_size=(3, 3), stride=2, padding=1, dim=128):
super(Max_AvgPoolNew, self).__init__()
self.Maxpool = nn.MaxPool2d(kernel_size=kernel_size, stride=stride,
padding=padding)
self.Avgpool = nn.AvgPool2d(kernel_size=kernel_size, stride=stride,
padding=padding)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
kooBH/EXTD_Pytorch
|
Max_AvgPool
| false | 10,435 |
[
"MIT"
] | 0 |
e93b196c87054684cc6c757e1dfd26f8b7dc57cf
|
https://github.com/kooBH/EXTD_Pytorch/tree/e93b196c87054684cc6c757e1dfd26f8b7dc57cf
|
my_AvgPool2d
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/vb/cvbbi4nnli3horxmlcv2azovajihodz7seyhvtrxgtxbdi7voix6.py
# Topologically Sorted Source Nodes: [input_2], Original ATen: [aten.avg_pool2d]
# Source node to ATen node mapping:
# input_2 => avg_pool2d
# Graph fragment:
# %avg_pool2d : [num_users=1] = call_function[target=torch.ops.aten.avg_pool2d.default](args = (%permute, [4, 4], [4, 4]), kwargs = {})
triton_poi_fused_avg_pool2d_0 = async_compile.triton('triton_poi_fused_avg_pool2d_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_avg_pool2d_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_avg_pool2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp7 = tl.load(in_ptr0 + (4 + x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr0 + (20 + x0 + (64*x1)), xmask)
tmp11 = tl.load(in_ptr0 + (36 + x0 + (64*x1)), xmask)
tmp13 = tl.load(in_ptr0 + (52 + x0 + (64*x1)), xmask)
tmp15 = tl.load(in_ptr0 + (8 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (24 + x0 + (64*x1)), xmask)
tmp19 = tl.load(in_ptr0 + (40 + x0 + (64*x1)), xmask)
tmp21 = tl.load(in_ptr0 + (56 + x0 + (64*x1)), xmask)
tmp23 = tl.load(in_ptr0 + (12 + x0 + (64*x1)), xmask)
tmp25 = tl.load(in_ptr0 + (28 + x0 + (64*x1)), xmask)
tmp27 = tl.load(in_ptr0 + (44 + x0 + (64*x1)), xmask)
tmp29 = tl.load(in_ptr0 + (60 + x0 + (64*x1)), xmask)
tmp2 = tmp1 + tmp0
tmp4 = tmp3 + tmp2
tmp6 = tmp5 + tmp4
tmp8 = tmp7 + tmp6
tmp10 = tmp9 + tmp8
tmp12 = tmp11 + tmp10
tmp14 = tmp13 + tmp12
tmp16 = tmp15 + tmp14
tmp18 = tmp17 + tmp16
tmp20 = tmp19 + tmp18
tmp22 = tmp21 + tmp20
tmp24 = tmp23 + tmp22
tmp26 = tmp25 + tmp24
tmp28 = tmp27 + tmp26
tmp30 = tmp29 + tmp28
tmp31 = 0.0625
tmp32 = tmp30 * tmp31
tl.store(out_ptr0 + (x2), tmp32, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_2], Original ATen: [aten.avg_pool2d]
stream0 = get_raw_stream(0)
triton_poi_fused_avg_pool2d_0.run(arg0_1, buf0, 16, grid=grid(16), stream=stream0)
del arg0_1
return (reinterpret_tensor(buf0, (4, 1, 1, 4), (4, 1, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
import torch.nn.functional as F
from torch.nn.modules.module import Module
class my_AvgPool2d(Module):
def __init__(self, kernel_size, stride=None, padding=0, ceil_mode=False,
count_include_pad=True):
super(my_AvgPool2d, self).__init__()
self.kernel_size = kernel_size
self.stride = stride or kernel_size
self.padding = padding
self.ceil_mode = ceil_mode
self.count_include_pad = count_include_pad
def forward(self, input):
input = input.transpose(3, 1)
input = F.avg_pool2d(input, self.kernel_size, self.stride, self.
padding, self.ceil_mode, self.count_include_pad)
input = input.transpose(3, 1).contiguous()
return input
def __repr__(self):
return self.__class__.__name__ + '(' + 'kernel_size=' + str(self.
kernel_size) + ', stride=' + str(self.stride) + ', padding=' + str(
self.padding) + ', ceil_mode=' + str(self.ceil_mode
) + ', count_include_pad=' + str(self.count_include_pad) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'kernel_size': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch.nn import Module
from torch.nn.modules.module import Module
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_avg_pool2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp7 = tl.load(in_ptr0 + (4 + x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr0 + (20 + x0 + 64 * x1), xmask)
tmp11 = tl.load(in_ptr0 + (36 + x0 + 64 * x1), xmask)
tmp13 = tl.load(in_ptr0 + (52 + x0 + 64 * x1), xmask)
tmp15 = tl.load(in_ptr0 + (8 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (24 + x0 + 64 * x1), xmask)
tmp19 = tl.load(in_ptr0 + (40 + x0 + 64 * x1), xmask)
tmp21 = tl.load(in_ptr0 + (56 + x0 + 64 * x1), xmask)
tmp23 = tl.load(in_ptr0 + (12 + x0 + 64 * x1), xmask)
tmp25 = tl.load(in_ptr0 + (28 + x0 + 64 * x1), xmask)
tmp27 = tl.load(in_ptr0 + (44 + x0 + 64 * x1), xmask)
tmp29 = tl.load(in_ptr0 + (60 + x0 + 64 * x1), xmask)
tmp2 = tmp1 + tmp0
tmp4 = tmp3 + tmp2
tmp6 = tmp5 + tmp4
tmp8 = tmp7 + tmp6
tmp10 = tmp9 + tmp8
tmp12 = tmp11 + tmp10
tmp14 = tmp13 + tmp12
tmp16 = tmp15 + tmp14
tmp18 = tmp17 + tmp16
tmp20 = tmp19 + tmp18
tmp22 = tmp21 + tmp20
tmp24 = tmp23 + tmp22
tmp26 = tmp25 + tmp24
tmp28 = tmp27 + tmp26
tmp30 = tmp29 + tmp28
tmp31 = 0.0625
tmp32 = tmp30 * tmp31
tl.store(out_ptr0 + x2, tmp32, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
get_raw_stream(0)
triton_poi_fused_avg_pool2d_0[grid(16)](arg0_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del arg0_1
return reinterpret_tensor(buf0, (4, 1, 1, 4), (4, 1, 1, 1), 0),
class my_AvgPool2dNew(Module):
def __init__(self, kernel_size, stride=None, padding=0, ceil_mode=False,
count_include_pad=True):
super(my_AvgPool2dNew, self).__init__()
self.kernel_size = kernel_size
self.stride = stride or kernel_size
self.padding = padding
self.ceil_mode = ceil_mode
self.count_include_pad = count_include_pad
def __repr__(self):
return self.__class__.__name__ + '(' + 'kernel_size=' + str(self.
kernel_size) + ', stride=' + str(self.stride) + ', padding=' + str(
self.padding) + ', ceil_mode=' + str(self.ceil_mode
) + ', count_include_pad=' + str(self.count_include_pad) + ')'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
likun97/Low_quality_classification_with_mobilenetv3
|
my_AvgPool2d
| false | 10,436 |
[
"Apache-2.0"
] | 0 |
a9e6f66caad937fc7c8e101cddb76f116219b255
|
https://github.com/likun97/Low_quality_classification_with_mobilenetv3/tree/a9e6f66caad937fc7c8e101cddb76f116219b255
|
Conv2dLayer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ue/cuecegnhgafe2dsjwb2idu7ooicbmsi2pwlqk5kxrayxsv6nzpux.py
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.convolution, aten.elu]
# Source node to ATen node mapping:
# x_1 => convolution
# x_2 => expm1, gt, mul, mul_2, where
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 1.0), kwargs = {})
# %expm1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1, 1.0), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul, %mul_2), kwargs = {})
triton_poi_fused_convolution_elu_0 = async_compile.triton('triton_poi_fused_convolution_elu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_elu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_elu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 1.0
tmp6 = tmp2 * tmp5
tmp7 = libdevice.expm1(tmp6)
tmp8 = tmp7 * tmp5
tmp9 = tl.where(tmp4, tmp6, tmp8)
tl.store(in_out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 1, 1), (4, 1, 1, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.convolution, aten.elu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_elu_0.run(buf1, primals_3, 16, grid=grid(16), stream=stream0)
del primals_3
return (buf1, primals_1, primals_2, buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import Parameter
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class Conv2dLayer(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='elu', norm=
'none', sn=False):
super(Conv2dLayer, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU(inplace=True)
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
def forward(self, x):
x = self.pad(x)
x = self.conv2d(x)
if self.norm:
x = self.norm(x)
if self.activation:
x = self.activation(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_poi_fused_convolution_elu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 1.0
tmp6 = tmp2 * tmp5
tmp7 = libdevice.expm1(tmp6)
tmp8 = tmp7 * tmp5
tmp9 = tl.where(tmp4, tmp6, tmp8)
tl.store(in_out_ptr0 + x2, tmp9, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 1, 1), (4, 1, 1, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_elu_0[grid(16)](buf1, primals_3, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_3
return buf1, primals_1, primals_2, buf1
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class Conv2dLayerNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='elu', norm=
'none', sn=False):
super(Conv2dLayerNew, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU(inplace=True)
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
def forward(self, input_0):
primals_1 = self.conv2d.weight
primals_3 = self.conv2d.bias
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
kangzhiq/DeepFillv2_Pytorch
|
Conv2dLayer
| false | 10,437 |
[
"MIT"
] | 0 |
9c7ed61b25bb995713f89108b712490737abe1b1
|
https://github.com/kangzhiq/DeepFillv2_Pytorch/tree/9c7ed61b25bb995713f89108b712490737abe1b1
|
GatedConv2d
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/r3/cr3hlg2dj2d3nmsli5wlcbgrfym3b6ux3uuxd7pl3rggj6domt5d.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.reflection_pad2d]
# Source node to ATen node mapping:
# x => _unsafe_index, _unsafe_index_1
# Graph fragment:
# %_unsafe_index : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %sub_1, None]), kwargs = {})
# %_unsafe_index_1 : [num_users=3] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index, [None, None, None, %sub_1]), kwargs = {})
triton_poi_fused_reflection_pad2d_0 = async_compile.triton('triton_poi_fused_reflection_pad2d_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_reflection_pad2d_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + ((-1)*(tl_math.abs((-3) + x0))) + ((-4)*(tl_math.abs((-3) + x1))) + (16*x2)), xmask)
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/e2/ce2cdkzdgfs7x35ywqy3cnp5gbzraaqnnk3zmuic27ohalb55dzx.py
# Topologically Sorted Source Nodes: [conv, mask, gated_mask, conv_1, x_1], Original ATen: [aten.convolution, aten.sigmoid, aten.elu, aten.mul]
# Source node to ATen node mapping:
# conv => convolution
# conv_1 => expm1, gt, mul, mul_2, where
# gated_mask => sigmoid
# mask => convolution_1
# x_1 => mul_3
# Graph fragment:
# %convolution : [num_users=3] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_1, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_1,), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 1.0), kwargs = {})
# %expm1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1, 1.0), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul, %mul_2), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%where, %sigmoid), kwargs = {})
triton_poi_fused_convolution_elu_mul_sigmoid_1 = async_compile.triton('triton_poi_fused_convolution_elu_mul_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_elu_mul_sigmoid_1', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_elu_mul_sigmoid_1(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + (x2), xmask)
tmp4 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = 0.0
tmp7 = tmp2 > tmp6
tmp8 = 1.0
tmp9 = tmp2 * tmp8
tmp10 = libdevice.expm1(tmp9)
tmp11 = tmp10 * tmp8
tmp12 = tl.where(tmp7, tmp9, tmp11)
tmp13 = tl.sigmoid(tmp5)
tmp14 = tmp12 * tmp13
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
tl.store(in_out_ptr1 + (x2), tmp5, xmask)
tl.store(out_ptr0 + (x2), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.reflection_pad2d]
stream0 = get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [conv], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1, 1), (4, 1, 1, 1))
# Topologically Sorted Source Nodes: [mask], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf0, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 1, 1), (4, 1, 1, 1))
buf2 = buf1; del buf1 # reuse
buf4 = buf3; del buf3 # reuse
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv, mask, gated_mask, conv_1, x_1], Original ATen: [aten.convolution, aten.sigmoid, aten.elu, aten.mul]
triton_poi_fused_convolution_elu_mul_sigmoid_1.run(buf2, buf4, primals_3, primals_5, buf5, 16, grid=grid(16), stream=stream0)
del primals_3
del primals_5
return (buf5, primals_2, primals_4, buf0, buf2, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import Parameter
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class GatedConv2d(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='reflect', activation='elu', norm=
'none', sn=False):
super(GatedConv2d, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
self.mask_conv2d = SpectralNorm(nn.Conv2d(in_channels,
out_channels, kernel_size, stride, padding=0, dilation=
dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
self.mask_conv2d = nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation)
self.sigmoid = torch.nn.Sigmoid()
def forward(self, x):
x = self.pad(x)
conv = self.conv2d(x)
mask = self.mask_conv2d(x)
gated_mask = self.sigmoid(mask)
if self.activation:
conv = self.activation(conv)
x = conv * gated_mask
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + -1 * tl_math.abs(-3 + x0) + -4 * tl_math
.abs(-3 + x1) + 16 * x2), xmask)
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_convolution_elu_mul_sigmoid_1(in_out_ptr0, in_out_ptr1,
in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + x2, xmask)
tmp4 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = 0.0
tmp7 = tmp2 > tmp6
tmp8 = 1.0
tmp9 = tmp2 * tmp8
tmp10 = libdevice.expm1(tmp9)
tmp11 = tmp10 * tmp8
tmp12 = tl.where(tmp7, tmp9, tmp11)
tmp13 = tl.sigmoid(tmp5)
tmp14 = tmp12 * tmp13
tl.store(in_out_ptr0 + x2, tmp2, xmask)
tl.store(in_out_ptr1 + x2, tmp5, xmask)
tl.store(out_ptr0 + x2, tmp14, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0[grid(256)](primals_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1, 1), (4, 1, 1, 1))
buf3 = extern_kernels.convolution(buf0, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 1, 1), (4, 1, 1, 1))
buf2 = buf1
del buf1
buf4 = buf3
del buf3
buf5 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
triton_poi_fused_convolution_elu_mul_sigmoid_1[grid(16)](buf2, buf4,
primals_3, primals_5, buf5, 16, XBLOCK=16, num_warps=1,
num_stages=1)
del primals_3
del primals_5
return buf5, primals_2, primals_4, buf0, buf2, buf4
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class GatedConv2dNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='reflect', activation='elu', norm=
'none', sn=False):
super(GatedConv2dNew, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
self.mask_conv2d = SpectralNorm(nn.Conv2d(in_channels,
out_channels, kernel_size, stride, padding=0, dilation=
dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
self.mask_conv2d = nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation)
self.sigmoid = torch.nn.Sigmoid()
def forward(self, input_0):
primals_1 = self.conv2d.weight
primals_3 = self.conv2d.bias
primals_2 = self.mask_conv2d.weight
primals_5 = self.mask_conv2d.bias
primals_4 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
kangzhiq/DeepFillv2_Pytorch
|
GatedConv2d
| false | 10,438 |
[
"MIT"
] | 0 |
9c7ed61b25bb995713f89108b712490737abe1b1
|
https://github.com/kangzhiq/DeepFillv2_Pytorch/tree/9c7ed61b25bb995713f89108b712490737abe1b1
|
my_MaxPool2d
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/25/c252mdjutv7khw5cfthjqjyr4ilal3cc73gpc5h3h2hs2czd7hwz.py
# Topologically Sorted Source Nodes: [input_2], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# input_2 => _low_memory_max_pool2d_with_offsets
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets : [num_users=1] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%permute, [4, 4], [4, 4], [0, 0], [1, 1], False), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_0 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp7 = tl.load(in_ptr0 + (4 + x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr0 + (20 + x0 + (64*x1)), xmask)
tmp11 = tl.load(in_ptr0 + (36 + x0 + (64*x1)), xmask)
tmp13 = tl.load(in_ptr0 + (52 + x0 + (64*x1)), xmask)
tmp15 = tl.load(in_ptr0 + (8 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (24 + x0 + (64*x1)), xmask)
tmp19 = tl.load(in_ptr0 + (40 + x0 + (64*x1)), xmask)
tmp21 = tl.load(in_ptr0 + (56 + x0 + (64*x1)), xmask)
tmp23 = tl.load(in_ptr0 + (12 + x0 + (64*x1)), xmask)
tmp25 = tl.load(in_ptr0 + (28 + x0 + (64*x1)), xmask)
tmp27 = tl.load(in_ptr0 + (44 + x0 + (64*x1)), xmask)
tmp29 = tl.load(in_ptr0 + (60 + x0 + (64*x1)), xmask)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp8 = triton_helpers.maximum(tmp7, tmp6)
tmp10 = triton_helpers.maximum(tmp9, tmp8)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp20 = triton_helpers.maximum(tmp19, tmp18)
tmp22 = triton_helpers.maximum(tmp21, tmp20)
tmp24 = triton_helpers.maximum(tmp23, tmp22)
tmp26 = triton_helpers.maximum(tmp25, tmp24)
tmp28 = triton_helpers.maximum(tmp27, tmp26)
tmp30 = triton_helpers.maximum(tmp29, tmp28)
tl.store(out_ptr0 + (x2), tmp30, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_2], Original ATen: [aten.max_pool2d_with_indices]
stream0 = get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0.run(arg0_1, buf0, 16, grid=grid(16), stream=stream0)
del arg0_1
return (reinterpret_tensor(buf0, (4, 1, 1, 4), (4, 1, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
import torch.nn.functional as F
from torch.nn.modules.module import Module
from torch.nn.modules.utils import _pair
class my_MaxPool2d(Module):
def __init__(self, kernel_size, stride=None, padding=0, dilation=1,
return_indices=False, ceil_mode=False):
super(my_MaxPool2d, self).__init__()
self.kernel_size = kernel_size
self.stride = stride or kernel_size
self.padding = padding
self.dilation = dilation
self.return_indices = return_indices
self.ceil_mode = ceil_mode
def forward(self, input):
input = input.transpose(3, 1)
input = F.max_pool2d(input, self.kernel_size, self.stride, self.
padding, self.dilation, self.ceil_mode, self.return_indices)
input = input.transpose(3, 1).contiguous()
return input
def __repr__(self):
kh, kw = _pair(self.kernel_size)
dh, dw = _pair(self.stride)
padh, padw = _pair(self.padding)
dilh, dilw = _pair(self.dilation)
padding_str = ', padding=(' + str(padh) + ', ' + str(padw
) + ')' if padh != 0 or padw != 0 else ''
dilation_str = ', dilation=(' + str(dilh) + ', ' + str(dilw
) + ')' if dilh != 0 and dilw != 0 else ''
ceil_str = ', ceil_mode=' + str(self.ceil_mode)
return self.__class__.__name__ + '(' + 'kernel_size=(' + str(kh
) + ', ' + str(kw) + ')' + ', stride=(' + str(dh) + ', ' + str(dw
) + ')' + padding_str + dilation_str + ceil_str + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'kernel_size': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch.nn import Module
from torch.nn.modules.module import Module
from torch.nn.modules.utils import _pair
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp7 = tl.load(in_ptr0 + (4 + x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr0 + (20 + x0 + 64 * x1), xmask)
tmp11 = tl.load(in_ptr0 + (36 + x0 + 64 * x1), xmask)
tmp13 = tl.load(in_ptr0 + (52 + x0 + 64 * x1), xmask)
tmp15 = tl.load(in_ptr0 + (8 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (24 + x0 + 64 * x1), xmask)
tmp19 = tl.load(in_ptr0 + (40 + x0 + 64 * x1), xmask)
tmp21 = tl.load(in_ptr0 + (56 + x0 + 64 * x1), xmask)
tmp23 = tl.load(in_ptr0 + (12 + x0 + 64 * x1), xmask)
tmp25 = tl.load(in_ptr0 + (28 + x0 + 64 * x1), xmask)
tmp27 = tl.load(in_ptr0 + (44 + x0 + 64 * x1), xmask)
tmp29 = tl.load(in_ptr0 + (60 + x0 + 64 * x1), xmask)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp8 = triton_helpers.maximum(tmp7, tmp6)
tmp10 = triton_helpers.maximum(tmp9, tmp8)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp20 = triton_helpers.maximum(tmp19, tmp18)
tmp22 = triton_helpers.maximum(tmp21, tmp20)
tmp24 = triton_helpers.maximum(tmp23, tmp22)
tmp26 = triton_helpers.maximum(tmp25, tmp24)
tmp28 = triton_helpers.maximum(tmp27, tmp26)
tmp30 = triton_helpers.maximum(tmp29, tmp28)
tl.store(out_ptr0 + x2, tmp30, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0[grid(16)](arg0_1, buf0,
16, XBLOCK=16, num_warps=1, num_stages=1)
del arg0_1
return reinterpret_tensor(buf0, (4, 1, 1, 4), (4, 1, 1, 1), 0),
class my_MaxPool2dNew(Module):
def __init__(self, kernel_size, stride=None, padding=0, dilation=1,
return_indices=False, ceil_mode=False):
super(my_MaxPool2dNew, self).__init__()
self.kernel_size = kernel_size
self.stride = stride or kernel_size
self.padding = padding
self.dilation = dilation
self.return_indices = return_indices
self.ceil_mode = ceil_mode
def __repr__(self):
kh, kw = _pair(self.kernel_size)
dh, dw = _pair(self.stride)
padh, padw = _pair(self.padding)
dilh, dilw = _pair(self.dilation)
padding_str = ', padding=(' + str(padh) + ', ' + str(padw
) + ')' if padh != 0 or padw != 0 else ''
dilation_str = ', dilation=(' + str(dilh) + ', ' + str(dilw
) + ')' if dilh != 0 and dilw != 0 else ''
ceil_str = ', ceil_mode=' + str(self.ceil_mode)
return self.__class__.__name__ + '(' + 'kernel_size=(' + str(kh
) + ', ' + str(kw) + ')' + ', stride=(' + str(dh) + ', ' + str(dw
) + ')' + padding_str + dilation_str + ceil_str + ')'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
likun97/Low_quality_classification_with_mobilenetv3
|
my_MaxPool2d
| false | 10,439 |
[
"Apache-2.0"
] | 0 |
a9e6f66caad937fc7c8e101cddb76f116219b255
|
https://github.com/likun97/Low_quality_classification_with_mobilenetv3/tree/a9e6f66caad937fc7c8e101cddb76f116219b255
|
BaselineActor
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ix/cixxyusyg44s2hkoufcgbrv3ix5ookwqjl4ia3xkv7bdqi4yrzus.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = (xindex // 1600)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (1664*x3)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hj/chjzotk5iydxvuetxetlv36s7car7cdb24whkuqihxwcy5kkr4o2.py
# Topologically Sorted Source Nodes: [act], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# act => tanh
# Graph fragment:
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%view_5,), kwargs = {})
triton_poi_fused_tanh_1 = async_compile.triton('triton_poi_fused_tanh_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_tanh_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_tanh_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (400, 4), (4, 1))
assert_size_stride(primals_2, (400, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (400, 400), (400, 1))
assert_size_stride(primals_5, (400, ), (1, ))
assert_size_stride(primals_6, (4, 400), (400, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 400), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 400), (6400, 1600, 400, 1), 0); del buf0 # reuse
buf7 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf7, 25600, grid=grid(25600), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 400), (400, 1), 0), reinterpret_tensor(primals_4, (400, 400), (1, 400), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 400), (6400, 1600, 400, 1), 0); del buf2 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf6, 25600, grid=grid(25600), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf3, (64, 400), (400, 1), 0), reinterpret_tensor(primals_6, (400, 4), (1, 400), 0), out=buf4)
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [act], Original ATen: [aten.tanh]
triton_poi_fused_tanh_1.run(buf5, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
return (buf5, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 400), (400, 1), 0), reinterpret_tensor(buf3, (64, 400), (400, 1), 0), buf5, primals_6, buf6, primals_4, buf7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((400, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((400, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.functional import F
from torch.nn import functional as F
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
class BaselineActor(nn.Module):
def __init__(self, state_size, action_size, hidden_size=400):
super().__init__()
self.fc1 = nn.Linear(state_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, hidden_size)
self.out = nn.Linear(hidden_size, action_size)
def forward(self, state):
x = F.relu(self.fc1(state))
x = F.relu(self.fc2(x))
act = torch.tanh(self.out(x))
return act
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'state_size': 4, 'action_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = xindex // 1600
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 1664 * x3), tmp6, xmask)
@triton.jit
def triton_poi_fused_tanh_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (400, 4), (4, 1))
assert_size_stride(primals_2, (400,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (400, 400), (400, 1))
assert_size_stride(primals_5, (400,), (1,))
assert_size_stride(primals_6, (4, 400), (400, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 400), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 400), (6400, 1600, 400, 1), 0
)
del buf0
buf7 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(25600)](buf1,
primals_2, buf7, 25600, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 400), (400, 1), 0),
reinterpret_tensor(primals_4, (400, 400), (1, 400), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 400), (6400, 1600, 400, 1), 0
)
del buf2
buf6 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(25600)](buf3,
primals_5, buf6, 25600, XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf3, (64, 400), (400, 1), 0),
reinterpret_tensor(primals_6, (400, 4), (1, 400), 0), out=buf4)
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf4
triton_poi_fused_tanh_1[grid(256)](buf5, primals_7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_7
return buf5, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 400), (400, 1), 0
), reinterpret_tensor(buf3, (64, 400), (400, 1), 0
), buf5, primals_6, buf6, primals_4, buf7
class BaselineActorNew(nn.Module):
def __init__(self, state_size, action_size, hidden_size=400):
super().__init__()
self.fc1 = nn.Linear(state_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, hidden_size)
self.out = nn.Linear(hidden_size, action_size)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.out.weight
primals_7 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
greenstar1151/pytorch-benchmark
|
BaselineActor
| false | 10,440 |
[
"BSD-3-Clause"
] | 0 |
8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
https://github.com/greenstar1151/pytorch-benchmark/tree/8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
PMA
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/cm/ccmcgo4hhocf76otuns232vkfdobmiyhbrbzce7zxp7kc5eree6u.py
# Topologically Sorted Source Nodes: [repeat], Original ATen: [aten.repeat]
# Source node to ATen node mapping:
# repeat => repeat
# Graph fragment:
# %repeat : [num_users=2] = call_function[target=torch.ops.aten.repeat.default](args = (%primals_2, [4, 1, 1]), kwargs = {})
triton_poi_fused_repeat_0 = async_compile.triton('triton_poi_fused_repeat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_repeat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_repeat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2i/c2ifovqaljskkztxrqpysdk52llsxvhxf2zatdd7lqt2wm3hqaan.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + (16*y0)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tt/cttmvktt3m2x2nl56afa7l3abaxt7wlehowakdzngkhgs35f3n7u.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ry/cryn7ntc2gpkbfzbre3xh7lffx7zkbskw6oihbzsekkgajmdbki6.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_3 = async_compile.triton('triton_poi_fused__softmax_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sl/csl42w366z4dq6axbiuto4v6ryii3tmyrcbas65opz5pllxuww4m.py
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_3 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_4 = async_compile.triton('triton_poi_fused_clone_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/56/c566ex7xddxc2fqpwqlmymdyd23nesbsyghxftm7cy73ebnuo3ke.py
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# h_1 => add
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_21, %repeat), kwargs = {})
triton_poi_fused_add_5 = async_compile.triton('triton_poi_fused_add_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ae/caemgqme4td5miovnvxrouyha6d75h7jrf5txbdjnejoxy7cc7ym.py
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_2 => add_1
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_23,), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %relu), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_6 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_6(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [repeat], Original ATen: [aten.repeat]
stream0 = get_raw_stream(0)
triton_poi_fused_repeat_0.run(primals_2, buf0, 64, grid=grid(64), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf2)
del primals_5
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf3)
del primals_7
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf1, primals_4, buf4, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_4
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf2, primals_6, buf5, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_6
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
buf8 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf7, buf8, 256, grid=grid(256), stream=stream0)
del buf7
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [contiguous_2], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf3, primals_8, buf9, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_8
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [o], Original ATen: [aten.bmm]
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (16, 4, 1), (4, 1, 0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf10, buf11, 16, 4, grid=grid(16, 4), stream=stream0)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0); del buf10 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0); del buf12 # reuse
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
triton_poi_fused_add_5.run(buf13, primals_10, buf0, 64, grid=grid(64), stream=stream0)
del primals_10
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0), reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf14)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf16 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_6.run(buf13, buf14, primals_12, buf15, buf16, 64, grid=grid(64), stream=stream0)
del buf14
del primals_12
return (buf15, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(buf13, (16, 4), (4, 1), 0), buf16, primals_11, primals_9, reinterpret_tensor(buf9, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0), primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class PMA(nn.Module):
def __init__(self, d, n_heads, n_seeds, layer_norm=False):
super().__init__()
self.mab = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.S = nn.Parameter(torch.zeros(1, n_seeds, d), requires_grad=True)
with torch.no_grad():
nn.init.xavier_uniform_(self.S)
def forward(self, x, presence=None):
batch_size = x.shape[0]
return self.mab(self.S.repeat(batch_size, 1, 1), x, presence)
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'d': 4, 'n_heads': 4, 'n_seeds': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_repeat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + 16 * y0), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_6(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12
) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_repeat_0[grid(64)](primals_2, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf2)
del primals_5
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf3)
del primals_7
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_1[grid(4, 16)](buf1, primals_4, buf4, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_4
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_1[grid(4, 16)](buf2, primals_6, buf5, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_6
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf8 = buf6
del buf6
triton_poi_fused__softmax_3[grid(256)](buf7, buf8, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf7
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf2
triton_poi_fused_clone_1[grid(4, 16)](buf3, primals_8, buf9, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_8
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0)
del buf3
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (16, 4, 1), (4, 1,
0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf10, buf11, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0)
del buf10
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0)
del buf12
triton_poi_fused_add_5[grid(64)](buf13, primals_10, buf0, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_10
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf14)
buf15 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf16 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_6[grid(64)](buf13,
buf14, primals_12, buf15, buf16, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf14
del primals_12
return buf15, reinterpret_tensor(buf0, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0
), reinterpret_tensor(buf13, (16, 4), (4, 1), 0
), buf16, primals_11, primals_9, reinterpret_tensor(buf9, (16, 1, 4
), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0), primals_3
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class PMANew(nn.Module):
def __init__(self, d, n_heads, n_seeds, layer_norm=False):
super().__init__()
self.mab = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.S = nn.Parameter(torch.zeros(1, n_seeds, d), requires_grad=True)
with torch.no_grad():
nn.init.xavier_uniform_(self.S)
def forward(self, input_0):
primals_2 = self.S
primals_3 = self.mab.mqkv.q_projector.weight
primals_4 = self.mab.mqkv.q_projector.bias
primals_5 = self.mab.mqkv.k_projector.weight
primals_6 = self.mab.mqkv.k_projector.bias
primals_7 = self.mab.mqkv.v_projector.weight
primals_8 = self.mab.mqkv.v_projector.bias
primals_9 = self.mab.mqkv.o_projector.weight
primals_10 = self.mab.mqkv.o_projector.bias
primals_11 = self.mab.fc.weight
primals_12 = self.mab.fc.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12])
return output[0]
|
karayanni/torch-scae
|
PMA
| false | 10,441 |
[
"Apache-2.0"
] | 0 |
e044662d8942d8d1923d13d071f375144cf4a1e8
|
https://github.com/karayanni/torch-scae/tree/e044662d8942d8d1923d13d071f375144cf4a1e8
|
ISAB
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/cm/ccmcgo4hhocf76otuns232vkfdobmiyhbrbzce7zxp7kc5eree6u.py
# Topologically Sorted Source Nodes: [repeat], Original ATen: [aten.repeat]
# Source node to ATen node mapping:
# repeat => repeat
# Graph fragment:
# %repeat : [num_users=2] = call_function[target=torch.ops.aten.repeat.default](args = (%primals_2, [4, 1, 1]), kwargs = {})
triton_poi_fused_repeat_0 = async_compile.triton('triton_poi_fused_repeat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_repeat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_repeat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2i/c2ifovqaljskkztxrqpysdk52llsxvhxf2zatdd7lqt2wm3hqaan.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + (16*y0)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tt/cttmvktt3m2x2nl56afa7l3abaxt7wlehowakdzngkhgs35f3n7u.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%bmm, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%bmm, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ry/cryn7ntc2gpkbfzbre3xh7lffx7zkbskw6oihbzsekkgajmdbki6.py
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# routing_1 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_3 = async_compile.triton('triton_poi_fused__softmax_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sl/csl42w366z4dq6axbiuto4v6ryii3tmyrcbas65opz5pllxuww4m.py
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_3 => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_4 = async_compile.triton('triton_poi_fused_clone_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/56/c566ex7xddxc2fqpwqlmymdyd23nesbsyghxftm7cy73ebnuo3ke.py
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# h_1 => add
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_21, %repeat), kwargs = {})
triton_poi_fused_add_5 = async_compile.triton('triton_poi_fused_add_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ae/caemgqme4td5miovnvxrouyha6d75h7jrf5txbdjnejoxy7cc7ym.py
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_2 => add_1
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_23,), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %relu), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_6 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_6(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4, ), (1, ))
assert_size_stride(primals_13, (4, 4), (4, 1))
assert_size_stride(primals_14, (4, ), (1, ))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4, ), (1, ))
assert_size_stride(primals_17, (4, 4), (4, 1))
assert_size_stride(primals_18, (4, ), (1, ))
assert_size_stride(primals_19, (4, 4), (4, 1))
assert_size_stride(primals_20, (4, ), (1, ))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [repeat], Original ATen: [aten.repeat]
stream0 = get_raw_stream(0)
triton_poi_fused_repeat_0.run(primals_2, buf0, 64, grid=grid(64), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf2)
del primals_5
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf3)
del primals_7
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf1, primals_4, buf4, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_4
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf2, primals_6, buf5, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_6
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
buf8 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [routing_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf7, buf8, 256, grid=grid(256), stream=stream0)
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [contiguous_2], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf3, primals_8, buf9, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_8
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [o], Original ATen: [aten.bmm]
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (16, 4, 1), (4, 1, 0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_3], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf10, buf11, 16, 4, grid=grid(16, 4), stream=stream0)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0); del buf10 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0); del buf12 # reuse
# Topologically Sorted Source Nodes: [h_1], Original ATen: [aten.add]
triton_poi_fused_add_5.run(buf13, primals_10, buf0, 64, grid=grid(64), stream=stream0)
del primals_10
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0), reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf14)
buf15 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_13, (4, 4), (1, 4), 0), out=buf15)
del primals_13
buf16 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf32 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_6.run(buf13, buf14, primals_12, buf16, buf32, 64, grid=grid(64), stream=stream0)
del primals_12
buf17 = buf14; del buf14 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf16, (16, 4), (4, 1), 0), reinterpret_tensor(primals_15, (4, 4), (1, 4), 0), out=buf17)
buf18 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf16, (16, 4), (4, 1), 0), reinterpret_tensor(primals_17, (4, 4), (1, 4), 0), out=buf18)
buf19 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_4], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf15, primals_14, buf19, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_14
buf20 = reinterpret_tensor(buf15, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf15 # reuse
# Topologically Sorted Source Nodes: [contiguous_5], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf17, primals_16, buf20, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_16
buf21 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [routing_2], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf19, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf20, (16, 1, 4), (4, 0, 1), 0), out=buf21)
buf22 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [routing_3], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf21, buf22, 256, grid=grid(256), stream=stream0)
buf23 = buf21; del buf21 # reuse
# Topologically Sorted Source Nodes: [routing_3], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf22, buf23, 256, grid=grid(256), stream=stream0)
del buf22
buf24 = reinterpret_tensor(buf17, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf17 # reuse
# Topologically Sorted Source Nodes: [contiguous_6], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf18, primals_18, buf24, 4, 16, grid=grid(4, 16), stream=stream0)
del primals_18
buf25 = reinterpret_tensor(buf18, (16, 4, 1), (4, 1, 1), 0); del buf18 # reuse
# Topologically Sorted Source Nodes: [o_2], Original ATen: [aten.bmm]
extern_kernels.bmm(buf23, reinterpret_tensor(buf24, (16, 4, 1), (4, 1, 0), 0), out=buf25)
buf26 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_7], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf25, buf26, 16, 4, grid=grid(16, 4), stream=stream0)
buf27 = reinterpret_tensor(buf25, (16, 4), (4, 1), 0); del buf25 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf26, (16, 4), (4, 1), 0), reinterpret_tensor(primals_19, (4, 4), (1, 4), 0), out=buf27)
buf28 = reinterpret_tensor(buf27, (4, 4, 4), (16, 4, 1), 0); del buf27 # reuse
# Topologically Sorted Source Nodes: [h_4], Original ATen: [aten.add]
triton_poi_fused_add_5.run(buf28, primals_20, primals_1, 64, grid=grid(64), stream=stream0)
del primals_20
buf29 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf28, (16, 4), (4, 1), 0), reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), out=buf29)
buf30 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf31 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu_1, h_5], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_6.run(buf28, buf29, primals_22, buf30, buf31, 64, grid=grid(64), stream=stream0)
del buf29
del primals_22
return (buf30, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(buf13, (16, 4), (4, 1), 0), reinterpret_tensor(buf16, (16, 4), (4, 1), 0), buf23, reinterpret_tensor(buf26, (16, 4), (4, 1), 0), reinterpret_tensor(buf28, (16, 4), (4, 1), 0), buf31, primals_21, primals_19, reinterpret_tensor(buf24, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf19, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf20, (16, 4, 1), (4, 1, 1), 0), primals_17, primals_15, buf32, primals_11, primals_9, reinterpret_tensor(buf9, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0), primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class ISAB(nn.Module):
def __init__(self, d, n_heads, n_inducing_points, layer_norm=False):
super().__init__()
self.mab0 = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.mab1 = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.I = nn.Parameter(torch.zeros(1, n_inducing_points, d),
requires_grad=True)
with torch.no_grad():
nn.init.xavier_uniform_(self.I)
def forward(self, x, presence=None):
batch_size = x.shape[0]
h = self.mab0(self.I.repeat(batch_size, 1, 1), x, presence)
return self.mab1(x, h)
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'d': 4, 'n_heads': 4, 'n_inducing_points': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_repeat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x1 + 16 * y0), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_6(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4,), (1,))
assert_size_stride(primals_13, (4, 4), (4, 1))
assert_size_stride(primals_14, (4,), (1,))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4,), (1,))
assert_size_stride(primals_17, (4, 4), (4, 1))
assert_size_stride(primals_18, (4,), (1,))
assert_size_stride(primals_19, (4, 4), (4, 1))
assert_size_stride(primals_20, (4,), (1,))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_repeat_0[grid(64)](primals_2, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf2)
del primals_5
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf3)
del primals_7
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_1[grid(4, 16)](buf1, primals_4, buf4, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_4
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_1[grid(4, 16)](buf2, primals_6, buf5, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_6
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf8 = buf6
del buf6
triton_poi_fused__softmax_3[grid(256)](buf7, buf8, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf2
triton_poi_fused_clone_1[grid(4, 16)](buf3, primals_8, buf9, 4, 16,
XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_8
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0)
del buf3
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (16, 4, 1), (4, 1,
0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf10, buf11, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0)
del buf10
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0)
del buf12
triton_poi_fused_add_5[grid(64)](buf13, primals_10, buf0, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_10
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), out=buf14)
buf15 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_13, (4, 4), (1, 4), 0), out=buf15)
del primals_13
buf16 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf32 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_6[grid(64)](buf13,
buf14, primals_12, buf16, buf32, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del primals_12
buf17 = buf14
del buf14
extern_kernels.mm(reinterpret_tensor(buf16, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_15, (4, 4), (1, 4), 0), out=buf17)
buf18 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf16, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_17, (4, 4), (1, 4), 0), out=buf18)
buf19 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_1[grid(4, 16)](buf15, primals_14, buf19, 4,
16, XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_14
buf20 = reinterpret_tensor(buf15, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf15
triton_poi_fused_clone_1[grid(4, 16)](buf17, primals_16, buf20, 4,
16, XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_16
buf21 = buf7
del buf7
extern_kernels.bmm(reinterpret_tensor(buf19, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf20, (16, 1, 4), (4, 0, 1), 0), out=buf21)
buf22 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(256)](buf21, buf22, 256, XBLOCK=
256, num_warps=4, num_stages=1)
buf23 = buf21
del buf21
triton_poi_fused__softmax_3[grid(256)](buf22, buf23, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del buf22
buf24 = reinterpret_tensor(buf17, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf17
triton_poi_fused_clone_1[grid(4, 16)](buf18, primals_18, buf24, 4,
16, XBLOCK=16, YBLOCK=2, num_warps=1, num_stages=1)
del primals_18
buf25 = reinterpret_tensor(buf18, (16, 4, 1), (4, 1, 1), 0)
del buf18
extern_kernels.bmm(buf23, reinterpret_tensor(buf24, (16, 4, 1), (4,
1, 0), 0), out=buf25)
buf26 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf25, buf26, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf27 = reinterpret_tensor(buf25, (16, 4), (4, 1), 0)
del buf25
extern_kernels.mm(reinterpret_tensor(buf26, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_19, (4, 4), (1, 4), 0), out=buf27)
buf28 = reinterpret_tensor(buf27, (4, 4, 4), (16, 4, 1), 0)
del buf27
triton_poi_fused_add_5[grid(64)](buf28, primals_20, primals_1, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_20
buf29 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf28, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), out=buf29)
buf30 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf31 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_add_relu_threshold_backward_6[grid(64)](buf28,
buf29, primals_22, buf30, buf31, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf29
del primals_22
return (buf30, reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), buf8,
reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(
buf13, (16, 4), (4, 1), 0), reinterpret_tensor(buf16, (16, 4), (4,
1), 0), buf23, reinterpret_tensor(buf26, (16, 4), (4, 1), 0),
reinterpret_tensor(buf28, (16, 4), (4, 1), 0), buf31, primals_21,
primals_19, reinterpret_tensor(buf24, (16, 1, 4), (4, 1, 1), 0),
reinterpret_tensor(buf19, (16, 1, 4), (4, 1, 1), 0),
reinterpret_tensor(buf20, (16, 4, 1), (4, 1, 1), 0), primals_17,
primals_15, buf32, primals_11, primals_9, reinterpret_tensor(buf9,
(16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 1, 4), (4,
1, 1), 0), reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0),
primals_3)
def qkv_attention(queries, keys, values, presence=None):
"""
Transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
d_k = queries.shape[-1]
routing = torch.matmul(queries, keys.transpose(1, 2))
if presence is not None:
routing -= (1.0 - presence.unsqueeze(-2)) * 1e+32
routing = F.softmax(routing / np.sqrt(d_k), -1)
return torch.matmul(routing, values)
class MultiHeadQKVAttention(nn.Module):
"""Multi-head version of Transformer-like attention."""
def __init__(self, d_k, d_v, n_heads):
super().__init__()
self.d_k = d_k
self.d_v = d_v
self.n_heads = n_heads
d_k_p = int(math.ceil(d_k / n_heads)) * n_heads
d_v_p = int(math.ceil(d_v / n_heads)) * n_heads
self.q_projector = nn.Linear(d_k, d_k_p)
self.k_projector = nn.Linear(d_k, d_k_p)
self.v_projector = nn.Linear(d_v, d_v_p)
self.o_projector = nn.Linear(d_v_p, d_v)
def forward(self, queries, keys, values, presence=None):
"""
Multi-head transformer-like self-attention.
Args:
queries: Tensor of shape [B, N, d_k].
keys: Tensor of shape [B, M, d_k].
values: : Tensor of shape [B, M, d_v].
presence: None or tensor of shape [B, M].
Returns:
Tensor of shape [B, N, d_v]
"""
assert queries.shape[2] == keys.shape[2]
assert keys.shape[1] == values.shape[1]
if presence is not None:
assert values.shape[:2] == presence.shape
B, N, _d_k = queries.shape
M, _d_v = values.shape[1:]
H = self.n_heads
q_p = self.q_projector(queries)
k_p = self.k_projector(keys)
v_p = self.v_projector(values)
del queries, keys, values
q = q_p.view(B, N, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, N, -1)
k = k_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
v = v_p.view(B, M, H, -1).permute(2, 0, 1, 3).contiguous().view(H *
B, M, -1)
if presence is not None:
presence = presence.repeat(self.n_heads, 1)
o = qkv_attention(q, k, v, presence)
o = o.view(H, B, N, -1).permute(1, 2, 0, 3).contiguous().view(B, N, -1)
return self.o_projector(o)
class MAB(nn.Module):
def __init__(self, d, n_heads, layer_norm=False):
super().__init__()
self.layer_norm = layer_norm
self.mqkv = MultiHeadQKVAttention(d_k=d, d_v=d, n_heads=n_heads)
if layer_norm:
self.ln0 = nn.LayerNorm(d)
self.ln1 = nn.LayerNorm(d)
self.fc = nn.Linear(d, d)
def forward(self, queries, keys, presence=None):
h = self.mqkv(queries, keys, keys, presence)
h = h + queries
if presence is not None:
assert presence.shape[1] == queries.shape[1] == keys.shape[1]
h = h * presence.unsqueeze(-1)
if self.layer_norm:
h = self.ln0(h)
h = h + F.relu(self.fc(h))
if self.layer_norm:
h = self.ln1(h)
return h
class ISABNew(nn.Module):
def __init__(self, d, n_heads, n_inducing_points, layer_norm=False):
super().__init__()
self.mab0 = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.mab1 = MAB(d=d, n_heads=n_heads, layer_norm=layer_norm)
self.I = nn.Parameter(torch.zeros(1, n_inducing_points, d),
requires_grad=True)
with torch.no_grad():
nn.init.xavier_uniform_(self.I)
def forward(self, input_0):
primals_2 = self.I
primals_3 = self.mab0.mqkv.q_projector.weight
primals_4 = self.mab0.mqkv.q_projector.bias
primals_5 = self.mab0.mqkv.k_projector.weight
primals_6 = self.mab0.mqkv.k_projector.bias
primals_7 = self.mab0.mqkv.v_projector.weight
primals_8 = self.mab0.mqkv.v_projector.bias
primals_9 = self.mab0.mqkv.o_projector.weight
primals_10 = self.mab0.mqkv.o_projector.bias
primals_11 = self.mab0.fc.weight
primals_12 = self.mab0.fc.bias
primals_13 = self.mab1.mqkv.q_projector.weight
primals_14 = self.mab1.mqkv.q_projector.bias
primals_15 = self.mab1.mqkv.k_projector.weight
primals_16 = self.mab1.mqkv.k_projector.bias
primals_17 = self.mab1.mqkv.v_projector.weight
primals_18 = self.mab1.mqkv.v_projector.bias
primals_19 = self.mab1.mqkv.o_projector.weight
primals_20 = self.mab1.mqkv.o_projector.bias
primals_21 = self.mab1.fc.weight
primals_22 = self.mab1.fc.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22])
return output[0]
|
karayanni/torch-scae
|
ISAB
| false | 10,442 |
[
"Apache-2.0"
] | 0 |
e044662d8942d8d1923d13d071f375144cf4a1e8
|
https://github.com/karayanni/torch-scae/tree/e044662d8942d8d1923d13d071f375144cf4a1e8
|
adder2d
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/zr/czriihulllg44xhaffhdwvpylq55lkftttdpdjcdqfcu3eujm5rf.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_1,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (64*x1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x1 + (4*y0)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/44/c44i7vnizkohfraswy2fybvcbu4j2tl54qnzscskv4lizxkoye7f.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.sub, aten.abs, aten.sum]
# Source node to ATen node mapping:
# out => abs_1, sub, sum_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%unsqueeze_6, %unsqueeze_7), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%abs_1, [1]), kwargs = {})
triton_per_fused_abs_sub_sum_1 = async_compile.triton('triton_per_fused_abs_sub_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 64],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_abs_sub_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_abs_sub_sum_1(in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r2 = rindex
x1 = (xindex // 4)
x0 = xindex % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + (r2 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (x0 + (4*r2)), xmask, eviction_policy='evict_last', other=0.0)
tmp2 = tmp0 - tmp1
tmp3 = tl_math.abs(tmp2)
tmp4 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tl.store(out_ptr0 + (x3), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/eh/ceh3drpczc5qvlpgea6jmkqhnvpfcv3u5sykd2ulnebudkpn4a6l.py
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# out_2 => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_2 = async_compile.triton('triton_poi_fused_clone_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x1)), xmask & ymask)
tmp1 = -tmp0
tl.store(out_ptr0 + (x1 + (4*y0)), tmp1, xmask & ymask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 1, 4), (4, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_2, buf0, 64, 4, grid=grid(64, 4), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.sub, aten.abs, aten.sum]
triton_per_fused_abs_sub_sum_1.run(primals_1, buf0, buf1, 16, 64, grid=grid(16), stream=stream0)
buf2 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.clone]
triton_poi_fused_clone_2.run(buf1, buf2, 4, 4, grid=grid(4, 4), stream=stream0)
del buf1
return (buf2, primals_1, reinterpret_tensor(buf0, (1, 64, 4), (256, 4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.autograd import Function
import math
import torch
import torch.nn as nn
def adder2d_function(X, W, stride=1, padding=0):
n_filters, _d_filter, h_filter, w_filter = W.size()
n_x, _d_x, h_x, w_x = X.size()
h_out = (h_x - h_filter + 2 * padding) / stride + 1
w_out = (w_x - w_filter + 2 * padding) / stride + 1
h_out, w_out = int(h_out), int(w_out)
X_col = torch.nn.functional.unfold(X.view(1, -1, h_x, w_x), h_filter,
dilation=1, padding=padding, stride=stride).view(n_x, -1, h_out * w_out
)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W.view(n_filters, -1)
out = adder.apply(W_col, X_col)
out = out.view(n_filters, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
return out
class adder(Function):
@staticmethod
def forward(ctx, W_col, X_col):
ctx.save_for_backward(W_col, X_col)
output = -(W_col.unsqueeze(2) - X_col.unsqueeze(0)).abs().sum(1)
return output
@staticmethod
def backward(ctx, grad_output):
W_col, X_col = ctx.saved_tensors
grad_W_col = ((X_col.unsqueeze(0) - W_col.unsqueeze(2)) *
grad_output.unsqueeze(1)).sum(2)
grad_W_col = grad_W_col / grad_W_col.norm(p=2).clamp(min=1e-12
) * math.sqrt(W_col.size(1) * W_col.size(0)) / 5
grad_X_col = (-(X_col.unsqueeze(0) - W_col.unsqueeze(2)).clamp(-1,
1) * grad_output.unsqueeze(1)).sum(0)
return grad_W_col, grad_X_col
class adder2d(nn.Module):
def __init__(self, input_channel, output_channel, kernel_size, stride=1,
padding=0, bias=False):
super(adder2d, self).__init__()
self.stride = stride
self.padding = padding
self.input_channel = input_channel
self.output_channel = output_channel
self.kernel_size = kernel_size
self.adder = torch.nn.Parameter(nn.init.normal_(torch.randn(
output_channel, input_channel, kernel_size, kernel_size)))
self.bias = bias
if bias:
self.b = torch.nn.Parameter(nn.init.uniform_(torch.zeros(
output_channel)))
def forward(self, x):
output = adder2d_function(x, self.adder, self.stride, self.padding)
if self.bias:
output += self.b.unsqueeze(0).unsqueeze(2).unsqueeze(3)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_channel': 4, 'output_channel': 4, 'kernel_size': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.autograd import Function
import math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 64 * x1), xmask & ymask, eviction_policy
='evict_last')
tl.store(out_ptr0 + (x1 + 4 * y0), tmp0, xmask & ymask)
@triton.jit
def triton_per_fused_abs_sub_sum_1(in_ptr0, in_ptr1, out_ptr0, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r2 = rindex
x1 = xindex // 4
x0 = xindex % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + (r2 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (x0 + 4 * r2), xmask, eviction_policy=
'evict_last', other=0.0)
tmp2 = tmp0 - tmp1
tmp3 = tl_math.abs(tmp2)
tmp4 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tl.store(out_ptr0 + x3, tmp7, xmask)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 4
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x1), xmask & ymask)
tmp1 = -tmp0
tl.store(out_ptr0 + (x1 + 4 * y0), tmp1, xmask & ymask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 1, 4), (4, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(64, 4)](primals_2, buf0, 64, 4,
XBLOCK=4, YBLOCK=32, num_warps=4, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_per_fused_abs_sub_sum_1[grid(16)](primals_1, buf0, buf1, 16,
64, XBLOCK=8, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
triton_poi_fused_clone_2[grid(4, 4)](buf1, buf2, 4, 4, XBLOCK=4,
YBLOCK=4, num_warps=1, num_stages=1)
del buf1
return buf2, primals_1, reinterpret_tensor(buf0, (1, 64, 4), (256, 4, 1), 0
)
def adder2d_function(X, W, stride=1, padding=0):
n_filters, _d_filter, h_filter, w_filter = W.size()
n_x, _d_x, h_x, w_x = X.size()
h_out = (h_x - h_filter + 2 * padding) / stride + 1
w_out = (w_x - w_filter + 2 * padding) / stride + 1
h_out, w_out = int(h_out), int(w_out)
X_col = torch.nn.functional.unfold(X.view(1, -1, h_x, w_x), h_filter,
dilation=1, padding=padding, stride=stride).view(n_x, -1, h_out * w_out
)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W.view(n_filters, -1)
out = adder.apply(W_col, X_col)
out = out.view(n_filters, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
return out
class adder(Function):
@staticmethod
def forward(ctx, W_col, X_col):
ctx.save_for_backward(W_col, X_col)
output = -(W_col.unsqueeze(2) - X_col.unsqueeze(0)).abs().sum(1)
return output
@staticmethod
def backward(ctx, grad_output):
W_col, X_col = ctx.saved_tensors
grad_W_col = ((X_col.unsqueeze(0) - W_col.unsqueeze(2)) *
grad_output.unsqueeze(1)).sum(2)
grad_W_col = grad_W_col / grad_W_col.norm(p=2).clamp(min=1e-12
) * math.sqrt(W_col.size(1) * W_col.size(0)) / 5
grad_X_col = (-(X_col.unsqueeze(0) - W_col.unsqueeze(2)).clamp(-1,
1) * grad_output.unsqueeze(1)).sum(0)
return grad_W_col, grad_X_col
class adder2dNew(nn.Module):
def __init__(self, input_channel, output_channel, kernel_size, stride=1,
padding=0, bias=False):
super(adder2dNew, self).__init__()
self.stride = stride
self.padding = padding
self.input_channel = input_channel
self.output_channel = output_channel
self.kernel_size = kernel_size
self.adder = torch.nn.Parameter(nn.init.normal_(torch.randn(
output_channel, input_channel, kernel_size, kernel_size)))
self.bias = bias
if bias:
self.b = torch.nn.Parameter(nn.init.uniform_(torch.zeros(
output_channel)))
def forward(self, input_0):
primals_1 = self.adder
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
|
mark531593296/AdderNet
|
adder2d
| false | 10,443 |
[
"BSD-3-Clause"
] | 0 |
2936728f537c0cceb8a47727630e5723af86df61
|
https://github.com/mark531593296/AdderNet/tree/2936728f537c0cceb8a47727630e5723af86df61
|
BaselineDiscreteCritic
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/fk/cfkpmzrz5fsihotvtb2iptrxsxsj2pu6jx4m3j5xhm4ptz5cd42c.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = (xindex // 1200)
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + (1216*x2)), tmp4, xmask)
tl.store(out_ptr1 + (x3 + (1280*x2)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6c/c6cbnt5xpzncru5uqhj6zzm2rgjtlv6ylmvnad4rfkm4h2d5ni5s.py
# Topologically Sorted Source Nodes: [x, linear_1], Original ATen: [aten.relu, aten.view]
# Source node to ATen node mapping:
# linear_1 => view_2
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %view_2 : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%relu, [64, 300]), kwargs = {})
triton_poi_fused_relu_view_1 = async_compile.triton('triton_poi_fused_relu_view_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_view_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_view_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = (xindex // 300)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (300*(x1 % 4)) + (1216*(x1 // 4))), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (300, 4), (4, 1))
assert_size_stride(primals_2, (300, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (300, 300), (300, 1))
assert_size_stride(primals_5, (300, ), (1, ))
assert_size_stride(primals_6, (4, 300), (300, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 300), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1), torch.float32)
buf8 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf0, primals_2, buf1, buf8, 19200, grid=grid(19200), stream=stream0)
del primals_2
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x, linear_1], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_1.run(buf1, buf2, 19200, grid=grid(19200), stream=stream0)
buf3 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf2, reinterpret_tensor(primals_4, (300, 300), (1, 300), 0), out=buf3)
buf4 = buf1; del buf1 # reuse
buf7 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf4, buf7, 19200, grid=grid(19200), stream=stream0)
del primals_5
buf5 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [x_1, vals], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_1.run(buf4, buf5, 19200, grid=grid(19200), stream=stream0)
del buf4
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [vals], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf5, reinterpret_tensor(primals_6, (300, 4), (1, 300), 0), alpha=1, beta=1, out=buf6)
del primals_7
return (reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf2, buf5, primals_6, buf7, primals_4, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((300, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((300, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.functional import F
from torch.nn import functional as F
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
class BaselineDiscreteCritic(nn.Module):
def __init__(self, obs_shape, action_shape, hidden_size=300):
super().__init__()
self.fc1 = nn.Linear(obs_shape, hidden_size)
self.fc2 = nn.Linear(hidden_size, hidden_size)
self.out = nn.Linear(hidden_size, action_shape)
def forward(self, state):
x = F.relu(self.fc1(state))
x = F.relu(self.fc2(x))
vals = self.out(x)
return vals
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'obs_shape': 4, 'action_shape': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = xindex // 1200
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + 1216 * x2), tmp4, xmask)
tl.store(out_ptr1 + (x3 + 1280 * x2), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_view_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = xindex // 300
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 300 * (x1 % 4) + 1216 * (x1 // 4)), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (300, 4), (4, 1))
assert_size_stride(primals_2, (300,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (300, 300), (300, 1))
assert_size_stride(primals_5, (300,), (1,))
assert_size_stride(primals_6, (4, 300), (300, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 300), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1),
torch.float32)
buf8 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(19200)](buf0,
primals_2, buf1, buf8, 19200, XBLOCK=256, num_warps=4, num_stages=1
)
del primals_2
buf2 = buf0
del buf0
triton_poi_fused_relu_view_1[grid(19200)](buf1, buf2, 19200, XBLOCK
=256, num_warps=4, num_stages=1)
buf3 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
extern_kernels.mm(buf2, reinterpret_tensor(primals_4, (300, 300), (
1, 300), 0), out=buf3)
buf4 = buf1
del buf1
buf7 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(19200)](buf3,
primals_5, buf4, buf7, 19200, XBLOCK=256, num_warps=4, num_stages=1
)
del primals_5
buf5 = buf3
del buf3
triton_poi_fused_relu_view_1[grid(19200)](buf4, buf5, 19200, XBLOCK
=256, num_warps=4, num_stages=1)
del buf4
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, buf5, reinterpret_tensor(primals_6,
(300, 4), (1, 300), 0), alpha=1, beta=1, out=buf6)
del primals_7
return reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf2, buf5, primals_6, buf7, primals_4, buf8
class BaselineDiscreteCriticNew(nn.Module):
def __init__(self, obs_shape, action_shape, hidden_size=300):
super().__init__()
self.fc1 = nn.Linear(obs_shape, hidden_size)
self.fc2 = nn.Linear(hidden_size, hidden_size)
self.out = nn.Linear(hidden_size, action_shape)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.out.weight
primals_7 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
greenstar1151/pytorch-benchmark
|
BaselineDiscreteCritic
| false | 10,444 |
[
"BSD-3-Clause"
] | 0 |
8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
https://github.com/greenstar1151/pytorch-benchmark/tree/8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
TransposeGatedConv2d
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/oj/cojl5mb3pzv5jbmfzjkbac5hekbmpvb72kof6ouyyasitrogdd6n.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._unsafe_index]
# Source node to ATen node mapping:
# x => _unsafe_index
# Graph fragment:
# %_unsafe_index : [num_users=3] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %unsqueeze, %convert_element_type_1]), kwargs = {})
triton_poi_fused__unsafe_index_0 = async_compile.triton('triton_poi_fused__unsafe_index_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_index_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_index_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 8) % 8
x0 = xindex % 8
x2 = (xindex // 64)
x4 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = x0
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + (4*tmp4) + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x4), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ya/cya2grnbhraytq2wzrkx5sd2ottwnbrnd5ohd2xstcxyryneuc25.py
# Topologically Sorted Source Nodes: [mv, norm, add, truediv], Original ATen: [aten.mv, aten.linalg_vector_norm, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add_4
# mv => mul_4, sum_1
# norm => pow_1, pow_2, sum_2
# truediv => div
# Graph fragment:
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%permute, %primals_2), kwargs = {})
# %sum_1 : [num_users=2] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_4, [1]), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, None), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_2, 0.5), kwargs = {})
# %add_4 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_2, 1e-12), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%sum_1, %add_4), kwargs = {})
triton_per_fused_add_div_linalg_vector_norm_mv_1 = async_compile.triton('triton_per_fused_add_div_linalg_vector_norm_mv_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 6), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_linalg_vector_norm_mv_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_linalg_vector_norm_mv_1(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.load(in_ptr0 + (64 + r0), None)
tmp5 = tl.load(in_ptr1 + (1))
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp9 = tl.load(in_ptr0 + (128 + r0), None)
tmp10 = tl.load(in_ptr1 + (2))
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tmp14 = tl.load(in_ptr0 + (192 + r0), None)
tmp15 = tl.load(in_ptr1 + (3))
tmp16 = tl.broadcast_to(tmp15, [XBLOCK, RBLOCK])
tmp3 = tmp0 * tmp2
tmp7 = tmp4 * tmp6
tmp8 = tmp3 + tmp7
tmp12 = tmp9 * tmp11
tmp13 = tmp8 + tmp12
tmp17 = tmp14 * tmp16
tmp18 = tmp13 + tmp17
tmp19 = tmp18 * tmp18
tmp20 = tl.broadcast_to(tmp19, [XBLOCK, RBLOCK])
tmp22 = tl.sum(tmp20, 1)[:, None]
tmp23 = libdevice.sqrt(tmp22)
tmp24 = 1e-12
tmp25 = tmp23 + tmp24
tmp26 = tmp18 / tmp25
tl.store(out_ptr0 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp18, None)
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp25, None)
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp26, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/qi/cqiozgecuvqtnurxrggbllqpuci3n65ycew5qi5gdqg44ypxzegy.py
# Topologically Sorted Source Nodes: [truediv, mv_1], Original ATen: [aten.div, aten.mv]
# Source node to ATen node mapping:
# mv_1 => mul_5, sum_3
# truediv => div
# Graph fragment:
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%sum_1, %add_4), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %div), kwargs = {})
# %sum_3 : [num_users=3] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_5, [1]), kwargs = {})
triton_per_fused_div_mv_2 = async_compile.triton('triton_per_fused_div_mv_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_mv_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_mv_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (r1), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp4 = tmp1 / tmp3
tmp5 = tmp0 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/qa/cqaed4ios3xqwlv4d3cciikkdz7d73vhwkegurd5cxca3y7htmvg.py
# Topologically Sorted Source Nodes: [norm_1, add_1, truediv_1], Original ATen: [aten.linalg_vector_norm, aten.add, aten.div]
# Source node to ATen node mapping:
# add_1 => add_5
# norm_1 => pow_3, pow_4, sum_4
# truediv_1 => div_1
# Graph fragment:
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_3, 2), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_3, None), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_4, 0.5), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_4, 1e-12), kwargs = {})
# %div_1 : [num_users=3] = call_function[target=torch.ops.aten.div.Tensor](args = (%sum_3, %add_5), kwargs = {})
triton_per_fused_add_div_linalg_vector_norm_3 = async_compile.triton('triton_per_fused_add_div_linalg_vector_norm_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_linalg_vector_norm_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_linalg_vector_norm_3(in_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp5 = libdevice.sqrt(tmp4)
tmp6 = 1e-12
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp8, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/c2/cc2arficwjs4sforhl25gdfmb3uzfg7hkw46gq3mxgv57jy52z32.py
# Topologically Sorted Source Nodes: [sigma], Original ATen: [aten.dot]
# Source node to ATen node mapping:
# sigma => mul_7, sum_6
# Graph fragment:
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %sum_3), kwargs = {})
# %sum_6 : [num_users=2] = call_function[target=torch.ops.aten.sum.default](args = (%mul_7,), kwargs = {})
triton_per_fused_dot_4 = async_compile.triton('triton_per_fused_dot_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_dot_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_dot_4(in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr1 + (r0), None)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/kw/ckwzptlssdpmtxi6pt23ik63xcuqar2giaakuqtgizxlg5weagc7.py
# Topologically Sorted Source Nodes: [truediv_2], Original ATen: [aten.div]
# Source node to ATen node mapping:
# truediv_2 => div_2
# Graph fragment:
# %div_2 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_4, %expand), kwargs = {})
triton_poi_fused_div_5 = async_compile.triton('triton_poi_fused_div_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_5(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 / tmp2
tl.store(out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2w/c2wlnlirhh2nibaimsmrfiriqyr7m3r6ij6r2vrxypktuy5hni2x.py
# Topologically Sorted Source Nodes: [conv, mask, gated_mask, conv_1, x_2], Original ATen: [aten.convolution, aten.sigmoid, aten.leaky_relu, aten.mul]
# Source node to ATen node mapping:
# conv => convolution
# conv_1 => gt, mul_12, where
# gated_mask => sigmoid
# mask => convolution_1
# x_2 => mul_13
# Graph fragment:
# %convolution : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index, %div_2, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index, %div_5, %primals_9, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_1,), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul_12 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 0.2), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %convolution, %mul_12), kwargs = {})
# %mul_13 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%where, %sigmoid), kwargs = {})
triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6 = async_compile.triton('triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 25) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + (x3), xmask)
tmp4 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = 0.0
tmp7 = tmp2 > tmp6
tmp8 = 0.2
tmp9 = tmp2 * tmp8
tmp10 = tl.where(tmp7, tmp2, tmp9)
tmp11 = tl.sigmoid(tmp5)
tmp12 = tmp10 * tmp11
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
tl.store(in_out_ptr1 + (x3), tmp5, xmask)
tl.store(out_ptr0 + (x3), tmp12, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (64, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (64, ), (1, ))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._unsafe_index]
stream0 = get_raw_stream(0)
triton_poi_fused__unsafe_index_0.run(primals_1, buf0, 1024, grid=grid(1024), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((64, ), (1, ), torch.float32)
buf2 = empty_strided_cuda((), (), torch.float32)
buf3 = buf2; del buf2 # reuse
buf27 = empty_strided_cuda((64, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mv, norm, add, truediv], Original ATen: [aten.mv, aten.linalg_vector_norm, aten.add, aten.div]
triton_per_fused_add_div_linalg_vector_norm_mv_1.run(buf3, primals_4, primals_2, buf1, buf27, 1, 64, grid=grid(1), stream=stream0)
buf4 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [truediv, mv_1], Original ATen: [aten.div, aten.mv]
triton_per_fused_div_mv_2.run(primals_4, buf1, buf3, buf4, 4, 64, grid=grid(4), stream=stream0)
buf6 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [norm_1, add_1, truediv_1], Original ATen: [aten.linalg_vector_norm, aten.add, aten.div]
triton_per_fused_add_div_linalg_vector_norm_3.run(buf4, buf6, 1, 4, grid=grid(1), stream=stream0)
buf7 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [sigma], Original ATen: [aten.dot]
triton_per_fused_dot_4.run(buf6, buf4, buf7, 1, 4, grid=grid(1), stream=stream0)
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [truediv_2], Original ATen: [aten.div]
triton_poi_fused_div_5.run(primals_4, buf7, buf8, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [conv], Original ATen: [aten.convolution]
buf9 = extern_kernels.convolution(buf0, buf8, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 4, 5, 5), (100, 25, 5, 1))
buf11 = empty_strided_cuda((64, ), (1, ), torch.float32)
buf12 = empty_strided_cuda((), (), torch.float32)
buf13 = buf12; del buf12 # reuse
buf36 = empty_strided_cuda((64, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mv_3, norm_2, add_2, truediv_3], Original ATen: [aten.mv, aten.linalg_vector_norm, aten.add, aten.div]
triton_per_fused_add_div_linalg_vector_norm_mv_1.run(buf13, primals_8, primals_6, buf11, buf36, 1, 64, grid=grid(1), stream=stream0)
buf14 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [truediv_3, mv_4], Original ATen: [aten.div, aten.mv]
triton_per_fused_div_mv_2.run(primals_8, buf11, buf13, buf14, 4, 64, grid=grid(4), stream=stream0)
buf16 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [norm_3, add_3, truediv_4], Original ATen: [aten.linalg_vector_norm, aten.add, aten.div]
triton_per_fused_add_div_linalg_vector_norm_3.run(buf14, buf16, 1, 4, grid=grid(1), stream=stream0)
buf17 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [sigma_1], Original ATen: [aten.dot]
triton_per_fused_dot_4.run(buf16, buf14, buf17, 1, 4, grid=grid(1), stream=stream0)
del buf14
buf18 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [truediv_5], Original ATen: [aten.div]
triton_poi_fused_div_5.run(primals_8, buf17, buf18, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [mask], Original ATen: [aten.convolution]
buf19 = extern_kernels.convolution(buf0, buf18, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 4, 5, 5), (100, 25, 5, 1))
buf10 = buf9; del buf9 # reuse
buf20 = buf19; del buf19 # reuse
buf21 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv, mask, gated_mask, conv_1, x_2], Original ATen: [aten.convolution, aten.sigmoid, aten.leaky_relu, aten.mul]
triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6.run(buf10, buf20, primals_5, primals_9, buf21, 400, grid=grid(400), stream=stream0)
del primals_5
del primals_9
# Topologically Sorted Source Nodes: [], Original ATen: []
buf22 = torch.ops.aten.set_.source_Tensor(primals_2, buf6)
assert_size_stride(buf22, (4, ), (1, ))
del buf1
# Topologically Sorted Source Nodes: [truediv], Original ATen: [aten.div]
buf28 = torch.ops.aten.set_.source_Tensor(primals_3, buf27)
assert_size_stride(buf28, (64, ), (1, ))
del primals_3
# Topologically Sorted Source Nodes: [], Original ATen: []
buf31 = torch.ops.aten.set_.source_Tensor(primals_6, buf16)
assert_size_stride(buf31, (4, ), (1, ))
del buf11
# Topologically Sorted Source Nodes: [truediv_3], Original ATen: [aten.div]
buf37 = torch.ops.aten.set_.source_Tensor(primals_7, buf36)
assert_size_stride(buf37, (64, ), (1, ))
del primals_7
return (buf21, buf8, buf18, primals_2, primals_4, primals_6, primals_8, buf0, buf3, buf6, buf7, buf8, buf10, buf13, buf16, buf17, buf18, buf20, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
from torch.nn import functional as F
from torch.nn import Parameter
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class GatedConv2d(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='reflect', activation='elu', norm=
'none', sn=False):
super(GatedConv2d, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
self.mask_conv2d = SpectralNorm(nn.Conv2d(in_channels,
out_channels, kernel_size, stride, padding=0, dilation=
dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
self.mask_conv2d = nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation)
self.sigmoid = torch.nn.Sigmoid()
def forward(self, x):
x = self.pad(x)
conv = self.conv2d(x)
mask = self.mask_conv2d(x)
gated_mask = self.sigmoid(mask)
if self.activation:
conv = self.activation(conv)
x = conv * gated_mask
return x
class TransposeGatedConv2d(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='lrelu', norm=
'none', sn=True, scale_factor=2):
super(TransposeGatedConv2d, self).__init__()
self.scale_factor = scale_factor
self.gated_conv2d = GatedConv2d(in_channels, out_channels,
kernel_size, stride, padding, dilation, pad_type, activation,
norm, sn)
def forward(self, x):
x = F.interpolate(x, scale_factor=self.scale_factor, mode='nearest')
x = self.gated_conv2d(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__unsafe_index_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 8 % 8
x0 = xindex % 8
x2 = xindex // 64
x4 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = x0
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + 4 * tmp4 + 16 * x2), xmask,
eviction_policy='evict_last')
tl.store(out_ptr0 + x4, tmp9, xmask)
@triton.jit
def triton_per_fused_add_div_linalg_vector_norm_mv_1(in_out_ptr0, in_ptr0,
in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.load(in_ptr0 + (64 + r0), None)
tmp5 = tl.load(in_ptr1 + 1)
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp9 = tl.load(in_ptr0 + (128 + r0), None)
tmp10 = tl.load(in_ptr1 + 2)
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tmp14 = tl.load(in_ptr0 + (192 + r0), None)
tmp15 = tl.load(in_ptr1 + 3)
tmp16 = tl.broadcast_to(tmp15, [XBLOCK, RBLOCK])
tmp3 = tmp0 * tmp2
tmp7 = tmp4 * tmp6
tmp8 = tmp3 + tmp7
tmp12 = tmp9 * tmp11
tmp13 = tmp8 + tmp12
tmp17 = tmp14 * tmp16
tmp18 = tmp13 + tmp17
tmp19 = tmp18 * tmp18
tmp20 = tl.broadcast_to(tmp19, [XBLOCK, RBLOCK])
tmp22 = tl.sum(tmp20, 1)[:, None]
tmp23 = libdevice.sqrt(tmp22)
tmp24 = 1e-12
tmp25 = tmp23 + tmp24
tmp26 = tmp18 / tmp25
tl.store(out_ptr0 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp18, None)
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp25, None)
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp26, None)
@triton.jit
def triton_per_fused_div_mv_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + r1, None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp4 = tmp1 / tmp3
tmp5 = tmp0 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tl.store(out_ptr0 + x0, tmp9, xmask)
@triton.jit
def triton_per_fused_add_div_linalg_vector_norm_3(in_ptr0, out_ptr1, xnumel,
rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp5 = libdevice.sqrt(tmp4)
tmp6 = 1e-12
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp8, None)
@triton.jit
def triton_per_fused_dot_4(in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr1 + r0, None)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
@triton.jit
def triton_poi_fused_div_5(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 / tmp2
tl.store(out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6(in_out_ptr0,
in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 25 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + x3, xmask)
tmp4 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = 0.0
tmp7 = tmp2 > tmp6
tmp8 = 0.2
tmp9 = tmp2 * tmp8
tmp10 = tl.where(tmp7, tmp2, tmp9)
tmp11 = tl.sigmoid(tmp5)
tmp12 = tmp10 * tmp11
tl.store(in_out_ptr0 + x3, tmp2, xmask)
tl.store(in_out_ptr1 + x3, tmp5, xmask)
tl.store(out_ptr0 + x3, tmp12, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (64,), (1,))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (64,), (1,))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__unsafe_index_0[grid(1024)](primals_1, buf0, 1024,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((64,), (1,), torch.float32)
buf2 = empty_strided_cuda((), (), torch.float32)
buf3 = buf2
del buf2
buf27 = empty_strided_cuda((64,), (1,), torch.float32)
triton_per_fused_add_div_linalg_vector_norm_mv_1[grid(1)](buf3,
primals_4, primals_2, buf1, buf27, 1, 64, XBLOCK=1, num_warps=2,
num_stages=1)
buf4 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_div_mv_2[grid(4)](primals_4, buf1, buf3, buf4, 4,
64, XBLOCK=1, num_warps=2, num_stages=1)
buf6 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_div_linalg_vector_norm_3[grid(1)](buf4, buf6,
1, 4, XBLOCK=1, num_warps=2, num_stages=1)
buf7 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_dot_4[grid(1)](buf6, buf4, buf7, 1, 4, XBLOCK=1,
num_warps=2, num_stages=1)
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_div_5[grid(256)](primals_4, buf7, buf8, 256,
XBLOCK=128, num_warps=4, num_stages=1)
buf9 = extern_kernels.convolution(buf0, buf8, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 4, 5, 5), (100, 25, 5, 1))
buf11 = empty_strided_cuda((64,), (1,), torch.float32)
buf12 = empty_strided_cuda((), (), torch.float32)
buf13 = buf12
del buf12
buf36 = empty_strided_cuda((64,), (1,), torch.float32)
triton_per_fused_add_div_linalg_vector_norm_mv_1[grid(1)](buf13,
primals_8, primals_6, buf11, buf36, 1, 64, XBLOCK=1, num_warps=
2, num_stages=1)
buf14 = buf4
del buf4
triton_per_fused_div_mv_2[grid(4)](primals_8, buf11, buf13, buf14,
4, 64, XBLOCK=1, num_warps=2, num_stages=1)
buf16 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_div_linalg_vector_norm_3[grid(1)](buf14, buf16,
1, 4, XBLOCK=1, num_warps=2, num_stages=1)
buf17 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_dot_4[grid(1)](buf16, buf14, buf17, 1, 4, XBLOCK=1,
num_warps=2, num_stages=1)
del buf14
buf18 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_div_5[grid(256)](primals_8, buf17, buf18, 256,
XBLOCK=128, num_warps=4, num_stages=1)
buf19 = extern_kernels.convolution(buf0, buf18, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 4, 5, 5), (100, 25, 5, 1))
buf10 = buf9
del buf9
buf20 = buf19
del buf19
buf21 = empty_strided_cuda((4, 4, 5, 5), (100, 25, 5, 1), torch.float32
)
triton_poi_fused_convolution_leaky_relu_mul_sigmoid_6[grid(400)](buf10,
buf20, primals_5, primals_9, buf21, 400, XBLOCK=256, num_warps=
4, num_stages=1)
del primals_5
del primals_9
buf22 = torch.ops.aten.set_.source_Tensor(primals_2, buf6)
assert_size_stride(buf22, (4,), (1,))
del buf1
buf28 = torch.ops.aten.set_.source_Tensor(primals_3, buf27)
assert_size_stride(buf28, (64,), (1,))
del primals_3
buf31 = torch.ops.aten.set_.source_Tensor(primals_6, buf16)
assert_size_stride(buf31, (4,), (1,))
del buf11
buf37 = torch.ops.aten.set_.source_Tensor(primals_7, buf36)
assert_size_stride(buf37, (64,), (1,))
del primals_7
return (buf21, buf8, buf18, primals_2, primals_4, primals_6, primals_8,
buf0, buf3, buf6, buf7, buf8, buf10, buf13, buf16, buf17, buf18, buf20)
def l2normalize(v, eps=1e-12):
return v / (v.norm() + eps)
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-08, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = Parameter(torch.Tensor(num_features).uniform_())
self.beta = Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
if x.size(0) == 1:
mean = x.view(-1).mean().view(*shape)
std = x.view(-1).std().view(*shape)
else:
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class SpectralNorm(nn.Module):
def __init__(self, module, name='weight', power_iterations=1):
super(SpectralNorm, self).__init__()
self.module = module
self.name = name
self.power_iterations = power_iterations
if not self._made_params():
self._make_params()
def _update_u_v(self):
u = getattr(self.module, self.name + '_u')
v = getattr(self.module, self.name + '_v')
w = getattr(self.module, self.name + '_bar')
height = w.data.shape[0]
for _ in range(self.power_iterations):
v.data = l2normalize(torch.mv(torch.t(w.view(height, -1).data),
u.data))
u.data = l2normalize(torch.mv(w.view(height, -1).data, v.data))
sigma = u.dot(w.view(height, -1).mv(v))
setattr(self.module, self.name, w / sigma.expand_as(w))
def _made_params(self):
try:
getattr(self.module, self.name + '_u')
getattr(self.module, self.name + '_v')
getattr(self.module, self.name + '_bar')
return True
except AttributeError:
return False
def _make_params(self):
w = getattr(self.module, self.name)
height = w.data.shape[0]
width = w.view(height, -1).data.shape[1]
u = Parameter(w.data.new(height).normal_(0, 1), requires_grad=False)
v = Parameter(w.data.new(width).normal_(0, 1), requires_grad=False)
u.data = l2normalize(u.data)
v.data = l2normalize(v.data)
w_bar = Parameter(w.data)
del self.module._parameters[self.name]
self.module.register_parameter(self.name + '_u', u)
self.module.register_parameter(self.name + '_v', v)
self.module.register_parameter(self.name + '_bar', w_bar)
def forward(self, *args):
self._update_u_v()
return self.module.forward(*args)
class GatedConv2d(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='reflect', activation='elu', norm=
'none', sn=False):
super(GatedConv2d, self).__init__()
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'replicate':
self.pad = nn.ReplicationPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
if norm == 'bn':
self.norm = nn.BatchNorm2d(out_channels)
elif norm == 'in':
self.norm = nn.InstanceNorm2d(out_channels)
elif norm == 'ln':
self.norm = LayerNorm(out_channels)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'elu':
self.activation = nn.ELU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'sigmoid':
self.activation = nn.Sigmoid()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
if sn:
self.conv2d = SpectralNorm(nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation))
self.mask_conv2d = SpectralNorm(nn.Conv2d(in_channels,
out_channels, kernel_size, stride, padding=0, dilation=
dilation))
else:
self.conv2d = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=0, dilation=dilation)
self.mask_conv2d = nn.Conv2d(in_channels, out_channels,
kernel_size, stride, padding=0, dilation=dilation)
self.sigmoid = torch.nn.Sigmoid()
def forward(self, x):
x = self.pad(x)
conv = self.conv2d(x)
mask = self.mask_conv2d(x)
gated_mask = self.sigmoid(mask)
if self.activation:
conv = self.activation(conv)
x = conv * gated_mask
return x
class TransposeGatedConv2dNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, pad_type='zero', activation='lrelu', norm=
'none', sn=True, scale_factor=2):
super(TransposeGatedConv2dNew, self).__init__()
self.scale_factor = scale_factor
self.gated_conv2d = GatedConv2d(in_channels, out_channels,
kernel_size, stride, padding, dilation, pad_type, activation,
norm, sn)
def forward(self, input_0):
primals_2 = self.gated_conv2d.conv2d.module.bias
primals_5 = self.gated_conv2d.conv2d.module.weight_u
primals_3 = self.gated_conv2d.conv2d.module.weight_v
primals_1 = self.gated_conv2d.conv2d.module.weight_bar
primals_6 = self.gated_conv2d.mask_conv2d.module.bias
primals_9 = self.gated_conv2d.mask_conv2d.module.weight_u
primals_7 = self.gated_conv2d.mask_conv2d.module.weight_v
primals_4 = self.gated_conv2d.mask_conv2d.module.weight_bar
primals_8 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
|
kangzhiq/DeepFillv2_Pytorch
|
TransposeGatedConv2d
| false | 10,445 |
[
"MIT"
] | 0 |
9c7ed61b25bb995713f89108b712490737abe1b1
|
https://github.com/kangzhiq/DeepFillv2_Pytorch/tree/9c7ed61b25bb995713f89108b712490737abe1b1
|
NextSentencePrediction
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yy/cyya3js6wt64vdji3sfisvrqyfvqxwkwqq5mzg5bqjl2crzjs4t3.py
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# linear => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%select,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/t6/ct6t4t4f6df2rswen66qmskrag4dcnuaoxkebk4pssna63yxl3v3.py
# Topologically Sorted Source Nodes: [linear, log_softmax], Original ATen: [aten.add, aten._log_softmax]
# Source node to ATen node mapping:
# linear => add
# log_softmax => amax, sub
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %primals_3), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%add, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %amax), kwargs = {})
triton_poi_fused__log_softmax_add_1 = async_compile.triton('triton_poi_fused__log_softmax_add_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_add_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_add_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 2
x1 = (xindex // 2)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2*x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (0))
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (1 + (2*x1)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (1))
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp6 = tmp3 + tmp5
tmp10 = tmp7 + tmp9
tmp11 = triton_helpers.maximum(tmp6, tmp10)
tmp12 = tmp2 - tmp11
tl.store(out_ptr0 + (x2), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jt/cjtlgs566wh4nqwf6wvcsracaovzta35wbm5vllgfuavp5ebkiq4.py
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax => exp, log, sub_1, sum_1
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
triton_poi_fused__log_softmax_2 = async_compile.triton('triton_poi_fused__log_softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 2)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (2*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (2*x1)), xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp6 = tl_math.log(tmp5)
tmp7 = tmp0 - tmp6
tl.store(out_ptr0 + (x2), tmp7, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4), (4, 1))
assert_size_stride(primals_3, (2, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_1, buf0, 64, grid=grid(64), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((16, 2), (2, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 2), (1, 4), 0), out=buf1)
del primals_2
buf2 = empty_strided_cuda((4, 4, 2), (8, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear, log_softmax], Original ATen: [aten.add, aten._log_softmax]
triton_poi_fused__log_softmax_add_1.run(buf1, primals_3, buf2, 32, grid=grid(32), stream=stream0)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 4, 2), (8, 2, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_2.run(buf2, buf3, 32, grid=grid(32), stream=stream0)
del buf2
return (buf3, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((2, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((2, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
class NextSentencePrediction(nn.Module):
"""
2-class classification model : is_next, is_not_next
"""
def __init__(self, hidden):
"""
:param hidden: BERT model output size
"""
super().__init__()
self.linear = nn.Linear(hidden, 2)
self.softmax = nn.LogSoftmax(dim=-1)
def forward(self, x):
return self.softmax(self.linear(x[:, 0]))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'hidden': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused__log_softmax_add_1(in_ptr0, in_ptr1, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 2
x1 = xindex // 2
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + 2 * x1, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + 0)
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (1 + 2 * x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + 1)
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp6 = tmp3 + tmp5
tmp10 = tmp7 + tmp9
tmp11 = triton_helpers.maximum(tmp6, tmp10)
tmp12 = tmp2 - tmp11
tl.store(out_ptr0 + x2, tmp12, xmask)
@triton.jit
def triton_poi_fused__log_softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 2
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 2 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 2 * x1), xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp6 = tl_math.log(tmp5)
tmp7 = tmp0 - tmp6
tl.store(out_ptr0 + x2, tmp7, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4), (4, 1))
assert_size_stride(primals_3, (2,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(64)](primals_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((16, 2), (2, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 2), (1, 4), 0), out=buf1)
del primals_2
buf2 = empty_strided_cuda((4, 4, 2), (8, 2, 1), torch.float32)
triton_poi_fused__log_softmax_add_1[grid(32)](buf1, primals_3, buf2,
32, XBLOCK=32, num_warps=1, num_stages=1)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 4, 2), (8, 2, 1), 0)
del buf1
triton_poi_fused__log_softmax_2[grid(32)](buf2, buf3, 32, XBLOCK=32,
num_warps=1, num_stages=1)
del buf2
return buf3, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf3
class NextSentencePredictionNew(nn.Module):
"""
2-class classification model : is_next, is_not_next
"""
def __init__(self, hidden):
"""
:param hidden: BERT model output size
"""
super().__init__()
self.linear = nn.Linear(hidden, 2)
self.softmax = nn.LogSoftmax(dim=-1)
def forward(self, input_0):
primals_2 = self.linear.weight
primals_3 = self.linear.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
greenstar1151/pytorch-benchmark
|
NextSentencePrediction
| false | 10,446 |
[
"BSD-3-Clause"
] | 0 |
8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
https://github.com/greenstar1151/pytorch-benchmark/tree/8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
AttentionPool2d
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ov/covbryzjnff2kb26c5gkcqbvct6kdwzanlx3iu6ee24itsit76o3.py
# Topologically Sorted Source Nodes: [mean], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# mean => mean
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%view, [-1], True), kwargs = {})
triton_per_fused_mean_0 = async_compile.triton('triton_per_fused_mean_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_0(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ct/cctgbe64jgxq3sxjjjqccvq653sunfecfcizp3jcofnl7uiib7wo.py
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.cat, aten.add]
# Source node to ATen node mapping:
# x_1 => cat
# x_2 => add
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean, %view], -1), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%cat, %unsqueeze), kwargs = {})
triton_poi_fused_add_cat_1 = async_compile.triton('triton_poi_fused_add_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_cat_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 17
x3 = (xindex // 17)
x4 = xindex % 68
x5 = xindex
tmp15 = tl.load(in_ptr2 + (x4), xmask, eviction_policy='evict_last')
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x3), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = 16.0
tmp7 = tmp5 / tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 17, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tl.load(in_ptr1 + ((16*x3) + ((-1) + x0)), tmp10 & xmask, eviction_policy='evict_last', other=0.0)
tmp14 = tl.where(tmp4, tmp9, tmp13)
tmp16 = tmp14 + tmp15
tl.store(out_ptr0 + (x5), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/rq/crqhbpxwwpvdowuqzuzvjdahxv45tx2y4dpxda2rurvr5kralgbn.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%getitem, 0.7071067811865475), kwargs = {})
triton_poi_fused_mul_2 = async_compile.triton('triton_poi_fused_mul_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 68)
x3 = xindex % 68
x1 = (xindex // 17) % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3 + (204*x2)), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x4), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/cr/ccrjqh2olqyevqox4t3kpoubo2s2m44cuvtxmf2k3slvdcgmmlcm.py
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%getitem_1, 0.7071067811865475), kwargs = {})
triton_poi_fused_mul_3 = async_compile.triton('triton_poi_fused_mul_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 68)
x3 = xindex % 68
x1 = (xindex // 17) % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (68 + x3 + (204*x2)), xmask)
tmp1 = tl.load(in_ptr1 + (4 + x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x4), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6p/c6p34hneq7lp7a3tjiwk44lxqe2hzbpdjgbgmanamv35xvznwb2j.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax, div, exp, sub, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_per_fused__softmax_4 = async_compile.triton('triton_per_fused__softmax_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[128, 32],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__softmax_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__softmax_4(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 68
rnumel = 17
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (17*x0)), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float("-inf"))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + (17*x0)), tmp11, rmask & xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wf/cwf36kt6t5p6sv4fjknukcyy4vz6ejfamnreogfggthvquklzkf6.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_3 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%add, %primals_3, %primals_4, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_5 = async_compile.triton('triton_poi_fused_convolution_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_5(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 816
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 17) % 12
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/oe/coet3g5rxo652nti4d4ogiwpej2mrpgh4cyidpv6aivegc4mda7s.py
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_5 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%view_9, %primals_5, %primals_6, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_6 = async_compile.triton('triton_poi_fused_convolution_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 32], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_6(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 17
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (68*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (17*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/uk/cuk32wvqx6lzak6biu2zut26pzpqa3ell4xsxs3qv3cpuuon7kkw.py
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_5 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%view_9, %primals_5, %primals_6, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_7 = async_compile.triton('triton_poi_fused_convolution_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_7', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_7(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 17) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 17), (17, 1))
assert_size_stride(primals_3, (12, 4, 1), (4, 1, 1))
assert_size_stride(primals_4, (12, ), (1, ))
assert_size_stride(primals_5, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_6, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [mean], Original ATen: [aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_mean_0.run(primals_1, buf0, 16, 16, grid=grid(16), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.cat, aten.add]
triton_poi_fused_add_cat_1.run(buf0, primals_1, primals_2, buf1, 272, grid=grid(272), stream=stream0)
del buf0
del primals_1
del primals_2
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_3, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf2, (4, 12, 17), (204, 17, 1))
buf3 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
triton_poi_fused_mul_2.run(buf2, primals_4, buf3, 272, grid=grid(272), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
triton_poi_fused_mul_3.run(buf2, primals_4, buf4, 272, grid=grid(272), stream=stream0)
buf5 = empty_strided_cuda((4, 17, 17), (289, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [weight], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (4, 17, 4), (68, 1, 17), 0), buf4, out=buf5)
buf8 = empty_strided_cuda((4, 17, 17), (289, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_per_fused__softmax_4.run(buf5, buf8, 68, 17, grid=grid(68), stream=stream0)
del buf5
buf9 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.convolution]
triton_poi_fused_convolution_5.run(buf9, primals_4, 816, grid=grid(816), stream=stream0)
del primals_4
buf10 = empty_strided_cuda((4, 17, 4), (68, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [a], Original ATen: [aten.bmm]
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (4, 17, 4), (204, 1, 17), 136), out=buf10)
buf11 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.convolution]
triton_poi_fused_convolution_6.run(buf10, buf11, 16, 17, grid=grid(16, 17), stream=stream0)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.convolution]
buf12 = extern_kernels.convolution(buf11, primals_5, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf12, (4, 4, 17), (68, 17, 1))
del buf11
buf13 = buf12; del buf12 # reuse
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.convolution]
triton_poi_fused_convolution_7.run(buf13, primals_6, 272, grid=grid(272), stream=stream0)
del primals_6
return (reinterpret_tensor(buf13, (4, 4), (68, 17), 0), primals_3, primals_5, buf1, buf8, reinterpret_tensor(buf10, (4, 4, 17), (68, 1, 4), 0), reinterpret_tensor(buf9, (4, 4, 17), (204, 17, 1), 136), buf3, reinterpret_tensor(buf4, (4, 17, 4), (68, 1, 17), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 17), (17, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((12, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((12, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
from torch import nn
import torch as th
def conv_nd(dims, *args, **kwargs):
"""
Create a 1D, 2D, or 3D convolution module.
"""
if dims == 1:
return nn.Conv1d(*args, **kwargs)
elif dims == 2:
return nn.Conv2d(*args, **kwargs)
elif dims == 3:
return nn.Conv3d(*args, **kwargs)
raise ValueError(f'unsupported dimensions: {dims}')
class QKVAttention(nn.Module):
"""
A module which performs QKV attention. Matches legacy QKVAttention + input/ouput heads shaping
"""
def __init__(self, n_heads):
super().__init__()
self.n_heads = n_heads
def forward(self, qkv, encoder_kv=None):
"""
Apply QKV attention.
:param qkv: an [N x (H * 3 * C) x T] tensor of Qs, Ks, and Vs.
:return: an [N x (H * C) x T] tensor after attention.
"""
bs, width, length = qkv.shape
assert width % (3 * self.n_heads) == 0
ch = width // (3 * self.n_heads)
q, k, v = qkv.reshape(bs * self.n_heads, ch * 3, length).split(ch,
dim=1)
if encoder_kv is not None:
assert encoder_kv.shape[1] == self.n_heads * ch * 2
ek, ev = encoder_kv.reshape(bs * self.n_heads, ch * 2, -1).split(ch
, dim=1)
k = th.cat([ek, k], dim=-1)
v = th.cat([ev, v], dim=-1)
scale = 1 / math.sqrt(math.sqrt(ch))
weight = th.einsum('bct,bcs->bts', q * scale, k * scale)
weight = th.softmax(weight.float(), dim=-1).type(weight.dtype)
a = th.einsum('bts,bcs->bct', weight, v)
return a.reshape(bs, -1, length)
class AttentionPool2d(nn.Module):
"""
Adapted from CLIP: https://github.com/openai/CLIP/blob/main/clip/model.py
"""
def __init__(self, spacial_dim: 'int', embed_dim: 'int',
num_heads_channels: 'int', output_dim: 'int'=None):
super().__init__()
self.positional_embedding = nn.Parameter(th.randn(embed_dim,
spacial_dim ** 2 + 1) / embed_dim ** 0.5)
self.qkv_proj = conv_nd(1, embed_dim, 3 * embed_dim, 1)
self.c_proj = conv_nd(1, embed_dim, output_dim or embed_dim, 1)
self.num_heads = embed_dim // num_heads_channels
self.attention = QKVAttention(self.num_heads)
def forward(self, x):
b, c, *_spatial = x.shape
x = x.reshape(b, c, -1)
x = th.cat([x.mean(dim=-1, keepdim=True), x], dim=-1)
x = x + self.positional_embedding[None, :, :]
x = self.qkv_proj(x)
x = self.attention(x)
x = self.c_proj(x)
return x[:, :, 0]
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'spacial_dim': 4, 'embed_dim': 4, 'num_heads_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
from torch import nn
import torch as th
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mean_0(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK: tl.
constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tl.store(out_ptr0 + x0, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_cat_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 17
x3 = xindex // 17
x4 = xindex % 68
x5 = xindex
tmp15 = tl.load(in_ptr2 + x4, xmask, eviction_policy='evict_last')
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + x3, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp6 = 16.0
tmp7 = tmp5 / tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tl.full([1], 17, tl.int64)
tmp13 = tl.load(in_ptr1 + (16 * x3 + (-1 + x0)), tmp10 & xmask,
eviction_policy='evict_last', other=0.0)
tmp14 = tl.where(tmp4, tmp9, tmp13)
tmp16 = tmp14 + tmp15
tl.store(out_ptr0 + x5, tmp16, xmask)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 68
x3 = xindex % 68
x1 = xindex // 17 % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3 + 204 * x2), xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + x4, tmp4, xmask)
@triton.jit
def triton_poi_fused_mul_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 68
x3 = xindex % 68
x1 = xindex // 17 % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (68 + x3 + 204 * x2), xmask)
tmp1 = tl.load(in_ptr1 + (4 + x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + x4, tmp4, xmask)
@triton.jit
def triton_per_fused__softmax_4(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 68
rnumel = 17
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 17 * x0), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float('-inf'))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tmp6 / tmp10
tl.store(out_ptr2 + (r1 + 17 * x0), tmp11, rmask & xmask)
@triton.jit
def triton_poi_fused_convolution_5(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 816
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 17 % 12
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_convolution_6(in_ptr0, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 17
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 68 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 17 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_convolution_7(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 272
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 17 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 17), (17, 1))
assert_size_stride(primals_3, (12, 4, 1), (4, 1, 1))
assert_size_stride(primals_4, (12,), (1,))
assert_size_stride(primals_5, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_6, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
get_raw_stream(0)
triton_per_fused_mean_0[grid(16)](primals_1, buf0, 16, 16, XBLOCK=1,
num_warps=2, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
triton_poi_fused_add_cat_1[grid(272)](buf0, primals_1, primals_2,
buf1, 272, XBLOCK=128, num_warps=4, num_stages=1)
del buf0
del primals_1
del primals_2
buf2 = extern_kernels.convolution(buf1, primals_3, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf2, (4, 12, 17), (204, 17, 1))
buf3 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
triton_poi_fused_mul_2[grid(272)](buf2, primals_4, buf3, 272,
XBLOCK=256, num_warps=4, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
triton_poi_fused_mul_3[grid(272)](buf2, primals_4, buf4, 272,
XBLOCK=256, num_warps=4, num_stages=1)
buf5 = empty_strided_cuda((4, 17, 17), (289, 17, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (4, 17, 4), (68, 1, 17),
0), buf4, out=buf5)
buf8 = empty_strided_cuda((4, 17, 17), (289, 17, 1), torch.float32)
triton_per_fused__softmax_4[grid(68)](buf5, buf8, 68, 17, XBLOCK=1,
num_warps=2, num_stages=1)
del buf5
buf9 = buf2
del buf2
triton_poi_fused_convolution_5[grid(816)](buf9, primals_4, 816,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_4
buf10 = empty_strided_cuda((4, 17, 4), (68, 4, 1), torch.float32)
extern_kernels.bmm(buf8, reinterpret_tensor(buf9, (4, 17, 4), (204,
1, 17), 136), out=buf10)
buf11 = empty_strided_cuda((4, 4, 17), (68, 17, 1), torch.float32)
triton_poi_fused_convolution_6[grid(16, 17)](buf10, buf11, 16, 17,
XBLOCK=32, YBLOCK=16, num_warps=4, num_stages=1)
buf12 = extern_kernels.convolution(buf11, primals_5, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf12, (4, 4, 17), (68, 17, 1))
del buf11
buf13 = buf12
del buf12
triton_poi_fused_convolution_7[grid(272)](buf13, primals_6, 272,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_6
return reinterpret_tensor(buf13, (4, 4), (68, 17), 0
), primals_3, primals_5, buf1, buf8, reinterpret_tensor(buf10, (4,
4, 17), (68, 1, 4), 0), reinterpret_tensor(buf9, (4, 4, 17), (204,
17, 1), 136), buf3, reinterpret_tensor(buf4, (4, 17, 4), (68, 1, 17), 0
)
def conv_nd(dims, *args, **kwargs):
"""
Create a 1D, 2D, or 3D convolution module.
"""
if dims == 1:
return nn.Conv1d(*args, **kwargs)
elif dims == 2:
return nn.Conv2d(*args, **kwargs)
elif dims == 3:
return nn.Conv3d(*args, **kwargs)
raise ValueError(f'unsupported dimensions: {dims}')
class QKVAttention(nn.Module):
"""
A module which performs QKV attention. Matches legacy QKVAttention + input/ouput heads shaping
"""
def __init__(self, n_heads):
super().__init__()
self.n_heads = n_heads
def forward(self, qkv, encoder_kv=None):
"""
Apply QKV attention.
:param qkv: an [N x (H * 3 * C) x T] tensor of Qs, Ks, and Vs.
:return: an [N x (H * C) x T] tensor after attention.
"""
bs, width, length = qkv.shape
assert width % (3 * self.n_heads) == 0
ch = width // (3 * self.n_heads)
q, k, v = qkv.reshape(bs * self.n_heads, ch * 3, length).split(ch,
dim=1)
if encoder_kv is not None:
assert encoder_kv.shape[1] == self.n_heads * ch * 2
ek, ev = encoder_kv.reshape(bs * self.n_heads, ch * 2, -1).split(ch
, dim=1)
k = th.cat([ek, k], dim=-1)
v = th.cat([ev, v], dim=-1)
scale = 1 / math.sqrt(math.sqrt(ch))
weight = th.einsum('bct,bcs->bts', q * scale, k * scale)
weight = th.softmax(weight.float(), dim=-1).type(weight.dtype)
a = th.einsum('bts,bcs->bct', weight, v)
return a.reshape(bs, -1, length)
class AttentionPool2dNew(nn.Module):
"""
Adapted from CLIP: https://github.com/openai/CLIP/blob/main/clip/model.py
"""
def __init__(self, spacial_dim: 'int', embed_dim: 'int',
num_heads_channels: 'int', output_dim: 'int'=None):
super().__init__()
self.positional_embedding = nn.Parameter(th.randn(embed_dim,
spacial_dim ** 2 + 1) / embed_dim ** 0.5)
self.qkv_proj = conv_nd(1, embed_dim, 3 * embed_dim, 1)
self.c_proj = conv_nd(1, embed_dim, output_dim or embed_dim, 1)
self.num_heads = embed_dim // num_heads_channels
self.attention = QKVAttention(self.num_heads)
def forward(self, input_0):
primals_2 = self.positional_embedding
primals_3 = self.qkv_proj.weight
primals_4 = self.qkv_proj.bias
primals_5 = self.c_proj.weight
primals_6 = self.c_proj.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0]
|
litevxx/glid-3
|
AttentionPool2d
| false | 10,447 |
[
"MIT"
] | 0 |
d7bd53e671d642b0cbc8af81197170b585c7e624
|
https://github.com/litevxx/glid-3/tree/d7bd53e671d642b0cbc8af81197170b585c7e624
|
ConcreteDenseMixture
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/5m/c5m6znlu4sybwbvy6htxjumpdsiyi5qroh6vctqzm4q4pynppftx.py
# Topologically Sorted Source Nodes: [p, add, log, sub, add_1, log_1, sub_1, add_2, log_2, add_3, sub_2, add_4, log_3, drop_prob, truediv, drop_prob_1, random_tensor, x, x_1], Original ATen: [aten.sigmoid, aten.add, aten.log, aten.rsub, aten.sub, aten.div, aten.mul]
# Source node to ATen node mapping:
# add => add
# add_1 => add_1
# add_2 => add_2
# add_3 => add_3
# add_4 => add_4
# drop_prob => sub_3
# drop_prob_1 => sigmoid_1
# log => log
# log_1 => log_1
# log_2 => log_2
# log_3 => log_3
# p => sigmoid
# random_tensor => sub_4
# sub => sub
# sub_1 => sub_1
# sub_2 => sub_2
# truediv => div
# x => mul
# x_1 => div_1
# Graph fragment:
# %sigmoid : [num_users=5] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_2,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sigmoid, 1e-07), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add,), kwargs = {})
# %sub : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub, 1e-07), kwargs = {})
# %log_1 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log, %log_1), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand, 1e-07), kwargs = {})
# %log_2 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_2,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_1, %log_2), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_2, 1e-07), kwargs = {})
# %log_3 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_4,), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_3, %log_3), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_3, 0.1), kwargs = {})
# %sigmoid_1 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%div,), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, %sub_4), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul, %sub), kwargs = {})
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0 = async_compile.triton('triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp12 = tl.load(in_ptr2 + (x0), xmask)
tmp3 = tl.sigmoid(tmp2)
tmp4 = 1e-07
tmp5 = tmp3 + tmp4
tmp6 = tl_math.log(tmp5)
tmp7 = 1.0
tmp8 = tmp7 - tmp3
tmp9 = tmp8 + tmp4
tmp10 = tl_math.log(tmp9)
tmp11 = tmp6 - tmp10
tmp13 = tmp12 + tmp4
tmp14 = tl_math.log(tmp13)
tmp15 = tmp11 + tmp14
tmp16 = tmp7 - tmp12
tmp17 = tmp16 + tmp4
tmp18 = tl_math.log(tmp17)
tmp19 = tmp15 - tmp18
tmp20 = 10.0
tmp21 = tmp19 * tmp20
tmp22 = tl.sigmoid(tmp21)
tmp23 = tmp7 - tmp22
tmp24 = tmp0 * tmp23
tmp25 = tmp24 / tmp8
tl.store(out_ptr0 + (x0), tmp25, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/oo/coow5so5su6vtsogkckryyczitsbdkncnh4l2dnly7t36twt6us6.py
# Topologically Sorted Source Nodes: [pow_1, sum_1], Original ATen: [aten.pow, aten.sum]
# Source node to ATen node mapping:
# pow_1 => pow_1
# sum_1 => sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_3, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_1,), kwargs = {})
triton_per_fused_pow_sum_1 = async_compile.triton('triton_per_fused_pow_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_pow_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_pow_sum_1(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/or/corzuezitsvrehhmp6i4nhbwxpmbztc6s6rcir6feyzjzvi6an4v.py
# Topologically Sorted Source Nodes: [p, sub, sum_of_square, pow_2, sum_2, sum_of_square_1, mul_1, weights_regularizer, log_4, dropout_regularizer, sub_7, log_5, mul_3, dropout_regularizer_1, dropout_regularizer_2, regularization_1], Original ATen: [aten.sigmoid, aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div, aten.log]
# Source node to ATen node mapping:
# dropout_regularizer => mul_2
# dropout_regularizer_1 => add_7
# dropout_regularizer_2 => mul_4
# log_4 => log_4
# log_5 => log_5
# mul_1 => mul_1
# mul_3 => mul_3
# p => sigmoid
# pow_2 => pow_2
# regularization_1 => add_8
# sub => sub
# sub_7 => sub_7
# sum_2 => sum_2
# sum_of_square => add_5
# sum_of_square_1 => add_6
# weights_regularizer => div_2
# Graph fragment:
# %sigmoid : [num_users=5] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_2,), kwargs = {})
# %sub : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 0), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_4, 2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_2,), kwargs = {})
# %add_6 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_5, %sum_2), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_6, 4), kwargs = {})
# %div_2 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_1, %sub), kwargs = {})
# %log_4 : [num_users=2] = call_function[target=torch.ops.aten.log.default](args = (%sigmoid,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %log_4), kwargs = {})
# %sub_7 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sigmoid), kwargs = {})
# %log_5 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sub_7,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_7, %log_5), kwargs = {})
# %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %mul_3), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_7, 32.0), kwargs = {})
# %add_8 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div_2, %mul_4), kwargs = {})
# %div_160 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_2, %sub), kwargs = {})
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2 = async_compile.triton('triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {7: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=(7,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2(in_ptr0, in_ptr1, in_ptr2, out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp5 = tl.load(in_ptr1 + (0))
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, 1])
tmp11 = tl.load(in_ptr2 + (0))
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp7 = tl.sigmoid(tmp6)
tmp8 = tl_math.log(tmp7)
tmp9 = 1.0
tmp10 = tmp9 - tmp7
tmp13 = 0.0
tmp14 = tmp12 + tmp13
tmp15 = tmp14 + tmp4
tmp16 = 4.0
tmp17 = tmp15 * tmp16
tmp18 = tmp17 / tmp10
tmp19 = tmp7 * tmp8
tmp20 = tl_math.log(tmp10)
tmp21 = tmp10 * tmp20
tmp22 = tmp19 + tmp21
tmp23 = 32.0
tmp24 = tmp22 * tmp23
tmp25 = tmp18 + tmp24
tmp26 = tmp18 / tmp10
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp8, None)
tl.store(out_ptr2 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp10, None)
tl.store(out_ptr3 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp25, None)
tl.store(out_ptr4 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp26, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ch/cchcvqeuz2kdsoaw53vwus5rv43qa3wvhkanfzsif6i6cuwympx5.py
# Topologically Sorted Source Nodes: [input_2, p_1, add_7, log_6, sub_9, add_8, log_7, sub_10, add_9, log_8, add_10, sub_11, add_11, log_9, drop_prob_2, truediv_2, drop_prob_3, random_tensor_1, x_2, x_3], Original ATen: [aten.tanh, aten.sigmoid, aten.add, aten.log, aten.rsub, aten.sub, aten.div, aten.mul]
# Source node to ATen node mapping:
# add_10 => add_12
# add_11 => add_13
# add_7 => add_9
# add_8 => add_10
# add_9 => add_11
# drop_prob_2 => sub_12
# drop_prob_3 => sigmoid_3
# input_2 => tanh
# log_6 => log_6
# log_7 => log_7
# log_8 => log_8
# log_9 => log_9
# p_1 => sigmoid_2
# random_tensor_1 => sub_13
# sub_10 => sub_10
# sub_11 => sub_11
# sub_9 => sub_9
# truediv_2 => div_3
# x_2 => mul_5
# x_3 => div_4
# Graph fragment:
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%view_2,), kwargs = {})
# %sigmoid_2 : [num_users=5] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_5,), kwargs = {})
# %add_9 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sigmoid_2, 1e-07), kwargs = {})
# %log_6 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_9,), kwargs = {})
# %sub_9 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid_2), kwargs = {})
# %add_10 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_9, 1e-07), kwargs = {})
# %log_7 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_10,), kwargs = {})
# %sub_10 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_6, %log_7), kwargs = {})
# %add_11 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_1, 1e-07), kwargs = {})
# %log_8 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_11,), kwargs = {})
# %add_12 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_10, %log_8), kwargs = {})
# %sub_11 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_1), kwargs = {})
# %add_13 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_11, 1e-07), kwargs = {})
# %log_9 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_13,), kwargs = {})
# %sub_12 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_12, %log_9), kwargs = {})
# %div_3 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_12, 0.1), kwargs = {})
# %sigmoid_3 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%div_3,), kwargs = {})
# %sub_13 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid_3), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh, %sub_13), kwargs = {})
# %div_4 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_5, %sub_9), kwargs = {})
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3 = async_compile.triton('triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp2 = tl.load(in_ptr1 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp13 = tl.load(in_ptr2 + (x0), xmask)
tmp1 = libdevice.tanh(tmp0)
tmp4 = tl.sigmoid(tmp3)
tmp5 = 1e-07
tmp6 = tmp4 + tmp5
tmp7 = tl_math.log(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp4
tmp10 = tmp9 + tmp5
tmp11 = tl_math.log(tmp10)
tmp12 = tmp7 - tmp11
tmp14 = tmp13 + tmp5
tmp15 = tl_math.log(tmp14)
tmp16 = tmp12 + tmp15
tmp17 = tmp8 - tmp13
tmp18 = tmp17 + tmp5
tmp19 = tl_math.log(tmp18)
tmp20 = tmp16 - tmp19
tmp21 = 10.0
tmp22 = tmp20 * tmp21
tmp23 = tl.sigmoid(tmp22)
tmp24 = tmp8 - tmp23
tmp25 = tmp1 * tmp24
tmp26 = tmp25 / tmp9
tl.store(out_ptr0 + (x0), tmp26, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ag/cagmdttkouzii56lcbd4mlc6oyw7xiz3hgky42obwbgdidfdzoqs.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_for_fused_4 = async_compile.triton('triton_for_fused_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.foreach(
num_warps=8,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: '*fp32', 15: '*fp32', 16: '*fp32', 17: '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), equal_to_1=())]},
inductor_meta={'kernel_name': 'triton_for_fused_4', 'mutated_arg_names': [], 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
)
@triton.jit
def triton_for_fused_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7, out_ptr8):
pid = tl.program_id(0)
XBLOCK: tl.constexpr = 1024
num_xblocks_0 = tl.cdiv(1, XBLOCK)
num_xblocks_1 = num_xblocks_0 + tl.cdiv(1, XBLOCK)
num_xblocks_2 = num_xblocks_1 + tl.cdiv(1, XBLOCK)
num_xblocks_3 = num_xblocks_2 + tl.cdiv(1, XBLOCK)
num_xblocks_4 = num_xblocks_3 + tl.cdiv(1, XBLOCK)
num_xblocks_5 = num_xblocks_4 + tl.cdiv(1, XBLOCK)
num_xblocks_6 = num_xblocks_5 + tl.cdiv(1, XBLOCK)
num_xblocks_7 = num_xblocks_6 + tl.cdiv(1, XBLOCK)
num_xblocks_8 = num_xblocks_7 + tl.cdiv(1, XBLOCK)
if pid < num_xblocks_0:
pid_offset = pid
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp0 = tl.load(in_ptr0 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tl.store(out_ptr0 + (tl.full([XBLOCK], 0, tl.int32)), tmp1, None)
elif pid < num_xblocks_1:
pid_offset = pid - num_xblocks_0
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp2 = tl.load(in_ptr1 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tl.store(out_ptr1 + (tl.full([XBLOCK], 0, tl.int32)), tmp3, None)
elif pid < num_xblocks_2:
pid_offset = pid - num_xblocks_1
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp4 = tl.load(in_ptr2 + (0))
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tl.store(out_ptr2 + (tl.full([XBLOCK], 0, tl.int32)), tmp5, None)
elif pid < num_xblocks_3:
pid_offset = pid - num_xblocks_2
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp6 = tl.load(in_ptr3 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tl.store(out_ptr3 + (tl.full([XBLOCK], 0, tl.int32)), tmp7, None)
elif pid < num_xblocks_4:
pid_offset = pid - num_xblocks_3
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp8 = tl.load(in_ptr4 + (0))
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tl.store(out_ptr4 + (tl.full([XBLOCK], 0, tl.int32)), tmp9, None)
elif pid < num_xblocks_5:
pid_offset = pid - num_xblocks_4
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp10 = tl.load(in_ptr5 + (0))
tmp11 = tl.broadcast_to(tmp10, [XBLOCK])
tl.store(out_ptr5 + (tl.full([XBLOCK], 0, tl.int32)), tmp11, None)
elif pid < num_xblocks_6:
pid_offset = pid - num_xblocks_5
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp12 = tl.load(in_ptr6 + (0))
tmp13 = tl.broadcast_to(tmp12, [XBLOCK])
tl.store(out_ptr6 + (tl.full([XBLOCK], 0, tl.int32)), tmp13, None)
elif pid < num_xblocks_7:
pid_offset = pid - num_xblocks_6
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp14 = tl.load(in_ptr7 + (0))
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tl.store(out_ptr7 + (tl.full([XBLOCK], 0, tl.int32)), tmp15, None)
elif pid < num_xblocks_8:
pid_offset = pid - num_xblocks_7
xnumel = 1
rnumel = 1
xoffset = pid_offset * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
tmp16 = tl.load(in_ptr8 + (0))
tmp17 = tl.broadcast_to(tmp16, [XBLOCK])
tl.store(out_ptr8 + (tl.full([XBLOCK], 0, tl.int32)), tmp17, None)
else:
pass
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zh/czh2ghogdlf4qx4mbbij2czxblnwgpba7hf4peek3nrfz3ke65y7.py
# Topologically Sorted Source Nodes: [sigmoid_24], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid_24 => sigmoid_6
# Graph fragment:
# %sigmoid_6 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%view_12,), kwargs = {})
triton_poi_fused_sigmoid_5 = async_compile.triton('triton_poi_fused_sigmoid_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_5(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 9
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/rd/crd7zpxjaubl3xwr6s3dxqbevflh4ofguejlgr4a32gpmzolkemf.py
# Topologically Sorted Source Nodes: [sub_99, pow_23, sum_23, sum_of_square_22, pow_24, sum_24, sum_of_square_23, mul_45, weights_regularizer_11, log_70, dropout_regularizer_33, sub_106, log_71, mul_47, dropout_regularizer_34, dropout_regularizer_35, regularization_12], Original ATen: [aten.rsub, aten.pow, aten.sum, aten.add, aten.mul, aten.div, aten.log]
# Source node to ATen node mapping:
# dropout_regularizer_33 => mul_57
# dropout_regularizer_34 => add_106
# dropout_regularizer_35 => mul_59
# log_70 => log_70
# log_71 => log_71
# mul_45 => mul_56
# mul_47 => mul_58
# pow_23 => pow_23
# pow_24 => pow_24
# regularization_12 => add_107
# sub_106 => sub_106
# sub_99 => sub_91
# sum_23 => sum_23
# sum_24 => sum_24
# sum_of_square_22 => add_104
# sum_of_square_23 => add_105
# weights_regularizer_11 => div_35
# Graph fragment:
# %sub_91 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_8), kwargs = {})
# %pow_23 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_36, 2), kwargs = {})
# %sum_23 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_23,), kwargs = {})
# %add_104 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_23, 0), kwargs = {})
# %pow_24 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_37, 2), kwargs = {})
# %sum_24 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_24,), kwargs = {})
# %add_105 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_104, %sum_24), kwargs = {})
# %mul_56 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_105, 4), kwargs = {})
# %div_35 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_56, %sub_91), kwargs = {})
# %log_70 : [num_users=2] = call_function[target=torch.ops.aten.log.default](args = (%getitem_8,), kwargs = {})
# %mul_57 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%getitem_8, %log_70), kwargs = {})
# %sub_106 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %getitem_8), kwargs = {})
# %log_71 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sub_106,), kwargs = {})
# %mul_58 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_106, %log_71), kwargs = {})
# %add_106 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_57, %mul_58), kwargs = {})
# %mul_59 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_106, 32.0), kwargs = {})
# %add_107 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div_35, %mul_59), kwargs = {})
# %div_39 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_35, %sub_91), kwargs = {})
triton_per_fused_add_div_log_mul_pow_rsub_sum_6 = async_compile.triton('triton_per_fused_add_div_log_mul_pow_rsub_sum_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {7: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=(7,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_log_mul_pow_rsub_sum_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_log_mul_pow_rsub_sum_6(in_ptr0, in_ptr1, in_ptr2, out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp5 = tl.load(in_ptr1 + (8))
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, 1])
tmp12 = tl.load(in_ptr2 + (0))
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp7 = tl_math.log(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp6
tmp10 = 0.0
tmp11 = tmp4 + tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp11 + tmp14
tmp16 = 4.0
tmp17 = tmp15 * tmp16
tmp18 = tmp17 / tmp9
tmp19 = tmp6 * tmp7
tmp20 = tl_math.log(tmp9)
tmp21 = tmp9 * tmp20
tmp22 = tmp19 + tmp21
tmp23 = 32.0
tmp24 = tmp22 * tmp23
tmp25 = tmp18 + tmp24
tmp26 = tmp18 / tmp9
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp7, None)
tl.store(out_ptr2 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, None)
tl.store(out_ptr3 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp25, None)
tl.store(out_ptr4 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp26, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tl/ctlayqq4hegcuvwscwjlkk3jvo4lradqxpndemm6kbslgl7aoahd.py
# Topologically Sorted Source Nodes: [pow_8, sum_8], Original ATen: [aten.pow, aten.sum]
# Source node to ATen node mapping:
# pow_8 => pow_8
# sum_8 => sum_8
# Graph fragment:
# %pow_8 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_13, 2), kwargs = {})
# %sum_8 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_8,), kwargs = {})
triton_per_fused_pow_sum_7 = async_compile.triton('triton_per_fused_pow_sum_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_pow_sum_7', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_pow_sum_7(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 3
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), rmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/yr/cyrcdmzymqmgy3uy4er6nkcwmlte6efc3shfpandhquk2eyvzs4u.py
# Topologically Sorted Source Nodes: [pow_11, sum_11], Original ATen: [aten.pow, aten.sum]
# Source node to ATen node mapping:
# pow_11 => pow_11
# sum_11 => sum_11
# Graph fragment:
# %pow_11 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_18, 2), kwargs = {})
# %sum_11 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_11,), kwargs = {})
triton_per_fused_pow_sum_8 = async_compile.triton('triton_per_fused_pow_sum_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 32],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_pow_sum_8', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_pow_sum_8(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), rmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/42/c42hwesai5nocof66i3vgu2h4yx5eyjkkhuw46lm4gp55cnlv5ry.py
# Topologically Sorted Source Nodes: [sub_45, sum_of_square_10, pow_12, sum_12, sum_of_square_11, mul_21, weights_regularizer_5], Original ATen: [aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div]
# Source node to ATen node mapping:
# mul_21 => mul_23
# pow_12 => pow_12
# sub_45 => sub_43
# sum_12 => sum_12
# sum_of_square_10 => add_50
# sum_of_square_11 => add_51
# weights_regularizer_5 => div_14
# Graph fragment:
# %sub_43 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_2), kwargs = {})
# %add_50 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_11, 0), kwargs = {})
# %pow_12 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_19, 2), kwargs = {})
# %sum_12 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_12,), kwargs = {})
# %add_51 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_50, %sum_12), kwargs = {})
# %mul_23 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_51, 4), kwargs = {})
# %div_14 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_23, %sub_43), kwargs = {})
# %div_114 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_14, %sub_43), kwargs = {})
triton_per_fused_add_div_mul_pow_rsub_sum_9 = async_compile.triton('triton_per_fused_add_div_mul_pow_rsub_sum_9', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 8],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_mul_pow_rsub_sum_9', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_9(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 6
RBLOCK: tl.constexpr = 8
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, 1])
tmp13 = tl.load(in_ptr2 + (2))
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp8 = 0.0
tmp9 = tmp7 + tmp8
tmp10 = tmp9 + tmp5
tmp11 = 4.0
tmp12 = tmp10 * tmp11
tmp15 = 1.0
tmp16 = tmp15 - tmp14
tmp17 = tmp12 / tmp16
tmp18 = tmp17 / tmp16
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp18, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ri/cridgry2xm245a23u75d65thgtjwllmoy3elpz33w7a4pwtrfnu6.py
# Topologically Sorted Source Nodes: [sub_72, sum_of_square_16, pow_18, sum_18, sum_of_square_17, mul_33, weights_regularizer_8], Original ATen: [aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div]
# Source node to ATen node mapping:
# mul_33 => mul_35
# pow_18 => pow_18
# sub_72 => sub_67
# sum_18 => sum_18
# sum_of_square_16 => add_77
# sum_of_square_17 => add_78
# weights_regularizer_8 => div_20
# Graph fragment:
# %sub_67 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_5), kwargs = {})
# %add_77 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_17, 0), kwargs = {})
# %pow_18 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_28, 2), kwargs = {})
# %sum_18 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_18,), kwargs = {})
# %add_78 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_77, %sum_18), kwargs = {})
# %mul_35 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_78, 4), kwargs = {})
# %div_20 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_35, %sub_67), kwargs = {})
# %div_90 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_20, %sub_67), kwargs = {})
triton_per_fused_add_div_mul_pow_rsub_sum_10 = async_compile.triton('triton_per_fused_add_div_mul_pow_rsub_sum_10', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 8],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_mul_pow_rsub_sum_10', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_10(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 6
RBLOCK: tl.constexpr = 8
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, 1])
tmp13 = tl.load(in_ptr2 + (5))
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp8 = 0.0
tmp9 = tmp7 + tmp8
tmp10 = tmp9 + tmp5
tmp11 = 4.0
tmp12 = tmp10 * tmp11
tmp15 = 1.0
tmp16 = tmp15 - tmp14
tmp17 = tmp12 / tmp16
tmp18 = tmp17 / tmp16
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp18, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zs/czs4gufsgrnyhhzlfgvfufobsmm6xraz7oweb72an5ufo7bjroyt.py
# Topologically Sorted Source Nodes: [pow_19, sum_19], Original ATen: [aten.pow, aten.sum]
# Source node to ATen node mapping:
# pow_19 => pow_19
# sum_19 => sum_19
# Graph fragment:
# %pow_19 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_30, 2), kwargs = {})
# %sum_19 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_19,), kwargs = {})
triton_per_fused_pow_sum_11 = async_compile.triton('triton_per_fused_pow_sum_11', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_pow_sum_11', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_pow_sum_11(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/dj/cdjdhwz5va5vtqe2nrqby3vvq5krnefikusczyguzrhiyp74dddp.py
# Topologically Sorted Source Nodes: [sub_27, pow_7, sum_7, sum_of_square_6, sum_of_square_7, mul_13, weights_regularizer_3, sub_36, pow_9, sum_9, sum_of_square_8, sum_of_square_9, mul_17, weights_regularizer_4, sub_54, pow_13, sum_13, sum_of_square_12, sum_of_square_13, mul_25, weights_regularizer_6, sub_63, pow_15, sum_15, sum_of_square_14, sum_of_square_15, mul_29, weights_regularizer_7, sum_25], Original ATen: [aten.rsub, aten.pow, aten.sum, aten.add, aten.mul, aten.div]
# Source node to ATen node mapping:
# mul_13 => mul_15
# mul_17 => mul_19
# mul_25 => mul_27
# mul_29 => mul_31
# pow_13 => pow_13
# pow_15 => pow_15
# pow_7 => pow_7
# pow_9 => pow_9
# sub_27 => sub_27
# sub_36 => sub_35
# sub_54 => sub_51
# sub_63 => sub_59
# sum_13 => sum_13
# sum_15 => sum_15
# sum_25 => sum_25
# sum_7 => sum_7
# sum_9 => sum_9
# sum_of_square_12 => add_59
# sum_of_square_13 => add_60
# sum_of_square_14 => add_68
# sum_of_square_15 => add_69
# sum_of_square_6 => add_32
# sum_of_square_7 => add_33
# sum_of_square_8 => add_41
# sum_of_square_9 => add_42
# weights_regularizer_3 => div_10
# weights_regularizer_4 => div_12
# weights_regularizer_6 => div_16
# weights_regularizer_7 => div_18
# Graph fragment:
# %select_scatter_default : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%empty, %view_3, 0, 0), kwargs = {})
# %select_scatter_default_1 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default, %view_7, 0, 1), kwargs = {})
# %select_scatter_default_2 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_1, %view_11, 0, 2), kwargs = {})
# %sub_27 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem), kwargs = {})
# %pow_7 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_12, 2), kwargs = {})
# %sum_7 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_7,), kwargs = {})
# %add_32 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_7, 0), kwargs = {})
# %add_33 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_32, %sum_8), kwargs = {})
# %mul_15 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_33, 4), kwargs = {})
# %div_10 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_15, %sub_27), kwargs = {})
# %select_scatter_default_3 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_2, %view_13, 0, 3), kwargs = {})
# %sub_35 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_1), kwargs = {})
# %pow_9 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_15, 2), kwargs = {})
# %sum_9 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_9,), kwargs = {})
# %add_41 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_9, 0), kwargs = {})
# %add_42 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_41, %sum_10), kwargs = {})
# %mul_19 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_42, 4), kwargs = {})
# %div_12 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_19, %sub_35), kwargs = {})
# %select_scatter_default_4 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_3, %view_14, 0, 4), kwargs = {})
# %select_scatter_default_5 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_4, %view_15, 0, 5), kwargs = {})
# %sub_51 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_3), kwargs = {})
# %pow_13 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_21, 2), kwargs = {})
# %sum_13 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_13,), kwargs = {})
# %add_59 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_13, 0), kwargs = {})
# %add_60 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_59, %sum_14), kwargs = {})
# %mul_27 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_60, 4), kwargs = {})
# %div_16 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_27, %sub_51), kwargs = {})
# %select_scatter_default_6 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_5, %view_16, 0, 6), kwargs = {})
# %sub_59 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_4), kwargs = {})
# %pow_15 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_24, 2), kwargs = {})
# %sum_15 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%pow_15,), kwargs = {})
# %add_68 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_15, 0), kwargs = {})
# %add_69 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_68, %sum_16), kwargs = {})
# %mul_31 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_69, 4), kwargs = {})
# %div_18 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_31, %sub_59), kwargs = {})
# %select_scatter_default_7 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_6, %view_17, 0, 7), kwargs = {})
# %select_scatter_default_8 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_7, %view_18, 0, 8), kwargs = {})
# %select_scatter_default_9 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_8, %view_19, 0, 9), kwargs = {})
# %select_scatter_default_10 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_9, %view_20, 0, 10), kwargs = {})
# %select_scatter_default_11 : [num_users=1] = call_function[target=torch.ops.aten.select_scatter.default](args = (%select_scatter_default_10, %view_49, 0, 11), kwargs = {})
# %sum_25 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%select_scatter_default_11,), kwargs = {})
# %div_98 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_18, %sub_59), kwargs = {})
# %div_106 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_16, %sub_51), kwargs = {})
# %div_122 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_12, %sub_35), kwargs = {})
# %div_130 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div_10, %sub_27), kwargs = {})
triton_per_fused_add_div_mul_pow_rsub_sum_12 = async_compile.triton('triton_per_fused_add_div_mul_pow_rsub_sum_12', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: '*fp32', 15: '*fp32', 16: '*fp32', 17: '*fp32', 18: '*fp32', 19: '*fp32', 20: '*fp32', 21: '*fp32', 22: '*fp32', 23: '*fp32', 24: '*fp32', 25: '*fp32', 26: 'i32', 27: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {26: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25), equal_to_1=(26,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_mul_pow_rsub_sum_12', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 36, 'num_reduction': 5, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_12(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10, in_ptr11, in_ptr12, in_ptr13, in_ptr14, in_ptr15, in_ptr16, in_ptr17, in_ptr18, in_ptr19, in_ptr20, out_ptr4, out_ptr5, out_ptr6, out_ptr7, out_ptr8, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 12
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + (r0), rmask, other=0.0)
tmp12 = tl.load(in_ptr2 + (r0), rmask, other=0.0)
tmp18 = tl.load(in_ptr3 + (r0), rmask, other=0.0)
tmp29 = tl.load(in_ptr4 + (0))
tmp30 = tl.broadcast_to(tmp29, [XBLOCK, RBLOCK])
tmp34 = tl.load(in_ptr5 + (0))
tmp35 = tl.broadcast_to(tmp34, [XBLOCK, RBLOCK])
tmp49 = tl.load(in_ptr6 + (0))
tmp50 = tl.broadcast_to(tmp49, [XBLOCK, RBLOCK])
tmp53 = tl.load(in_ptr7 + (0))
tmp54 = tl.broadcast_to(tmp53, [XBLOCK, RBLOCK])
tmp57 = tl.load(in_ptr8 + (0))
tmp58 = tl.broadcast_to(tmp57, [XBLOCK, RBLOCK])
tmp66 = tl.load(in_ptr9 + (0))
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = tl.load(in_ptr10 + (0))
tmp70 = tl.broadcast_to(tmp69, [XBLOCK, RBLOCK])
tmp73 = tl.load(in_ptr5 + (2))
tmp74 = tl.broadcast_to(tmp73, [XBLOCK, RBLOCK])
tmp87 = tl.load(in_ptr11 + (0))
tmp88 = tl.broadcast_to(tmp87, [XBLOCK, RBLOCK])
tmp91 = tl.load(in_ptr5 + (1))
tmp92 = tl.broadcast_to(tmp91, [XBLOCK, RBLOCK])
tmp107 = tl.load(in_ptr12 + (0))
tmp108 = tl.broadcast_to(tmp107, [XBLOCK, RBLOCK])
tmp111 = tl.load(in_ptr5 + (4))
tmp112 = tl.broadcast_to(tmp111, [XBLOCK, RBLOCK])
tmp125 = tl.load(in_ptr13 + (0))
tmp126 = tl.broadcast_to(tmp125, [XBLOCK, RBLOCK])
tmp129 = tl.load(in_ptr5 + (3))
tmp130 = tl.broadcast_to(tmp129, [XBLOCK, RBLOCK])
tmp144 = tl.load(in_ptr14 + (0))
tmp145 = tl.broadcast_to(tmp144, [XBLOCK, RBLOCK])
tmp147 = tl.load(in_ptr15 + (0))
tmp148 = tl.broadcast_to(tmp147, [XBLOCK, RBLOCK])
tmp152 = tl.load(in_ptr5 + (6))
tmp153 = tl.broadcast_to(tmp152, [XBLOCK, RBLOCK])
tmp165 = tl.load(in_ptr16 + (0))
tmp166 = tl.broadcast_to(tmp165, [XBLOCK, RBLOCK])
tmp168 = tl.load(in_ptr17 + (0))
tmp169 = tl.broadcast_to(tmp168, [XBLOCK, RBLOCK])
tmp172 = tl.load(in_ptr5 + (5))
tmp173 = tl.broadcast_to(tmp172, [XBLOCK, RBLOCK])
tmp187 = tl.load(in_ptr18 + (0))
tmp188 = tl.broadcast_to(tmp187, [XBLOCK, RBLOCK])
tmp191 = tl.load(in_ptr19 + (0))
tmp192 = tl.broadcast_to(tmp191, [XBLOCK, RBLOCK])
tmp194 = tl.load(in_ptr20 + (0))
tmp195 = tl.broadcast_to(tmp194, [XBLOCK, RBLOCK])
tmp199 = tl.load(in_ptr5 + (7))
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp216 = tl.broadcast_to(tmp107, [XBLOCK, 1])
tmp219 = tl.broadcast_to(tmp111, [XBLOCK, 1])
tmp223 = tl.broadcast_to(tmp125, [XBLOCK, 1])
tmp226 = tl.broadcast_to(tmp129, [XBLOCK, 1])
tmp230 = tl.broadcast_to(tmp87, [XBLOCK, 1])
tmp233 = tl.broadcast_to(tmp91, [XBLOCK, 1])
tmp237 = tl.broadcast_to(tmp29, [XBLOCK, 1])
tmp240 = tl.broadcast_to(tmp34, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp7 = tmp6 * tmp6
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp10 = tl.where(rmask, tmp8, 0)
tmp11 = tl.sum(tmp10, 1)[:, None]
tmp13 = tmp12 * tmp12
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, RBLOCK])
tmp16 = tl.where(rmask, tmp14, 0)
tmp17 = tl.sum(tmp16, 1)[:, None]
tmp19 = tmp18 * tmp18
tmp20 = tl.broadcast_to(tmp19, [XBLOCK, RBLOCK])
tmp22 = tl.where(rmask, tmp20, 0)
tmp23 = tl.sum(tmp22, 1)[:, None]
tmp24 = r0
tmp25 = tl.full([1, 1], 3, tl.int32)
tmp26 = tmp24 == tmp25
tmp27 = 0.0
tmp28 = tmp23 + tmp27
tmp31 = tmp28 + tmp30
tmp32 = 4.0
tmp33 = tmp31 * tmp32
tmp36 = 1.0
tmp37 = tmp36 - tmp35
tmp38 = tmp33 / tmp37
tmp39 = tl_math.log(tmp35)
tmp40 = tmp35 * tmp39
tmp41 = tl_math.log(tmp37)
tmp42 = tmp37 * tmp41
tmp43 = tmp40 + tmp42
tmp44 = 32.0
tmp45 = tmp43 * tmp44
tmp46 = tmp38 + tmp45
tmp47 = tl.full([1, 1], 2, tl.int32)
tmp48 = tmp24 == tmp47
tmp51 = tl.full([1, 1], 1, tl.int32)
tmp52 = tmp24 == tmp51
tmp55 = tl.full([1, 1], 0, tl.int32)
tmp56 = tmp24 == tmp55
tmp59 = float("nan")
tmp60 = tl.where(tmp56, tmp58, tmp59)
tmp61 = tl.where(tmp52, tmp54, tmp60)
tmp62 = tl.where(tmp48, tmp50, tmp61)
tmp63 = tl.where(tmp26, tmp46, tmp62)
tmp64 = tl.full([1, 1], 5, tl.int32)
tmp65 = tmp24 == tmp64
tmp68 = tmp67 + tmp27
tmp71 = tmp68 + tmp70
tmp72 = tmp71 * tmp32
tmp75 = tmp36 - tmp74
tmp76 = tmp72 / tmp75
tmp77 = tl_math.log(tmp74)
tmp78 = tmp74 * tmp77
tmp79 = tl_math.log(tmp75)
tmp80 = tmp75 * tmp79
tmp81 = tmp78 + tmp80
tmp82 = tmp81 * tmp44
tmp83 = tmp76 + tmp82
tmp84 = tl.full([1, 1], 4, tl.int32)
tmp85 = tmp24 == tmp84
tmp86 = tmp11 + tmp27
tmp89 = tmp86 + tmp88
tmp90 = tmp89 * tmp32
tmp93 = tmp36 - tmp92
tmp94 = tmp90 / tmp93
tmp95 = tl_math.log(tmp92)
tmp96 = tmp92 * tmp95
tmp97 = tl_math.log(tmp93)
tmp98 = tmp93 * tmp97
tmp99 = tmp96 + tmp98
tmp100 = tmp99 * tmp44
tmp101 = tmp94 + tmp100
tmp102 = tl.where(tmp85, tmp101, tmp63)
tmp103 = tl.where(tmp65, tmp83, tmp102)
tmp104 = tl.full([1, 1], 7, tl.int32)
tmp105 = tmp24 == tmp104
tmp106 = tmp17 + tmp27
tmp109 = tmp106 + tmp108
tmp110 = tmp109 * tmp32
tmp113 = tmp36 - tmp112
tmp114 = tmp110 / tmp113
tmp115 = tl_math.log(tmp112)
tmp116 = tmp112 * tmp115
tmp117 = tl_math.log(tmp113)
tmp118 = tmp113 * tmp117
tmp119 = tmp116 + tmp118
tmp120 = tmp119 * tmp44
tmp121 = tmp114 + tmp120
tmp122 = tl.full([1, 1], 6, tl.int32)
tmp123 = tmp24 == tmp122
tmp124 = tmp5 + tmp27
tmp127 = tmp124 + tmp126
tmp128 = tmp127 * tmp32
tmp131 = tmp36 - tmp130
tmp132 = tmp128 / tmp131
tmp133 = tl_math.log(tmp130)
tmp134 = tmp130 * tmp133
tmp135 = tl_math.log(tmp131)
tmp136 = tmp131 * tmp135
tmp137 = tmp134 + tmp136
tmp138 = tmp137 * tmp44
tmp139 = tmp132 + tmp138
tmp140 = tl.where(tmp123, tmp139, tmp103)
tmp141 = tl.where(tmp105, tmp121, tmp140)
tmp142 = tl.full([1, 1], 9, tl.int32)
tmp143 = tmp24 == tmp142
tmp146 = tmp145 + tmp27
tmp149 = tmp148 * tmp148
tmp150 = tmp146 + tmp149
tmp151 = tmp150 * tmp32
tmp154 = tmp36 - tmp153
tmp155 = tmp151 / tmp154
tmp156 = tl_math.log(tmp153)
tmp157 = tmp153 * tmp156
tmp158 = tl_math.log(tmp154)
tmp159 = tmp154 * tmp158
tmp160 = tmp157 + tmp159
tmp161 = tmp160 * tmp44
tmp162 = tmp155 + tmp161
tmp163 = tl.full([1, 1], 8, tl.int32)
tmp164 = tmp24 == tmp163
tmp167 = tmp166 + tmp27
tmp170 = tmp167 + tmp169
tmp171 = tmp170 * tmp32
tmp174 = tmp36 - tmp173
tmp175 = tmp171 / tmp174
tmp176 = tl_math.log(tmp173)
tmp177 = tmp173 * tmp176
tmp178 = tl_math.log(tmp174)
tmp179 = tmp174 * tmp178
tmp180 = tmp177 + tmp179
tmp181 = tmp180 * tmp44
tmp182 = tmp175 + tmp181
tmp183 = tl.where(tmp164, tmp182, tmp141)
tmp184 = tl.where(tmp143, tmp162, tmp183)
tmp185 = tl.full([1, 1], 11, tl.int32)
tmp186 = tmp24 == tmp185
tmp189 = tl.full([1, 1], 10, tl.int32)
tmp190 = tmp24 == tmp189
tmp193 = tmp192 + tmp27
tmp196 = tmp195 * tmp195
tmp197 = tmp193 + tmp196
tmp198 = tmp197 * tmp32
tmp201 = tmp36 - tmp200
tmp202 = tmp198 / tmp201
tmp203 = tl_math.log(tmp200)
tmp204 = tmp200 * tmp203
tmp205 = tl_math.log(tmp201)
tmp206 = tmp201 * tmp205
tmp207 = tmp204 + tmp206
tmp208 = tmp207 * tmp44
tmp209 = tmp202 + tmp208
tmp210 = tl.where(tmp190, tmp209, tmp184)
tmp211 = tl.where(tmp186, tmp188, tmp210)
tmp212 = tl.broadcast_to(tmp211, [XBLOCK, RBLOCK])
tmp214 = tl.where(rmask, tmp212, 0)
tmp215 = tl.sum(tmp214, 1)[:, None]
tmp217 = tmp106 + tmp216
tmp218 = tmp217 * tmp32
tmp220 = tmp36 - tmp219
tmp221 = tmp218 / tmp220
tmp222 = tmp221 / tmp220
tmp224 = tmp124 + tmp223
tmp225 = tmp224 * tmp32
tmp227 = tmp36 - tmp226
tmp228 = tmp225 / tmp227
tmp229 = tmp228 / tmp227
tmp231 = tmp86 + tmp230
tmp232 = tmp231 * tmp32
tmp234 = tmp36 - tmp233
tmp235 = tmp232 / tmp234
tmp236 = tmp235 / tmp234
tmp238 = tmp28 + tmp237
tmp239 = tmp238 * tmp32
tmp241 = tmp36 - tmp240
tmp242 = tmp239 / tmp241
tmp243 = tmp242 / tmp241
tl.store(out_ptr5 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp222, None)
tl.store(out_ptr6 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp229, None)
tl.store(out_ptr7 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp236, None)
tl.store(out_ptr8 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp243, None)
tl.store(out_ptr4 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp215, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/oq/coqmsyjhloguyy7mzq6xqosrkboymzslyzp7xjwrvac32ifw5u2p.py
# Topologically Sorted Source Nodes: [add_21, log_18, sub_27, add_22, log_19, sub_28, add_23, log_20, add_24, sub_29, add_25, log_21, drop_prob_6, truediv_6], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_21 => add_27
# add_22 => add_28
# add_23 => add_29
# add_24 => add_30
# add_25 => add_31
# drop_prob_6 => sub_30
# log_18 => log_18
# log_19 => log_19
# log_20 => log_20
# log_21 => log_21
# sub_27 => sub_27
# sub_28 => sub_28
# sub_29 => sub_29
# truediv_6 => div_9
# Graph fragment:
# %add_27 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-07), kwargs = {})
# %log_18 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_27,), kwargs = {})
# %sub_27 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem), kwargs = {})
# %add_28 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_27, 1e-07), kwargs = {})
# %log_19 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_28,), kwargs = {})
# %sub_28 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_18, %log_19), kwargs = {})
# %add_29 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_3, 1e-07), kwargs = {})
# %log_20 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_29,), kwargs = {})
# %add_30 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_28, %log_20), kwargs = {})
# %sub_29 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_3), kwargs = {})
# %add_31 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_29, 1e-07), kwargs = {})
# %log_21 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_31,), kwargs = {})
# %sub_30 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_30, %log_21), kwargs = {})
# %div_9 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_30, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_13 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_13', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_13', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_13(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/pv/cpvyxfqatwcudsrjlgbints7pg627ivbkfsarevqyxgr44ckrgyi.py
# Topologically Sorted Source Nodes: [add_28, log_24, sub_36, add_29, log_25, sub_37, add_30, log_26, add_31, sub_38, add_32, log_27, drop_prob_8, truediv_8], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_28 => add_36
# add_29 => add_37
# add_30 => add_38
# add_31 => add_39
# add_32 => add_40
# drop_prob_8 => sub_38
# log_24 => log_24
# log_25 => log_25
# log_26 => log_26
# log_27 => log_27
# sub_36 => sub_35
# sub_37 => sub_36
# sub_38 => sub_37
# truediv_8 => div_11
# Graph fragment:
# %add_36 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_1, 1e-07), kwargs = {})
# %log_24 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_36,), kwargs = {})
# %sub_35 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_1), kwargs = {})
# %add_37 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_35, 1e-07), kwargs = {})
# %log_25 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_37,), kwargs = {})
# %sub_36 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_24, %log_25), kwargs = {})
# %add_38 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_4, 1e-07), kwargs = {})
# %log_26 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_38,), kwargs = {})
# %add_39 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_36, %log_26), kwargs = {})
# %sub_37 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_4), kwargs = {})
# %add_40 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_37, 1e-07), kwargs = {})
# %log_27 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_40,), kwargs = {})
# %sub_38 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_39, %log_27), kwargs = {})
# %div_11 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_38, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_14 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_14', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_14', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_14(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (1))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/fm/cfmk2f3ravvmga7irng47525qqbr67cjqqfl2llqxdaidd4slbjd.py
# Topologically Sorted Source Nodes: [add_35, log_30, sub_45, add_36, log_31, sub_46, add_37, log_32, add_38, sub_47, add_39, log_33, drop_prob_10, truediv_10], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_35 => add_45
# add_36 => add_46
# add_37 => add_47
# add_38 => add_48
# add_39 => add_49
# drop_prob_10 => sub_46
# log_30 => log_30
# log_31 => log_31
# log_32 => log_32
# log_33 => log_33
# sub_45 => sub_43
# sub_46 => sub_44
# sub_47 => sub_45
# truediv_10 => div_13
# Graph fragment:
# %add_45 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-07), kwargs = {})
# %log_30 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_45,), kwargs = {})
# %sub_43 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_2), kwargs = {})
# %add_46 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_43, 1e-07), kwargs = {})
# %log_31 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_46,), kwargs = {})
# %sub_44 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_30, %log_31), kwargs = {})
# %add_47 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_5, 1e-07), kwargs = {})
# %log_32 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_47,), kwargs = {})
# %add_48 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_44, %log_32), kwargs = {})
# %sub_45 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_5), kwargs = {})
# %add_49 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_45, 1e-07), kwargs = {})
# %log_33 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_49,), kwargs = {})
# %sub_46 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_48, %log_33), kwargs = {})
# %div_13 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_46, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_15 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_15', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_15', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_15(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (2))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/di/cdirwn2ukwxdh2a3nou24fj4l2rchs36wnfpwym5xdmekuijp226.py
# Topologically Sorted Source Nodes: [add_42, log_36, sub_54, add_43, log_37, sub_55, add_44, log_38, add_45, sub_56, add_46, log_39, drop_prob_12, truediv_12], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_42 => add_54
# add_43 => add_55
# add_44 => add_56
# add_45 => add_57
# add_46 => add_58
# drop_prob_12 => sub_54
# log_36 => log_36
# log_37 => log_37
# log_38 => log_38
# log_39 => log_39
# sub_54 => sub_51
# sub_55 => sub_52
# sub_56 => sub_53
# truediv_12 => div_15
# Graph fragment:
# %add_54 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_3, 1e-07), kwargs = {})
# %log_36 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_54,), kwargs = {})
# %sub_51 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_3), kwargs = {})
# %add_55 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_51, 1e-07), kwargs = {})
# %log_37 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_55,), kwargs = {})
# %sub_52 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_36, %log_37), kwargs = {})
# %add_56 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_6, 1e-07), kwargs = {})
# %log_38 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_56,), kwargs = {})
# %add_57 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_52, %log_38), kwargs = {})
# %sub_53 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_6), kwargs = {})
# %add_58 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_53, 1e-07), kwargs = {})
# %log_39 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_58,), kwargs = {})
# %sub_54 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_57, %log_39), kwargs = {})
# %div_15 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_54, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_16 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_16', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_16', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_16(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (3))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zl/czldcaqd3p7spunsfhj5tn3qiabkrv5zzmrv2muwuhegjchnaqx6.py
# Topologically Sorted Source Nodes: [add_49, log_42, sub_63, add_50, log_43, sub_64, add_51, log_44, add_52, sub_65, add_53, log_45, drop_prob_14, truediv_14], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_49 => add_63
# add_50 => add_64
# add_51 => add_65
# add_52 => add_66
# add_53 => add_67
# drop_prob_14 => sub_62
# log_42 => log_42
# log_43 => log_43
# log_44 => log_44
# log_45 => log_45
# sub_63 => sub_59
# sub_64 => sub_60
# sub_65 => sub_61
# truediv_14 => div_17
# Graph fragment:
# %add_63 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_4, 1e-07), kwargs = {})
# %log_42 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_63,), kwargs = {})
# %sub_59 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_4), kwargs = {})
# %add_64 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_59, 1e-07), kwargs = {})
# %log_43 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_64,), kwargs = {})
# %sub_60 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_42, %log_43), kwargs = {})
# %add_65 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_7, 1e-07), kwargs = {})
# %log_44 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_65,), kwargs = {})
# %add_66 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_60, %log_44), kwargs = {})
# %sub_61 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_7), kwargs = {})
# %add_67 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_61, 1e-07), kwargs = {})
# %log_45 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_67,), kwargs = {})
# %sub_62 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_66, %log_45), kwargs = {})
# %div_17 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_62, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_17 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_17', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_17', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_17(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ve/cve2ojp6dw2uoupp6dvcy2z2p7x3ft3gwihdyamzwsqxcdh44vmc.py
# Topologically Sorted Source Nodes: [add_56, log_48, sub_72, add_57, log_49, sub_73, add_58, log_50, add_59, sub_74, add_60, log_51, drop_prob_16, truediv_16], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_56 => add_72
# add_57 => add_73
# add_58 => add_74
# add_59 => add_75
# add_60 => add_76
# drop_prob_16 => sub_70
# log_48 => log_48
# log_49 => log_49
# log_50 => log_50
# log_51 => log_51
# sub_72 => sub_67
# sub_73 => sub_68
# sub_74 => sub_69
# truediv_16 => div_19
# Graph fragment:
# %add_72 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_5, 1e-07), kwargs = {})
# %log_48 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_72,), kwargs = {})
# %sub_67 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_5), kwargs = {})
# %add_73 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_67, 1e-07), kwargs = {})
# %log_49 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_73,), kwargs = {})
# %sub_68 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_48, %log_49), kwargs = {})
# %add_74 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_8, 1e-07), kwargs = {})
# %log_50 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_74,), kwargs = {})
# %add_75 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_68, %log_50), kwargs = {})
# %sub_69 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_8), kwargs = {})
# %add_76 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_69, 1e-07), kwargs = {})
# %log_51 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_76,), kwargs = {})
# %sub_70 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_75, %log_51), kwargs = {})
# %div_19 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_70, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_18 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_18', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_18', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_18(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (5))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/dv/cdvczblwwzl7vx22llaqgu7l27h5ryk2igyz7o7vzyj2h2wpuhbr.py
# Topologically Sorted Source Nodes: [add_63, log_54, sub_81, add_64, log_55, sub_82, add_65, log_56, add_66, sub_83, add_67, log_57, drop_prob_18, truediv_18], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_63 => add_81
# add_64 => add_82
# add_65 => add_83
# add_66 => add_84
# add_67 => add_85
# drop_prob_18 => sub_78
# log_54 => log_54
# log_55 => log_55
# log_56 => log_56
# log_57 => log_57
# sub_81 => sub_75
# sub_82 => sub_76
# sub_83 => sub_77
# truediv_18 => div_21
# Graph fragment:
# %add_81 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_6, 1e-07), kwargs = {})
# %log_54 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_81,), kwargs = {})
# %sub_75 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_6), kwargs = {})
# %add_82 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_75, 1e-07), kwargs = {})
# %log_55 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_82,), kwargs = {})
# %sub_76 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_54, %log_55), kwargs = {})
# %add_83 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_9, 1e-07), kwargs = {})
# %log_56 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_83,), kwargs = {})
# %add_84 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_76, %log_56), kwargs = {})
# %sub_77 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_9), kwargs = {})
# %add_85 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_77, 1e-07), kwargs = {})
# %log_57 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_85,), kwargs = {})
# %sub_78 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_84, %log_57), kwargs = {})
# %div_21 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_78, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_19 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_19', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_19', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_19(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (6))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/bu/cbukwzy456kwupizkedavdwjwq2yotsyoyelboy57r6kiqkhuzrp.py
# Topologically Sorted Source Nodes: [add_70, log_60, sub_90, add_71, log_61, sub_91, add_72, log_62, add_73, sub_92, add_74, log_63, drop_prob_20, truediv_20], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_70 => add_90
# add_71 => add_91
# add_72 => add_92
# add_73 => add_93
# add_74 => add_94
# drop_prob_20 => sub_86
# log_60 => log_60
# log_61 => log_61
# log_62 => log_62
# log_63 => log_63
# sub_90 => sub_83
# sub_91 => sub_84
# sub_92 => sub_85
# truediv_20 => div_23
# Graph fragment:
# %add_90 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_7, 1e-07), kwargs = {})
# %log_60 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_90,), kwargs = {})
# %sub_83 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_7), kwargs = {})
# %add_91 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_83, 1e-07), kwargs = {})
# %log_61 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_91,), kwargs = {})
# %sub_84 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_60, %log_61), kwargs = {})
# %add_92 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_10, 1e-07), kwargs = {})
# %log_62 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_92,), kwargs = {})
# %add_93 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_84, %log_62), kwargs = {})
# %sub_85 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_10), kwargs = {})
# %add_94 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_85, 1e-07), kwargs = {})
# %log_63 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_94,), kwargs = {})
# %sub_86 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_93, %log_63), kwargs = {})
# %div_23 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_86, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_20 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_20', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_20', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_20(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (7))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7b/c7biwoo3uunmfxmxwnjem6pmhwqf3h4tnj4qkv7a4p7ino7xfhim.py
# Topologically Sorted Source Nodes: [add_77, log_66, sub_99, add_78, log_67, sub_100, add_79, log_68, add_80, sub_101, add_81, log_69, drop_prob_22, truediv_22], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
# Source node to ATen node mapping:
# add_77 => add_99
# add_78 => add_100
# add_79 => add_101
# add_80 => add_102
# add_81 => add_103
# drop_prob_22 => sub_94
# log_66 => log_66
# log_67 => log_67
# log_68 => log_68
# log_69 => log_69
# sub_100 => sub_92
# sub_101 => sub_93
# sub_99 => sub_91
# truediv_22 => div_25
# Graph fragment:
# %add_99 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_8, 1e-07), kwargs = {})
# %log_66 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_99,), kwargs = {})
# %sub_91 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_8), kwargs = {})
# %add_100 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_91, 1e-07), kwargs = {})
# %log_67 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_100,), kwargs = {})
# %sub_92 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%log_66, %log_67), kwargs = {})
# %add_101 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%rand_11, 1e-07), kwargs = {})
# %log_68 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_101,), kwargs = {})
# %add_102 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_92, %log_68), kwargs = {})
# %sub_93 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %rand_11), kwargs = {})
# %add_103 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_93, 1e-07), kwargs = {})
# %log_69 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add_103,), kwargs = {})
# %sub_94 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_102, %log_69), kwargs = {})
# %div_25 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_94, 0.1), kwargs = {})
triton_poi_fused_add_div_log_rsub_sub_21 = async_compile.triton('triton_poi_fused_add_div_log_rsub_sub_21', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_log_rsub_sub_21', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_21(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (8))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + (x0), tmp19, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ve/cve6bbmu3zgzd2l4enod45renr7qtdhz7ereyugmqisymx3ejt34.py
# Topologically Sorted Source Nodes: [sigmoid_25], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid_25 => sigmoid_7
# Graph fragment:
# %sigmoid_7 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%view_21,), kwargs = {})
triton_poi_fused_sigmoid_22 = async_compile.triton('triton_poi_fused_sigmoid_22', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_22', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_22(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/x5/cx5pxsptdjvwp3olpp3xz4ixq7wqxopnurob5t2kxqz2tkrh33gd.py
# Topologically Sorted Source Nodes: [input_6, sub_27, sub_36, sub_45, sub_54, sub_63, sub_72, sub_81, sub_90, sub_99, random_tensor_10, x_20, x_21, random_tensor_9, x_18, x_19, random_tensor_8, x_16, x_17, random_tensor_7, x_14, x_15, random_tensor_6, x_12, x_13, random_tensor_5, x_10, x_11, random_tensor_4, x_8, x_9, random_tensor_3, x_6, x_7, random_tensor_11, x_22, x_23], Original ATen: [aten.tanh, aten.rsub, aten.mul, aten.div]
# Source node to ATen node mapping:
# input_6 => tanh_2
# random_tensor_10 => sub_95
# random_tensor_11 => sub_103
# random_tensor_3 => sub_102
# random_tensor_4 => sub_101
# random_tensor_5 => sub_100
# random_tensor_6 => sub_99
# random_tensor_7 => sub_98
# random_tensor_8 => sub_97
# random_tensor_9 => sub_96
# sub_27 => sub_27
# sub_36 => sub_35
# sub_45 => sub_43
# sub_54 => sub_51
# sub_63 => sub_59
# sub_72 => sub_67
# sub_81 => sub_75
# sub_90 => sub_83
# sub_99 => sub_91
# x_10 => mul_52
# x_11 => div_31
# x_12 => mul_51
# x_13 => div_30
# x_14 => mul_50
# x_15 => div_29
# x_16 => mul_49
# x_17 => div_28
# x_18 => mul_48
# x_19 => div_27
# x_20 => mul_47
# x_21 => div_26
# x_22 => mul_55
# x_23 => div_34
# x_6 => mul_54
# x_7 => div_33
# x_8 => mul_53
# x_9 => div_32
# Graph fragment:
# %tanh_2 : [num_users=9] = call_function[target=torch.ops.aten.tanh.default](args = (%view_10,), kwargs = {})
# %sub_27 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem), kwargs = {})
# %sub_35 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_1), kwargs = {})
# %sub_43 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_2), kwargs = {})
# %sub_51 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_3), kwargs = {})
# %sub_59 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_4), kwargs = {})
# %sub_67 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_5), kwargs = {})
# %sub_75 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_6), kwargs = {})
# %sub_83 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_7), kwargs = {})
# %sub_91 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_8), kwargs = {})
# %sub_95 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_16), kwargs = {})
# %mul_47 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_95), kwargs = {})
# %div_26 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_47, %sub_83), kwargs = {})
# %sub_96 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_15), kwargs = {})
# %mul_48 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_96), kwargs = {})
# %div_27 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_48, %sub_75), kwargs = {})
# %sub_97 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_14), kwargs = {})
# %mul_49 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_97), kwargs = {})
# %div_28 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_49, %sub_67), kwargs = {})
# %sub_98 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_13), kwargs = {})
# %mul_50 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_98), kwargs = {})
# %div_29 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_50, %sub_59), kwargs = {})
# %sub_99 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_12), kwargs = {})
# %mul_51 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_99), kwargs = {})
# %div_30 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_51, %sub_51), kwargs = {})
# %sub_100 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_11), kwargs = {})
# %mul_52 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_100), kwargs = {})
# %div_31 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_52, %sub_43), kwargs = {})
# %sub_101 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_10), kwargs = {})
# %mul_53 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_101), kwargs = {})
# %div_32 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_53, %sub_35), kwargs = {})
# %sub_102 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_9), kwargs = {})
# %mul_54 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_102), kwargs = {})
# %div_33 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_54, %sub_27), kwargs = {})
# %sub_103 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %getitem_17), kwargs = {})
# %mul_55 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, %sub_103), kwargs = {})
# %div_34 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_55, %sub_91), kwargs = {})
triton_poi_fused_div_mul_rsub_tanh_23 = async_compile.triton('triton_poi_fused_div_mul_rsub_tanh_23', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: '*fp32', 15: '*fp32', 16: '*fp32', 17: '*fp32', 18: '*fp32', 19: '*fp32', 20: '*fp32', 21: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_mul_rsub_tanh_23', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 19, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_mul_rsub_tanh_23(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7, out_ptr8, out_ptr9, out_ptr10, out_ptr11, out_ptr12, out_ptr13, out_ptr14, out_ptr15, out_ptr16, out_ptr17, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (1792 + x0), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask)
tmp6 = tl.load(in_ptr2 + (7))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp10 = tl.load(in_ptr0 + (1536 + x0), xmask)
tmp13 = tl.load(in_ptr2 + (6))
tmp14 = tl.broadcast_to(tmp13, [XBLOCK])
tmp17 = tl.load(in_ptr0 + (1280 + x0), xmask)
tmp20 = tl.load(in_ptr2 + (5))
tmp21 = tl.broadcast_to(tmp20, [XBLOCK])
tmp24 = tl.load(in_ptr0 + (1024 + x0), xmask)
tmp27 = tl.load(in_ptr2 + (4))
tmp28 = tl.broadcast_to(tmp27, [XBLOCK])
tmp31 = tl.load(in_ptr0 + (768 + x0), xmask)
tmp34 = tl.load(in_ptr2 + (3))
tmp35 = tl.broadcast_to(tmp34, [XBLOCK])
tmp38 = tl.load(in_ptr0 + (512 + x0), xmask)
tmp41 = tl.load(in_ptr2 + (2))
tmp42 = tl.broadcast_to(tmp41, [XBLOCK])
tmp45 = tl.load(in_ptr0 + (256 + x0), xmask)
tmp48 = tl.load(in_ptr2 + (1))
tmp49 = tl.broadcast_to(tmp48, [XBLOCK])
tmp52 = tl.load(in_ptr0 + (x0), xmask)
tmp55 = tl.load(in_ptr2 + (0))
tmp56 = tl.broadcast_to(tmp55, [XBLOCK])
tmp59 = tl.load(in_ptr0 + (2048 + x0), xmask)
tmp62 = tl.load(in_ptr2 + (8))
tmp63 = tl.broadcast_to(tmp62, [XBLOCK])
tmp1 = 1.0
tmp2 = tmp1 - tmp0
tmp4 = libdevice.tanh(tmp3)
tmp5 = tmp4 * tmp2
tmp8 = tmp1 - tmp7
tmp9 = tmp5 / tmp8
tmp11 = tmp1 - tmp10
tmp12 = tmp4 * tmp11
tmp15 = tmp1 - tmp14
tmp16 = tmp12 / tmp15
tmp18 = tmp1 - tmp17
tmp19 = tmp4 * tmp18
tmp22 = tmp1 - tmp21
tmp23 = tmp19 / tmp22
tmp25 = tmp1 - tmp24
tmp26 = tmp4 * tmp25
tmp29 = tmp1 - tmp28
tmp30 = tmp26 / tmp29
tmp32 = tmp1 - tmp31
tmp33 = tmp4 * tmp32
tmp36 = tmp1 - tmp35
tmp37 = tmp33 / tmp36
tmp39 = tmp1 - tmp38
tmp40 = tmp4 * tmp39
tmp43 = tmp1 - tmp42
tmp44 = tmp40 / tmp43
tmp46 = tmp1 - tmp45
tmp47 = tmp4 * tmp46
tmp50 = tmp1 - tmp49
tmp51 = tmp47 / tmp50
tmp53 = tmp1 - tmp52
tmp54 = tmp4 * tmp53
tmp57 = tmp1 - tmp56
tmp58 = tmp54 / tmp57
tmp60 = tmp1 - tmp59
tmp61 = tmp4 * tmp60
tmp64 = tmp1 - tmp63
tmp65 = tmp61 / tmp64
tl.store(out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr1 + (x0), tmp9, xmask)
tl.store(out_ptr2 + (x0), tmp11, xmask)
tl.store(out_ptr3 + (x0), tmp16, xmask)
tl.store(out_ptr4 + (x0), tmp18, xmask)
tl.store(out_ptr5 + (x0), tmp23, xmask)
tl.store(out_ptr6 + (x0), tmp25, xmask)
tl.store(out_ptr7 + (x0), tmp30, xmask)
tl.store(out_ptr8 + (x0), tmp32, xmask)
tl.store(out_ptr9 + (x0), tmp37, xmask)
tl.store(out_ptr10 + (x0), tmp39, xmask)
tl.store(out_ptr11 + (x0), tmp44, xmask)
tl.store(out_ptr12 + (x0), tmp46, xmask)
tl.store(out_ptr13 + (x0), tmp51, xmask)
tl.store(out_ptr14 + (x0), tmp53, xmask)
tl.store(out_ptr15 + (x0), tmp58, xmask)
tl.store(out_ptr16 + (x0), tmp60, xmask)
tl.store(out_ptr17 + (x0), tmp65, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30, primals_31, primals_32, primals_33, primals_34, primals_35, primals_36, primals_37 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (1, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (1, ), (1, ))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (1, ), (1, ))
assert_size_stride(primals_12, (3, 4), (4, 1))
assert_size_stride(primals_13, (3, ), (1, ))
assert_size_stride(primals_14, (1, ), (1, ))
assert_size_stride(primals_15, (3, 4), (4, 1))
assert_size_stride(primals_16, (3, ), (1, ))
assert_size_stride(primals_17, (1, ), (1, ))
assert_size_stride(primals_18, (6, 4), (4, 1))
assert_size_stride(primals_19, (6, ), (1, ))
assert_size_stride(primals_20, (1, ), (1, ))
assert_size_stride(primals_21, (3, 4), (4, 1))
assert_size_stride(primals_22, (3, ), (1, ))
assert_size_stride(primals_23, (1, ), (1, ))
assert_size_stride(primals_24, (3, 4), (4, 1))
assert_size_stride(primals_25, (3, ), (1, ))
assert_size_stride(primals_26, (1, ), (1, ))
assert_size_stride(primals_27, (6, 4), (4, 1))
assert_size_stride(primals_28, (6, ), (1, ))
assert_size_stride(primals_29, (1, ), (1, ))
assert_size_stride(primals_30, (1, 4), (4, 1))
assert_size_stride(primals_31, (1, ), (1, ))
assert_size_stride(primals_32, (1, ), (1, ))
assert_size_stride(primals_33, (1, 4), (4, 1))
assert_size_stride(primals_34, (1, ), (1, ))
assert_size_stride(primals_35, (1, ), (1, ))
assert_size_stride(primals_36, (1, 4), (4, 1))
assert_size_stride(primals_37, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [unif_noise], Original ATen: [aten.rand_like]
buf1 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf2 = buf1
del buf1
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [p, add, log, sub, add_1, log_1, sub_1, add_2, log_2, add_3, sub_2, add_4, log_3, drop_prob, truediv, drop_prob_1, random_tensor, x, x_1], Original ATen: [aten.sigmoid, aten.add, aten.log, aten.rsub, aten.sub, aten.div, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0.run(primals_1, primals_2, buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, reinterpret_tensor(buf3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf4)
buf5 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_1, sum_1], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_1.run(primals_3, buf5, 1, 16, grid=grid(1), stream=stream0)
buf7 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf8 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf9 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf131 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [p, sub, sum_of_square, pow_2, sum_2, sum_of_square_1, mul_1, weights_regularizer, log_4, dropout_regularizer, sub_7, log_5, mul_3, dropout_regularizer_1, dropout_regularizer_2, regularization_1], Original ATen: [aten.sigmoid, aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div, aten.log]
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2.run(primals_4, primals_2, buf5, buf7, buf8, buf9, buf131, 1, 4, grid=grid(1), stream=stream0)
# Topologically Sorted Source Nodes: [unif_noise_1], Original ATen: [aten.rand_like]
buf10 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf11 = buf10
del buf10
buf12 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_2, p_1, add_7, log_6, sub_9, add_8, log_7, sub_10, add_9, log_8, add_10, sub_11, add_11, log_9, drop_prob_2, truediv_2, drop_prob_3, random_tensor_1, x_2, x_3], Original ATen: [aten.tanh, aten.sigmoid, aten.add, aten.log, aten.rsub, aten.sub, aten.div, aten.mul]
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3.run(buf4, primals_5, buf11, buf12, 256, grid=grid(256), stream=stream0)
buf13 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf12, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf13)
buf14 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [pow_3, sum_3], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_1.run(primals_6, buf14, 1, 16, grid=grid(1), stream=stream0)
buf16 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf17 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf18 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf130 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [p_1, sub_9, sum_of_square_2, pow_4, sum_4, sum_of_square_3, mul_5, weights_regularizer_1, log_10, dropout_regularizer_3, sub_16, log_11, mul_7, dropout_regularizer_4, dropout_regularizer_5, regularization_2], Original ATen: [aten.sigmoid, aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div, aten.log]
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2.run(primals_7, primals_5, buf14, buf16, buf17, buf18, buf130, 1, 4, grid=grid(1), stream=stream0)
# Topologically Sorted Source Nodes: [unif_noise_2], Original ATen: [aten.rand_like]
buf19 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf20 = buf19
del buf19
buf21 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_4, p_2, add_14, log_12, sub_18, add_15, log_13, sub_19, add_16, log_14, add_17, sub_20, add_18, log_15, drop_prob_4, truediv_4, drop_prob_5, random_tensor_2, x_4, x_5], Original ATen: [aten.tanh, aten.sigmoid, aten.add, aten.log, aten.rsub, aten.sub, aten.div, aten.mul]
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3.run(buf13, primals_8, buf20, buf21, 256, grid=grid(256), stream=stream0)
buf22 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_10, reinterpret_tensor(buf21, (64, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf22)
buf23 = buf14; del buf14 # reuse
# Topologically Sorted Source Nodes: [pow_5, sum_5], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_1.run(primals_9, buf23, 1, 16, grid=grid(1), stream=stream0)
buf25 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf26 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf27 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf129 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [p_2, sub_18, sum_of_square_4, pow_6, sum_6, sum_of_square_5, mul_9, weights_regularizer_2, log_16, dropout_regularizer_6, sub_25, log_17, mul_11, dropout_regularizer_7, dropout_regularizer_8, regularization_3], Original ATen: [aten.sigmoid, aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div, aten.log]
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2.run(primals_10, primals_8, buf23, buf25, buf26, buf27, buf129, 1, 4, grid=grid(1), stream=stream0)
buf37 = empty_strided_cuda((9, ), (1, ), torch.float32)
buf28 = reinterpret_tensor(buf37, (1, ), (1, ), 0) # alias
buf29 = reinterpret_tensor(buf37, (1, ), (1, ), 1) # alias
buf30 = reinterpret_tensor(buf37, (1, ), (1, ), 2) # alias
buf31 = reinterpret_tensor(buf37, (1, ), (1, ), 3) # alias
buf32 = reinterpret_tensor(buf37, (1, ), (1, ), 4) # alias
buf33 = reinterpret_tensor(buf37, (1, ), (1, ), 5) # alias
buf34 = reinterpret_tensor(buf37, (1, ), (1, ), 6) # alias
buf35 = reinterpret_tensor(buf37, (1, ), (1, ), 7) # alias
buf36 = reinterpret_tensor(buf37, (1, ), (1, ), 8) # alias
# Unsorted Source Nodes: [], Original ATen: []
triton_for_fused_4.run(primals_11, primals_14, primals_17, primals_20, primals_23, primals_26, primals_29, primals_32, primals_35, buf28, buf29, buf30, buf31, buf32, buf33, buf34, buf35, buf36, grid=(9, 1, 1), stream=stream0)
del primals_11
del primals_14
del primals_17
del primals_20
del primals_23
del primals_26
del primals_29
del primals_32
del primals_35
buf38 = empty_strided_cuda((9, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [sigmoid_24], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_5.run(buf37, buf38, 9, grid=grid(9), stream=stream0)
del buf28
del buf29
del buf30
del buf31
del buf32
del buf33
del buf34
del buf35
del buf36
del buf37
# Topologically Sorted Source Nodes: [unif_noise_3], Original ATen: [aten.rand_like]
buf39 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf40 = buf39
del buf39
buf117 = reinterpret_tensor(buf23, (1, ), (1, ), 0); del buf23 # reuse
buf118 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf119 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf122 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [sub_99, pow_23, sum_23, sum_of_square_22, pow_24, sum_24, sum_of_square_23, mul_45, weights_regularizer_11, log_70, dropout_regularizer_33, sub_106, log_71, mul_47, dropout_regularizer_34, dropout_regularizer_35, regularization_12], Original ATen: [aten.rsub, aten.pow, aten.sum, aten.add, aten.mul, aten.div, aten.log]
triton_per_fused_add_div_log_mul_pow_rsub_sum_6.run(primals_36, buf38, primals_37, buf117, buf118, buf119, buf122, 1, 4, grid=grid(1), stream=stream0)
buf42 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_8, sum_8], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_7.run(primals_13, buf42, 1, 3, grid=grid(1), stream=stream0)
buf47 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_10, sum_10], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_7.run(primals_16, buf47, 1, 3, grid=grid(1), stream=stream0)
buf50 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_11, sum_11], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_8.run(primals_18, buf50, 1, 24, grid=grid(1), stream=stream0)
buf51 = empty_strided_cuda((), (), torch.float32)
buf126 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [sub_45, sum_of_square_10, pow_12, sum_12, sum_of_square_11, mul_21, weights_regularizer_5], Original ATen: [aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div]
triton_per_fused_add_div_mul_pow_rsub_sum_9.run(primals_19, buf50, buf38, buf51, buf126, 1, 6, grid=grid(1), stream=stream0)
buf56 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_14, sum_14], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_7.run(primals_22, buf56, 1, 3, grid=grid(1), stream=stream0)
buf60 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_16, sum_16], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_7.run(primals_25, buf60, 1, 3, grid=grid(1), stream=stream0)
buf64 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_17, sum_17], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_8.run(primals_27, buf64, 1, 24, grid=grid(1), stream=stream0)
buf65 = empty_strided_cuda((), (), torch.float32)
buf123 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [sub_72, sum_of_square_16, pow_18, sum_18, sum_of_square_17, mul_33, weights_regularizer_8], Original ATen: [aten.rsub, aten.add, aten.pow, aten.sum, aten.mul, aten.div]
triton_per_fused_add_div_mul_pow_rsub_sum_10.run(primals_28, buf64, buf38, buf65, buf123, 1, 6, grid=grid(1), stream=stream0)
buf68 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_19, sum_19], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_11.run(primals_30, buf68, 1, 4, grid=grid(1), stream=stream0)
buf72 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [pow_21, sum_21], Original ATen: [aten.pow, aten.sum]
triton_per_fused_pow_sum_11.run(primals_33, buf72, 1, 4, grid=grid(1), stream=stream0)
buf121 = empty_strided_cuda((), (), torch.float32)
buf124 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf125 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf127 = empty_strided_cuda((1, ), (1, ), torch.float32)
buf128 = empty_strided_cuda((1, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [sub_27, pow_7, sum_7, sum_of_square_6, sum_of_square_7, mul_13, weights_regularizer_3, sub_36, pow_9, sum_9, sum_of_square_8, sum_of_square_9, mul_17, weights_regularizer_4, sub_54, pow_13, sum_13, sum_of_square_12, sum_of_square_13, mul_25, weights_regularizer_6, sub_63, pow_15, sum_15, sum_of_square_14, sum_of_square_15, mul_29, weights_regularizer_7, sum_25], Original ATen: [aten.rsub, aten.pow, aten.sum, aten.add, aten.mul, aten.div]
triton_per_fused_add_div_mul_pow_rsub_sum_12.run(primals_21, primals_15, primals_24, primals_12, buf42, buf38, buf27, buf18, buf9, buf50, buf51, buf47, buf60, buf56, buf68, primals_31, buf64, buf65, buf119, buf72, primals_34, buf121, buf124, buf125, buf127, buf128, 1, 12, grid=grid(1), stream=stream0)
del buf119
del buf18
del buf27
del buf42
del buf47
del buf50
del buf51
del buf56
del buf60
del buf64
del buf65
del buf68
del buf72
del buf9
# Topologically Sorted Source Nodes: [unif_noise_4], Original ATen: [aten.rand_like]
buf44 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf45 = buf44
del buf44
# Topologically Sorted Source Nodes: [unif_noise_5], Original ATen: [aten.rand_like]
buf48 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf49 = buf48
del buf48
# Topologically Sorted Source Nodes: [unif_noise_6], Original ATen: [aten.rand_like]
buf53 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf54 = buf53
del buf53
# Topologically Sorted Source Nodes: [unif_noise_7], Original ATen: [aten.rand_like]
buf57 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf58 = buf57
del buf57
# Topologically Sorted Source Nodes: [unif_noise_8], Original ATen: [aten.rand_like]
buf62 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf63 = buf62
del buf62
# Topologically Sorted Source Nodes: [unif_noise_9], Original ATen: [aten.rand_like]
buf66 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf67 = buf66
del buf66
# Topologically Sorted Source Nodes: [unif_noise_10], Original ATen: [aten.rand_like]
buf70 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf71 = buf70
del buf70
# Topologically Sorted Source Nodes: [unif_noise_11], Original ATen: [aten.rand_like]
buf73 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf74 = buf73
del buf73
buf84 = empty_strided_cuda((36, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf75 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 0) # alias
# Topologically Sorted Source Nodes: [add_21, log_18, sub_27, add_22, log_19, sub_28, add_23, log_20, add_24, sub_29, add_25, log_21, drop_prob_6, truediv_6], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_13.run(buf38, buf40, buf75, 256, grid=grid(256), stream=stream0)
buf76 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 256) # alias
# Topologically Sorted Source Nodes: [add_28, log_24, sub_36, add_29, log_25, sub_37, add_30, log_26, add_31, sub_38, add_32, log_27, drop_prob_8, truediv_8], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_14.run(buf38, buf45, buf76, 256, grid=grid(256), stream=stream0)
buf77 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 512) # alias
# Topologically Sorted Source Nodes: [add_35, log_30, sub_45, add_36, log_31, sub_46, add_37, log_32, add_38, sub_47, add_39, log_33, drop_prob_10, truediv_10], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_15.run(buf38, buf49, buf77, 256, grid=grid(256), stream=stream0)
buf78 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 768) # alias
# Topologically Sorted Source Nodes: [add_42, log_36, sub_54, add_43, log_37, sub_55, add_44, log_38, add_45, sub_56, add_46, log_39, drop_prob_12, truediv_12], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_16.run(buf38, buf54, buf78, 256, grid=grid(256), stream=stream0)
buf79 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1024) # alias
# Topologically Sorted Source Nodes: [add_49, log_42, sub_63, add_50, log_43, sub_64, add_51, log_44, add_52, sub_65, add_53, log_45, drop_prob_14, truediv_14], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_17.run(buf38, buf58, buf79, 256, grid=grid(256), stream=stream0)
buf80 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1280) # alias
# Topologically Sorted Source Nodes: [add_56, log_48, sub_72, add_57, log_49, sub_73, add_58, log_50, add_59, sub_74, add_60, log_51, drop_prob_16, truediv_16], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_18.run(buf38, buf63, buf80, 256, grid=grid(256), stream=stream0)
buf81 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1536) # alias
# Topologically Sorted Source Nodes: [add_63, log_54, sub_81, add_64, log_55, sub_82, add_65, log_56, add_66, sub_83, add_67, log_57, drop_prob_18, truediv_18], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_19.run(buf38, buf67, buf81, 256, grid=grid(256), stream=stream0)
buf82 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1792) # alias
# Topologically Sorted Source Nodes: [add_70, log_60, sub_90, add_71, log_61, sub_91, add_72, log_62, add_73, sub_92, add_74, log_63, drop_prob_20, truediv_20], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_20.run(buf38, buf71, buf82, 256, grid=grid(256), stream=stream0)
buf83 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 2048) # alias
# Topologically Sorted Source Nodes: [add_77, log_66, sub_99, add_78, log_67, sub_100, add_79, log_68, add_80, sub_101, add_81, log_69, drop_prob_22, truediv_22], Original ATen: [aten.add, aten.log, aten.rsub, aten.sub, aten.div]
triton_poi_fused_add_div_log_rsub_sub_21.run(buf38, buf74, buf83, 256, grid=grid(256), stream=stream0)
buf85 = empty_strided_cuda((9, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sigmoid_25], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_22.run(buf84, buf85, 2304, grid=grid(2304), stream=stream0)
del buf75
del buf76
del buf77
del buf78
del buf79
del buf80
del buf81
del buf82
del buf83
del buf84
buf86 = buf74; del buf74 # reuse
buf87 = buf71; del buf71 # reuse
buf90 = buf67; del buf67 # reuse
buf91 = buf63; del buf63 # reuse
buf94 = buf58; del buf58 # reuse
buf95 = buf54; del buf54 # reuse
buf97 = buf49; del buf49 # reuse
buf98 = buf45; del buf45 # reuse
buf100 = buf40; del buf40 # reuse
buf101 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf103 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf104 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf106 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf107 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf109 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf110 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf112 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf113 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [input_6, sub_27, sub_36, sub_45, sub_54, sub_63, sub_72, sub_81, sub_90, sub_99, random_tensor_10, x_20, x_21, random_tensor_9, x_18, x_19, random_tensor_8, x_16, x_17, random_tensor_7, x_14, x_15, random_tensor_6, x_12, x_13, random_tensor_5, x_10, x_11, random_tensor_4, x_8, x_9, random_tensor_3, x_6, x_7, random_tensor_11, x_22, x_23], Original ATen: [aten.tanh, aten.rsub, aten.mul, aten.div]
triton_poi_fused_div_mul_rsub_tanh_23.run(buf85, buf22, buf38, buf86, buf87, buf90, buf91, buf94, buf95, buf97, buf98, buf100, buf101, buf103, buf104, buf106, buf107, buf109, buf110, buf112, buf113, 256, grid=grid(256), stream=stream0)
buf89 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_7], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_34, reinterpret_tensor(buf87, (64, 4), (4, 1), 0), reinterpret_tensor(primals_33, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf89)
buf93 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_6], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_31, reinterpret_tensor(buf91, (64, 4), (4, 1), 0), reinterpret_tensor(primals_30, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf93)
buf96 = empty_strided_cuda((64, 6), (6, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_28, reinterpret_tensor(buf95, (64, 4), (4, 1), 0), reinterpret_tensor(primals_27, (4, 6), (1, 4), 0), alpha=1, beta=1, out=buf96)
buf99 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_4], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_25, reinterpret_tensor(buf98, (64, 4), (4, 1), 0), reinterpret_tensor(primals_24, (4, 3), (1, 4), 0), alpha=1, beta=1, out=buf99)
buf102 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_22, reinterpret_tensor(buf101, (64, 4), (4, 1), 0), reinterpret_tensor(primals_21, (4, 3), (1, 4), 0), alpha=1, beta=1, out=buf102)
buf105 = empty_strided_cuda((64, 6), (6, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_19, reinterpret_tensor(buf104, (64, 4), (4, 1), 0), reinterpret_tensor(primals_18, (4, 6), (1, 4), 0), alpha=1, beta=1, out=buf105)
buf108 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_16, reinterpret_tensor(buf107, (64, 4), (4, 1), 0), reinterpret_tensor(primals_15, (4, 3), (1, 4), 0), alpha=1, beta=1, out=buf108)
buf111 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_13, reinterpret_tensor(buf110, (64, 4), (4, 1), 0), reinterpret_tensor(primals_12, (4, 3), (1, 4), 0), alpha=1, beta=1, out=buf111)
buf115 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_8], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_37, reinterpret_tensor(buf113, (64, 4), (4, 1), 0), reinterpret_tensor(primals_36, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf115)
return (reinterpret_tensor(buf111, (4, 4, 4, 3), (48, 12, 3, 1), 0), reinterpret_tensor(buf108, (4, 4, 4, 3), (48, 12, 3, 1), 0), reinterpret_tensor(buf105, (4, 4, 4, 6), (96, 24, 6, 1), 0), reinterpret_tensor(buf102, (4, 4, 4, 3), (48, 12, 3, 1), 0), reinterpret_tensor(buf99, (4, 4, 4, 3), (48, 12, 3, 1), 0), reinterpret_tensor(buf96, (4, 4, 4, 6), (96, 24, 6, 1), 0), reinterpret_tensor(buf93, (4, 4, 4, 1), (16, 4, 1, 1), 0), reinterpret_tensor(buf89, (4, 4, 4, 1), (16, 4, 1, 1), 0), reinterpret_tensor(buf115, (4, 4, 4, 1), (16, 4, 1, 1), 0), buf121, primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_12, primals_13, primals_15, primals_16, primals_18, primals_19, primals_21, primals_22, primals_24, primals_25, primals_27, primals_28, primals_30, primals_31, primals_33, primals_34, primals_36, primals_37, buf2, buf3, buf4, buf7, buf8, buf11, buf12, buf13, buf16, buf17, buf20, buf21, buf22, buf25, buf26, buf38, reinterpret_tensor(buf38, (1, ), (1, ), 0), reinterpret_tensor(buf38, (1, ), (1, ), 1), reinterpret_tensor(buf38, (1, ), (1, ), 2), reinterpret_tensor(buf38, (1, ), (1, ), 3), reinterpret_tensor(buf38, (1, ), (1, ), 4), reinterpret_tensor(buf38, (1, ), (1, ), 5), reinterpret_tensor(buf38, (1, ), (1, ), 6), reinterpret_tensor(buf38, (1, ), (1, ), 7), reinterpret_tensor(buf38, (1, ), (1, ), 8), buf85, buf86, buf87, buf90, buf91, buf94, buf95, buf97, buf98, buf100, buf101, buf103, buf104, buf106, buf107, buf109, buf110, buf112, buf113, buf117, buf118, buf122, buf123, buf124, buf125, buf126, buf127, buf128, buf129, buf130, buf131, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((3, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((3, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((3, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((3, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((6, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((6, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((3, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((3, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_23 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_24 = rand_strided((3, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_25 = rand_strided((3, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_26 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_27 = rand_strided((6, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_28 = rand_strided((6, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_29 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_30 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_31 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_32 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_33 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_34 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_35 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_36 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_37 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30, primals_31, primals_32, primals_33, primals_34, primals_35, primals_36, primals_37])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import numpy as np
from torch import nn
class ConcreteDropout(nn.Module):
def __init__(self, weight_regularizer=1e-06, dropout_regularizer=1e-05,
init_min=0.1, init_max=0.1):
super(ConcreteDropout, self).__init__()
self.weight_regularizer = weight_regularizer
self.dropout_regularizer = dropout_regularizer
init_min = np.log(init_min) - np.log(1.0 - init_min)
init_max = np.log(init_max) - np.log(1.0 - init_max)
self.p_logit = nn.Parameter(torch.empty(1).uniform_(init_min, init_max)
)
def forward(self, x, layer):
p = torch.sigmoid(self.p_logit)
out = layer(self._concrete_dropout(x, p))
sum_of_square = 0
for param in layer.parameters():
sum_of_square += torch.sum(torch.pow(param, 2))
weights_regularizer = self.weight_regularizer * sum_of_square / (1 - p)
dropout_regularizer = p * torch.log(p)
dropout_regularizer += (1.0 - p) * torch.log(1.0 - p)
input_dimensionality = x[0].numel()
dropout_regularizer *= self.dropout_regularizer * input_dimensionality
regularization = weights_regularizer + dropout_regularizer
return out, regularization
def _concrete_dropout(self, x, p):
eps = 1e-07
temp = 0.1
unif_noise = torch.rand_like(x)
drop_prob = torch.log(p + eps) - torch.log(1 - p + eps) + torch.log(
unif_noise + eps) - torch.log(1 - unif_noise + eps)
drop_prob = torch.sigmoid(drop_prob / temp)
random_tensor = 1 - drop_prob
retain_prob = 1 - p
x = torch.mul(x, random_tensor)
x /= retain_prob
return x
class ConcreteDenseMixture(nn.Module):
def __init__(self, X_dim, Y_dim, nb_features, weight_regularizer,
dropout_regularizer, verbose=True):
super(ConcreteDenseMixture, self).__init__()
self.verbose = verbose
self.rank = 2
self.linear1 = nn.Linear(X_dim, nb_features)
self.linear2 = nn.Linear(nb_features, nb_features)
self.linear3 = nn.Linear(nb_features, nb_features)
self.linear4_mu = nn.Linear(nb_features, Y_dim - 1)
self.linear4_logvar = nn.Linear(nb_features, Y_dim - 1)
self.linear4_F = nn.Linear(nb_features, (Y_dim - 1) * self.rank)
self.linear4_mu2 = nn.Linear(nb_features, Y_dim - 1)
self.linear4_logvar2 = nn.Linear(nb_features, Y_dim - 1)
self.linear4_F2 = nn.Linear(nb_features, (Y_dim - 1) * self.rank)
self.linear4_alpha = nn.Linear(nb_features, 1)
self.linear4_mu_classifier = nn.Linear(nb_features, 1)
self.linear4_logvar_classifier = nn.Linear(nb_features, 1)
self.conc_drop1 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop3 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_F = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_F2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_alpha = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu_classifier = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar_classifier = ConcreteDropout(weight_regularizer
=weight_regularizer, dropout_regularizer=dropout_regularizer)
self.tanh = nn.Tanh()
def forward(self, x):
regularization = torch.empty(12, device=x.device)
x1, regularization[0] = self.conc_drop1(x, nn.Sequential(self.
linear1, self.tanh))
x2, regularization[1] = self.conc_drop2(x1, nn.Sequential(self.
linear2, self.tanh))
x3, regularization[2] = self.conc_drop3(x2, nn.Sequential(self.
linear3, self.tanh))
mean, regularization[3] = self.conc_drop_mu(x3, self.linear4_mu)
logvar, regularization[4] = self.conc_drop_logvar(x3, self.
linear4_logvar)
F, regularization[5] = self.conc_drop_F(x3, self.linear4_F)
mean2, regularization[6] = self.conc_drop_mu2(x3, self.linear4_mu2)
logvar2, regularization[7] = self.conc_drop_logvar2(x3, self.
linear4_logvar2)
F2, regularization[8] = self.conc_drop_F2(x3, self.linear4_F2)
alpha, regularization[9] = self.conc_drop_alpha(x3, self.linear4_alpha)
mean_classifier, regularization[10] = self.conc_drop_mu_classifier(x3,
self.linear4_mu_classifier)
logvar_classifier, regularization[11
] = self.conc_drop_logvar_classifier(x3, self.
linear4_logvar_classifier)
return (mean, logvar, F, mean2, logvar2, F2, alpha, mean_classifier,
logvar_classifier, regularization.sum())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'X_dim': 4, 'Y_dim': 4, 'nb_features': 4,
'weight_regularizer': 4, 'dropout_regularizer': 0.5}]
|
import torch
from torch import device
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import numpy as np
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp12 = tl.load(in_ptr2 + x0, xmask)
tmp3 = tl.sigmoid(tmp2)
tmp4 = 1e-07
tmp5 = tmp3 + tmp4
tmp6 = tl_math.log(tmp5)
tmp7 = 1.0
tmp8 = tmp7 - tmp3
tmp9 = tmp8 + tmp4
tmp10 = tl_math.log(tmp9)
tmp11 = tmp6 - tmp10
tmp13 = tmp12 + tmp4
tmp14 = tl_math.log(tmp13)
tmp15 = tmp11 + tmp14
tmp16 = tmp7 - tmp12
tmp17 = tmp16 + tmp4
tmp18 = tl_math.log(tmp17)
tmp19 = tmp15 - tmp18
tmp20 = 10.0
tmp21 = tmp19 * tmp20
tmp22 = tl.sigmoid(tmp21)
tmp23 = tmp7 - tmp22
tmp24 = tmp0 * tmp23
tmp25 = tmp24 / tmp8
tl.store(out_ptr0 + x0, tmp25, xmask)
@triton.jit
def triton_per_fused_pow_sum_1(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2(in_ptr0,
in_ptr1, in_ptr2, out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel,
rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp5 = tl.load(in_ptr1 + 0)
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, 1])
tmp11 = tl.load(in_ptr2 + 0)
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp7 = tl.sigmoid(tmp6)
tmp8 = tl_math.log(tmp7)
tmp9 = 1.0
tmp10 = tmp9 - tmp7
tmp13 = 0.0
tmp14 = tmp12 + tmp13
tmp15 = tmp14 + tmp4
tmp16 = 4.0
tmp17 = tmp15 * tmp16
tmp18 = tmp17 / tmp10
tmp19 = tmp7 * tmp8
tmp20 = tl_math.log(tmp10)
tmp21 = tmp10 * tmp20
tmp22 = tmp19 + tmp21
tmp23 = 32.0
tmp24 = tmp22 * tmp23
tmp25 = tmp18 + tmp24
tmp26 = tmp18 / tmp10
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp8, None)
tl.store(out_ptr2 + tl.full([XBLOCK, 1], 0, tl.int32), tmp10, None)
tl.store(out_ptr3 + tl.full([XBLOCK, 1], 0, tl.int32), tmp25, None)
tl.store(out_ptr4 + tl.full([XBLOCK, 1], 0, tl.int32), tmp26, None)
@triton.jit
def triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3(in_ptr0,
in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp2 = tl.load(in_ptr1 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp13 = tl.load(in_ptr2 + x0, xmask)
tmp1 = libdevice.tanh(tmp0)
tmp4 = tl.sigmoid(tmp3)
tmp5 = 1e-07
tmp6 = tmp4 + tmp5
tmp7 = tl_math.log(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp4
tmp10 = tmp9 + tmp5
tmp11 = tl_math.log(tmp10)
tmp12 = tmp7 - tmp11
tmp14 = tmp13 + tmp5
tmp15 = tl_math.log(tmp14)
tmp16 = tmp12 + tmp15
tmp17 = tmp8 - tmp13
tmp18 = tmp17 + tmp5
tmp19 = tl_math.log(tmp18)
tmp20 = tmp16 - tmp19
tmp21 = 10.0
tmp22 = tmp20 * tmp21
tmp23 = tl.sigmoid(tmp22)
tmp24 = tmp8 - tmp23
tmp25 = tmp1 * tmp24
tmp26 = tmp25 / tmp9
tl.store(out_ptr0 + x0, tmp26, xmask)
@triton.jit
def triton_for_fused_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5,
in_ptr6, in_ptr7, in_ptr8, out_ptr0, out_ptr1, out_ptr2, out_ptr3,
out_ptr4, out_ptr5, out_ptr6, out_ptr7, out_ptr8):
pid = tl.program_id(0)
XBLOCK: tl.constexpr = 1024
num_xblocks_0 = tl.cdiv(1, XBLOCK)
num_xblocks_1 = num_xblocks_0 + tl.cdiv(1, XBLOCK)
num_xblocks_2 = num_xblocks_1 + tl.cdiv(1, XBLOCK)
num_xblocks_3 = num_xblocks_2 + tl.cdiv(1, XBLOCK)
num_xblocks_4 = num_xblocks_3 + tl.cdiv(1, XBLOCK)
num_xblocks_5 = num_xblocks_4 + tl.cdiv(1, XBLOCK)
num_xblocks_6 = num_xblocks_5 + tl.cdiv(1, XBLOCK)
num_xblocks_7 = num_xblocks_6 + tl.cdiv(1, XBLOCK)
num_xblocks_8 = num_xblocks_7 + tl.cdiv(1, XBLOCK)
if pid < num_xblocks_0:
pid_offset = pid
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp0 = tl.load(in_ptr0 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tl.store(out_ptr0 + tl.full([XBLOCK], 0, tl.int32), tmp1, None)
elif pid < num_xblocks_1:
pid_offset = pid - num_xblocks_0
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp2 = tl.load(in_ptr1 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tl.store(out_ptr1 + tl.full([XBLOCK], 0, tl.int32), tmp3, None)
elif pid < num_xblocks_2:
pid_offset = pid - num_xblocks_1
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp4 = tl.load(in_ptr2 + 0)
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tl.store(out_ptr2 + tl.full([XBLOCK], 0, tl.int32), tmp5, None)
elif pid < num_xblocks_3:
pid_offset = pid - num_xblocks_2
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp6 = tl.load(in_ptr3 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tl.store(out_ptr3 + tl.full([XBLOCK], 0, tl.int32), tmp7, None)
elif pid < num_xblocks_4:
pid_offset = pid - num_xblocks_3
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp8 = tl.load(in_ptr4 + 0)
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tl.store(out_ptr4 + tl.full([XBLOCK], 0, tl.int32), tmp9, None)
elif pid < num_xblocks_5:
pid_offset = pid - num_xblocks_4
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp10 = tl.load(in_ptr5 + 0)
tmp11 = tl.broadcast_to(tmp10, [XBLOCK])
tl.store(out_ptr5 + tl.full([XBLOCK], 0, tl.int32), tmp11, None)
elif pid < num_xblocks_6:
pid_offset = pid - num_xblocks_5
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp12 = tl.load(in_ptr6 + 0)
tmp13 = tl.broadcast_to(tmp12, [XBLOCK])
tl.store(out_ptr6 + tl.full([XBLOCK], 0, tl.int32), tmp13, None)
elif pid < num_xblocks_7:
pid_offset = pid - num_xblocks_6
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp14 = tl.load(in_ptr7 + 0)
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tl.store(out_ptr7 + tl.full([XBLOCK], 0, tl.int32), tmp15, None)
elif pid < num_xblocks_8:
pid_offset = pid - num_xblocks_7
xoffset = pid_offset * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
tmp16 = tl.load(in_ptr8 + 0)
tmp17 = tl.broadcast_to(tmp16, [XBLOCK])
tl.store(out_ptr8 + tl.full([XBLOCK], 0, tl.int32), tmp17, None)
else:
pass
@triton.jit
def triton_poi_fused_sigmoid_5(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 9
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(out_ptr0 + x0, tmp1, xmask)
@triton.jit
def triton_per_fused_add_div_log_mul_pow_rsub_sum_6(in_ptr0, in_ptr1,
in_ptr2, out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp5 = tl.load(in_ptr1 + 8)
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, 1])
tmp12 = tl.load(in_ptr2 + 0)
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp7 = tl_math.log(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp6
tmp10 = 0.0
tmp11 = tmp4 + tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp11 + tmp14
tmp16 = 4.0
tmp17 = tmp15 * tmp16
tmp18 = tmp17 / tmp9
tmp19 = tmp6 * tmp7
tmp20 = tl_math.log(tmp9)
tmp21 = tmp9 * tmp20
tmp22 = tmp19 + tmp21
tmp23 = 32.0
tmp24 = tmp22 * tmp23
tmp25 = tmp18 + tmp24
tmp26 = tmp18 / tmp9
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp7, None)
tl.store(out_ptr2 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, None)
tl.store(out_ptr3 + tl.full([XBLOCK, 1], 0, tl.int32), tmp25, None)
tl.store(out_ptr4 + tl.full([XBLOCK, 1], 0, tl.int32), tmp26, None)
@triton.jit
def triton_per_fused_pow_sum_7(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
rnumel = 3
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, rmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
@triton.jit
def triton_per_fused_pow_sum_8(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
rnumel = 24
RBLOCK: tl.constexpr = 32
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, rmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_9(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
rnumel = 6
RBLOCK: tl.constexpr = 8
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, 1])
tmp13 = tl.load(in_ptr2 + 2)
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp8 = 0.0
tmp9 = tmp7 + tmp8
tmp10 = tmp9 + tmp5
tmp11 = 4.0
tmp12 = tmp10 * tmp11
tmp15 = 1.0
tmp16 = tmp15 - tmp14
tmp17 = tmp12 / tmp16
tmp18 = tmp17 / tmp16
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp18, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_10(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
rnumel = 6
RBLOCK: tl.constexpr = 8
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, 1])
tmp13 = tl.load(in_ptr2 + 5)
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp8 = 0.0
tmp9 = tmp7 + tmp8
tmp10 = tmp9 + tmp5
tmp11 = 4.0
tmp12 = tmp10 * tmp11
tmp15 = 1.0
tmp16 = tmp15 - tmp14
tmp17 = tmp12 / tmp16
tmp18 = tmp17 / tmp16
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp18, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
@triton.jit
def triton_per_fused_pow_sum_11(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_add_div_mul_pow_rsub_sum_12(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10,
in_ptr11, in_ptr12, in_ptr13, in_ptr14, in_ptr15, in_ptr16, in_ptr17,
in_ptr18, in_ptr19, in_ptr20, out_ptr4, out_ptr5, out_ptr6, out_ptr7,
out_ptr8, xnumel, rnumel, XBLOCK: tl.constexpr):
rnumel = 12
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, rmask, other=0.0)
tmp6 = tl.load(in_ptr1 + r0, rmask, other=0.0)
tmp12 = tl.load(in_ptr2 + r0, rmask, other=0.0)
tmp18 = tl.load(in_ptr3 + r0, rmask, other=0.0)
tmp29 = tl.load(in_ptr4 + 0)
tmp30 = tl.broadcast_to(tmp29, [XBLOCK, RBLOCK])
tmp34 = tl.load(in_ptr5 + 0)
tmp35 = tl.broadcast_to(tmp34, [XBLOCK, RBLOCK])
tmp49 = tl.load(in_ptr6 + 0)
tmp50 = tl.broadcast_to(tmp49, [XBLOCK, RBLOCK])
tmp53 = tl.load(in_ptr7 + 0)
tmp54 = tl.broadcast_to(tmp53, [XBLOCK, RBLOCK])
tmp57 = tl.load(in_ptr8 + 0)
tmp58 = tl.broadcast_to(tmp57, [XBLOCK, RBLOCK])
tmp66 = tl.load(in_ptr9 + 0)
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = tl.load(in_ptr10 + 0)
tmp70 = tl.broadcast_to(tmp69, [XBLOCK, RBLOCK])
tmp73 = tl.load(in_ptr5 + 2)
tmp74 = tl.broadcast_to(tmp73, [XBLOCK, RBLOCK])
tmp87 = tl.load(in_ptr11 + 0)
tmp88 = tl.broadcast_to(tmp87, [XBLOCK, RBLOCK])
tmp91 = tl.load(in_ptr5 + 1)
tmp92 = tl.broadcast_to(tmp91, [XBLOCK, RBLOCK])
tmp107 = tl.load(in_ptr12 + 0)
tmp108 = tl.broadcast_to(tmp107, [XBLOCK, RBLOCK])
tmp111 = tl.load(in_ptr5 + 4)
tmp112 = tl.broadcast_to(tmp111, [XBLOCK, RBLOCK])
tmp125 = tl.load(in_ptr13 + 0)
tmp126 = tl.broadcast_to(tmp125, [XBLOCK, RBLOCK])
tmp129 = tl.load(in_ptr5 + 3)
tmp130 = tl.broadcast_to(tmp129, [XBLOCK, RBLOCK])
tmp144 = tl.load(in_ptr14 + 0)
tmp145 = tl.broadcast_to(tmp144, [XBLOCK, RBLOCK])
tmp147 = tl.load(in_ptr15 + 0)
tmp148 = tl.broadcast_to(tmp147, [XBLOCK, RBLOCK])
tmp152 = tl.load(in_ptr5 + 6)
tmp153 = tl.broadcast_to(tmp152, [XBLOCK, RBLOCK])
tmp165 = tl.load(in_ptr16 + 0)
tmp166 = tl.broadcast_to(tmp165, [XBLOCK, RBLOCK])
tmp168 = tl.load(in_ptr17 + 0)
tmp169 = tl.broadcast_to(tmp168, [XBLOCK, RBLOCK])
tmp172 = tl.load(in_ptr5 + 5)
tmp173 = tl.broadcast_to(tmp172, [XBLOCK, RBLOCK])
tmp187 = tl.load(in_ptr18 + 0)
tmp188 = tl.broadcast_to(tmp187, [XBLOCK, RBLOCK])
tmp191 = tl.load(in_ptr19 + 0)
tmp192 = tl.broadcast_to(tmp191, [XBLOCK, RBLOCK])
tmp194 = tl.load(in_ptr20 + 0)
tmp195 = tl.broadcast_to(tmp194, [XBLOCK, RBLOCK])
tmp199 = tl.load(in_ptr5 + 7)
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp216 = tl.broadcast_to(tmp107, [XBLOCK, 1])
tmp219 = tl.broadcast_to(tmp111, [XBLOCK, 1])
tmp223 = tl.broadcast_to(tmp125, [XBLOCK, 1])
tmp226 = tl.broadcast_to(tmp129, [XBLOCK, 1])
tmp230 = tl.broadcast_to(tmp87, [XBLOCK, 1])
tmp233 = tl.broadcast_to(tmp91, [XBLOCK, 1])
tmp237 = tl.broadcast_to(tmp29, [XBLOCK, 1])
tmp240 = tl.broadcast_to(tmp34, [XBLOCK, 1])
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(rmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp7 = tmp6 * tmp6
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp10 = tl.where(rmask, tmp8, 0)
tmp11 = tl.sum(tmp10, 1)[:, None]
tmp13 = tmp12 * tmp12
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, RBLOCK])
tmp16 = tl.where(rmask, tmp14, 0)
tmp17 = tl.sum(tmp16, 1)[:, None]
tmp19 = tmp18 * tmp18
tmp20 = tl.broadcast_to(tmp19, [XBLOCK, RBLOCK])
tmp22 = tl.where(rmask, tmp20, 0)
tmp23 = tl.sum(tmp22, 1)[:, None]
tmp24 = r0
tmp25 = tl.full([1, 1], 3, tl.int32)
tmp26 = tmp24 == tmp25
tmp27 = 0.0
tmp28 = tmp23 + tmp27
tmp31 = tmp28 + tmp30
tmp32 = 4.0
tmp33 = tmp31 * tmp32
tmp36 = 1.0
tmp37 = tmp36 - tmp35
tmp38 = tmp33 / tmp37
tmp39 = tl_math.log(tmp35)
tmp40 = tmp35 * tmp39
tmp41 = tl_math.log(tmp37)
tmp42 = tmp37 * tmp41
tmp43 = tmp40 + tmp42
tmp44 = 32.0
tmp45 = tmp43 * tmp44
tmp46 = tmp38 + tmp45
tmp47 = tl.full([1, 1], 2, tl.int32)
tmp48 = tmp24 == tmp47
tmp51 = tl.full([1, 1], 1, tl.int32)
tmp52 = tmp24 == tmp51
tmp55 = tl.full([1, 1], 0, tl.int32)
tmp56 = tmp24 == tmp55
tmp59 = float('nan')
tmp60 = tl.where(tmp56, tmp58, tmp59)
tmp61 = tl.where(tmp52, tmp54, tmp60)
tmp62 = tl.where(tmp48, tmp50, tmp61)
tmp63 = tl.where(tmp26, tmp46, tmp62)
tmp64 = tl.full([1, 1], 5, tl.int32)
tmp65 = tmp24 == tmp64
tmp68 = tmp67 + tmp27
tmp71 = tmp68 + tmp70
tmp72 = tmp71 * tmp32
tmp75 = tmp36 - tmp74
tmp76 = tmp72 / tmp75
tmp77 = tl_math.log(tmp74)
tmp78 = tmp74 * tmp77
tmp79 = tl_math.log(tmp75)
tmp80 = tmp75 * tmp79
tmp81 = tmp78 + tmp80
tmp82 = tmp81 * tmp44
tmp83 = tmp76 + tmp82
tmp84 = tl.full([1, 1], 4, tl.int32)
tmp85 = tmp24 == tmp84
tmp86 = tmp11 + tmp27
tmp89 = tmp86 + tmp88
tmp90 = tmp89 * tmp32
tmp93 = tmp36 - tmp92
tmp94 = tmp90 / tmp93
tmp95 = tl_math.log(tmp92)
tmp96 = tmp92 * tmp95
tmp97 = tl_math.log(tmp93)
tmp98 = tmp93 * tmp97
tmp99 = tmp96 + tmp98
tmp100 = tmp99 * tmp44
tmp101 = tmp94 + tmp100
tmp102 = tl.where(tmp85, tmp101, tmp63)
tmp103 = tl.where(tmp65, tmp83, tmp102)
tmp104 = tl.full([1, 1], 7, tl.int32)
tmp105 = tmp24 == tmp104
tmp106 = tmp17 + tmp27
tmp109 = tmp106 + tmp108
tmp110 = tmp109 * tmp32
tmp113 = tmp36 - tmp112
tmp114 = tmp110 / tmp113
tmp115 = tl_math.log(tmp112)
tmp116 = tmp112 * tmp115
tmp117 = tl_math.log(tmp113)
tmp118 = tmp113 * tmp117
tmp119 = tmp116 + tmp118
tmp120 = tmp119 * tmp44
tmp121 = tmp114 + tmp120
tmp122 = tl.full([1, 1], 6, tl.int32)
tmp123 = tmp24 == tmp122
tmp124 = tmp5 + tmp27
tmp127 = tmp124 + tmp126
tmp128 = tmp127 * tmp32
tmp131 = tmp36 - tmp130
tmp132 = tmp128 / tmp131
tmp133 = tl_math.log(tmp130)
tmp134 = tmp130 * tmp133
tmp135 = tl_math.log(tmp131)
tmp136 = tmp131 * tmp135
tmp137 = tmp134 + tmp136
tmp138 = tmp137 * tmp44
tmp139 = tmp132 + tmp138
tmp140 = tl.where(tmp123, tmp139, tmp103)
tmp141 = tl.where(tmp105, tmp121, tmp140)
tmp142 = tl.full([1, 1], 9, tl.int32)
tmp143 = tmp24 == tmp142
tmp146 = tmp145 + tmp27
tmp149 = tmp148 * tmp148
tmp150 = tmp146 + tmp149
tmp151 = tmp150 * tmp32
tmp154 = tmp36 - tmp153
tmp155 = tmp151 / tmp154
tmp156 = tl_math.log(tmp153)
tmp157 = tmp153 * tmp156
tmp158 = tl_math.log(tmp154)
tmp159 = tmp154 * tmp158
tmp160 = tmp157 + tmp159
tmp161 = tmp160 * tmp44
tmp162 = tmp155 + tmp161
tmp163 = tl.full([1, 1], 8, tl.int32)
tmp164 = tmp24 == tmp163
tmp167 = tmp166 + tmp27
tmp170 = tmp167 + tmp169
tmp171 = tmp170 * tmp32
tmp174 = tmp36 - tmp173
tmp175 = tmp171 / tmp174
tmp176 = tl_math.log(tmp173)
tmp177 = tmp173 * tmp176
tmp178 = tl_math.log(tmp174)
tmp179 = tmp174 * tmp178
tmp180 = tmp177 + tmp179
tmp181 = tmp180 * tmp44
tmp182 = tmp175 + tmp181
tmp183 = tl.where(tmp164, tmp182, tmp141)
tmp184 = tl.where(tmp143, tmp162, tmp183)
tmp185 = tl.full([1, 1], 11, tl.int32)
tmp186 = tmp24 == tmp185
tmp189 = tl.full([1, 1], 10, tl.int32)
tmp190 = tmp24 == tmp189
tmp193 = tmp192 + tmp27
tmp196 = tmp195 * tmp195
tmp197 = tmp193 + tmp196
tmp198 = tmp197 * tmp32
tmp201 = tmp36 - tmp200
tmp202 = tmp198 / tmp201
tmp203 = tl_math.log(tmp200)
tmp204 = tmp200 * tmp203
tmp205 = tl_math.log(tmp201)
tmp206 = tmp201 * tmp205
tmp207 = tmp204 + tmp206
tmp208 = tmp207 * tmp44
tmp209 = tmp202 + tmp208
tmp210 = tl.where(tmp190, tmp209, tmp184)
tmp211 = tl.where(tmp186, tmp188, tmp210)
tmp212 = tl.broadcast_to(tmp211, [XBLOCK, RBLOCK])
tmp214 = tl.where(rmask, tmp212, 0)
tmp215 = tl.sum(tmp214, 1)[:, None]
tmp217 = tmp106 + tmp216
tmp218 = tmp217 * tmp32
tmp220 = tmp36 - tmp219
tmp221 = tmp218 / tmp220
tmp222 = tmp221 / tmp220
tmp224 = tmp124 + tmp223
tmp225 = tmp224 * tmp32
tmp227 = tmp36 - tmp226
tmp228 = tmp225 / tmp227
tmp229 = tmp228 / tmp227
tmp231 = tmp86 + tmp230
tmp232 = tmp231 * tmp32
tmp234 = tmp36 - tmp233
tmp235 = tmp232 / tmp234
tmp236 = tmp235 / tmp234
tmp238 = tmp28 + tmp237
tmp239 = tmp238 * tmp32
tmp241 = tmp36 - tmp240
tmp242 = tmp239 / tmp241
tmp243 = tmp242 / tmp241
tl.store(out_ptr5 + tl.full([XBLOCK, 1], 0, tl.int32), tmp222, None)
tl.store(out_ptr6 + tl.full([XBLOCK, 1], 0, tl.int32), tmp229, None)
tl.store(out_ptr7 + tl.full([XBLOCK, 1], 0, tl.int32), tmp236, None)
tl.store(out_ptr8 + tl.full([XBLOCK, 1], 0, tl.int32), tmp243, None)
tl.store(out_ptr4 + tl.full([XBLOCK, 1], 0, tl.int32), tmp215, None)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_13(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_14(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 1)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_15(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 2)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_16(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 3)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_17(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_18(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 5)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_19(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 6)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_20(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 7)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_add_div_log_rsub_sub_21(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 8)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK])
tmp10 = tl.load(in_ptr1 + x0, xmask)
tmp2 = 1e-07
tmp3 = tmp1 + tmp2
tmp4 = tl_math.log(tmp3)
tmp5 = 1.0
tmp6 = tmp5 - tmp1
tmp7 = tmp6 + tmp2
tmp8 = tl_math.log(tmp7)
tmp9 = tmp4 - tmp8
tmp11 = tmp10 + tmp2
tmp12 = tl_math.log(tmp11)
tmp13 = tmp9 + tmp12
tmp14 = tmp5 - tmp10
tmp15 = tmp14 + tmp2
tmp16 = tl_math.log(tmp15)
tmp17 = tmp13 - tmp16
tmp18 = 10.0
tmp19 = tmp17 * tmp18
tl.store(out_ptr0 + x0, tmp19, xmask)
@triton.jit
def triton_poi_fused_sigmoid_22(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(out_ptr0 + x0, tmp1, xmask)
@triton.jit
def triton_poi_fused_div_mul_rsub_tanh_23(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6,
out_ptr7, out_ptr8, out_ptr9, out_ptr10, out_ptr11, out_ptr12,
out_ptr13, out_ptr14, out_ptr15, out_ptr16, out_ptr17, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (1792 + x0), xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask)
tmp6 = tl.load(in_ptr2 + 7)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp10 = tl.load(in_ptr0 + (1536 + x0), xmask)
tmp13 = tl.load(in_ptr2 + 6)
tmp14 = tl.broadcast_to(tmp13, [XBLOCK])
tmp17 = tl.load(in_ptr0 + (1280 + x0), xmask)
tmp20 = tl.load(in_ptr2 + 5)
tmp21 = tl.broadcast_to(tmp20, [XBLOCK])
tmp24 = tl.load(in_ptr0 + (1024 + x0), xmask)
tmp27 = tl.load(in_ptr2 + 4)
tmp28 = tl.broadcast_to(tmp27, [XBLOCK])
tmp31 = tl.load(in_ptr0 + (768 + x0), xmask)
tmp34 = tl.load(in_ptr2 + 3)
tmp35 = tl.broadcast_to(tmp34, [XBLOCK])
tmp38 = tl.load(in_ptr0 + (512 + x0), xmask)
tmp41 = tl.load(in_ptr2 + 2)
tmp42 = tl.broadcast_to(tmp41, [XBLOCK])
tmp45 = tl.load(in_ptr0 + (256 + x0), xmask)
tmp48 = tl.load(in_ptr2 + 1)
tmp49 = tl.broadcast_to(tmp48, [XBLOCK])
tmp52 = tl.load(in_ptr0 + x0, xmask)
tmp55 = tl.load(in_ptr2 + 0)
tmp56 = tl.broadcast_to(tmp55, [XBLOCK])
tmp59 = tl.load(in_ptr0 + (2048 + x0), xmask)
tmp62 = tl.load(in_ptr2 + 8)
tmp63 = tl.broadcast_to(tmp62, [XBLOCK])
tmp1 = 1.0
tmp2 = tmp1 - tmp0
tmp4 = libdevice.tanh(tmp3)
tmp5 = tmp4 * tmp2
tmp8 = tmp1 - tmp7
tmp9 = tmp5 / tmp8
tmp11 = tmp1 - tmp10
tmp12 = tmp4 * tmp11
tmp15 = tmp1 - tmp14
tmp16 = tmp12 / tmp15
tmp18 = tmp1 - tmp17
tmp19 = tmp4 * tmp18
tmp22 = tmp1 - tmp21
tmp23 = tmp19 / tmp22
tmp25 = tmp1 - tmp24
tmp26 = tmp4 * tmp25
tmp29 = tmp1 - tmp28
tmp30 = tmp26 / tmp29
tmp32 = tmp1 - tmp31
tmp33 = tmp4 * tmp32
tmp36 = tmp1 - tmp35
tmp37 = tmp33 / tmp36
tmp39 = tmp1 - tmp38
tmp40 = tmp4 * tmp39
tmp43 = tmp1 - tmp42
tmp44 = tmp40 / tmp43
tmp46 = tmp1 - tmp45
tmp47 = tmp4 * tmp46
tmp50 = tmp1 - tmp49
tmp51 = tmp47 / tmp50
tmp53 = tmp1 - tmp52
tmp54 = tmp4 * tmp53
tmp57 = tmp1 - tmp56
tmp58 = tmp54 / tmp57
tmp60 = tmp1 - tmp59
tmp61 = tmp4 * tmp60
tmp64 = tmp1 - tmp63
tmp65 = tmp61 / tmp64
tl.store(out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr1 + x0, tmp9, xmask)
tl.store(out_ptr2 + x0, tmp11, xmask)
tl.store(out_ptr3 + x0, tmp16, xmask)
tl.store(out_ptr4 + x0, tmp18, xmask)
tl.store(out_ptr5 + x0, tmp23, xmask)
tl.store(out_ptr6 + x0, tmp25, xmask)
tl.store(out_ptr7 + x0, tmp30, xmask)
tl.store(out_ptr8 + x0, tmp32, xmask)
tl.store(out_ptr9 + x0, tmp37, xmask)
tl.store(out_ptr10 + x0, tmp39, xmask)
tl.store(out_ptr11 + x0, tmp44, xmask)
tl.store(out_ptr12 + x0, tmp46, xmask)
tl.store(out_ptr13 + x0, tmp51, xmask)
tl.store(out_ptr14 + x0, tmp53, xmask)
tl.store(out_ptr15 + x0, tmp58, xmask)
tl.store(out_ptr16 + x0, tmp60, xmask)
tl.store(out_ptr17 + x0, tmp65, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22,
primals_23, primals_24, primals_25, primals_26, primals_27,
primals_28, primals_29, primals_30, primals_31, primals_32,
primals_33, primals_34, primals_35, primals_36, primals_37) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1,), (1,))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (1,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (1,), (1,))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (1,), (1,))
assert_size_stride(primals_12, (3, 4), (4, 1))
assert_size_stride(primals_13, (3,), (1,))
assert_size_stride(primals_14, (1,), (1,))
assert_size_stride(primals_15, (3, 4), (4, 1))
assert_size_stride(primals_16, (3,), (1,))
assert_size_stride(primals_17, (1,), (1,))
assert_size_stride(primals_18, (6, 4), (4, 1))
assert_size_stride(primals_19, (6,), (1,))
assert_size_stride(primals_20, (1,), (1,))
assert_size_stride(primals_21, (3, 4), (4, 1))
assert_size_stride(primals_22, (3,), (1,))
assert_size_stride(primals_23, (1,), (1,))
assert_size_stride(primals_24, (3, 4), (4, 1))
assert_size_stride(primals_25, (3,), (1,))
assert_size_stride(primals_26, (1,), (1,))
assert_size_stride(primals_27, (6, 4), (4, 1))
assert_size_stride(primals_28, (6,), (1,))
assert_size_stride(primals_29, (1,), (1,))
assert_size_stride(primals_30, (1, 4), (4, 1))
assert_size_stride(primals_31, (1,), (1,))
assert_size_stride(primals_32, (1,), (1,))
assert_size_stride(primals_33, (1, 4), (4, 1))
assert_size_stride(primals_34, (1,), (1,))
assert_size_stride(primals_35, (1,), (1,))
assert_size_stride(primals_36, (1, 4), (4, 1))
assert_size_stride(primals_37, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf1 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf2 = buf1
del buf1
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_0[grid(256)](
primals_1, primals_2, buf2, buf3, 256, XBLOCK=128, num_warps=4,
num_stages=1)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_4, reinterpret_tensor(buf3, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf4)
buf5 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_1[grid(1)](primals_3, buf5, 1, 16, XBLOCK=
1, num_warps=2, num_stages=1)
buf7 = empty_strided_cuda((1,), (1,), torch.float32)
buf8 = empty_strided_cuda((1,), (1,), torch.float32)
buf9 = empty_strided_cuda((1,), (1,), torch.float32)
buf131 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2[grid(1)](
primals_4, primals_2, buf5, buf7, buf8, buf9, buf131, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf10 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf11 = buf10
del buf10
buf12 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3[grid(256)](
buf4, primals_5, buf11, buf12, 256, XBLOCK=128, num_warps=4,
num_stages=1)
buf13 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf12, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf13)
buf14 = buf5
del buf5
triton_per_fused_pow_sum_1[grid(1)](primals_6, buf14, 1, 16, XBLOCK
=1, num_warps=2, num_stages=1)
buf16 = empty_strided_cuda((1,), (1,), torch.float32)
buf17 = empty_strided_cuda((1,), (1,), torch.float32)
buf18 = empty_strided_cuda((1,), (1,), torch.float32)
buf130 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2[grid(1)](
primals_7, primals_5, buf14, buf16, buf17, buf18, buf130, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf19 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf20 = buf19
del buf19
buf21 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_div_log_mul_rsub_sigmoid_sub_tanh_3[grid(256)](
buf13, primals_8, buf20, buf21, 256, XBLOCK=128, num_warps=4,
num_stages=1)
buf22 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_10, reinterpret_tensor(buf21, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf22)
buf23 = buf14
del buf14
triton_per_fused_pow_sum_1[grid(1)](primals_9, buf23, 1, 16, XBLOCK
=1, num_warps=2, num_stages=1)
buf25 = empty_strided_cuda((1,), (1,), torch.float32)
buf26 = empty_strided_cuda((1,), (1,), torch.float32)
buf27 = empty_strided_cuda((1,), (1,), torch.float32)
buf129 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_log_mul_pow_rsub_sigmoid_sum_2[grid(1)](
primals_10, primals_8, buf23, buf25, buf26, buf27, buf129, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf37 = empty_strided_cuda((9,), (1,), torch.float32)
buf28 = reinterpret_tensor(buf37, (1,), (1,), 0)
buf29 = reinterpret_tensor(buf37, (1,), (1,), 1)
buf30 = reinterpret_tensor(buf37, (1,), (1,), 2)
buf31 = reinterpret_tensor(buf37, (1,), (1,), 3)
buf32 = reinterpret_tensor(buf37, (1,), (1,), 4)
buf33 = reinterpret_tensor(buf37, (1,), (1,), 5)
buf34 = reinterpret_tensor(buf37, (1,), (1,), 6)
buf35 = reinterpret_tensor(buf37, (1,), (1,), 7)
buf36 = reinterpret_tensor(buf37, (1,), (1,), 8)
triton_for_fused_4[9, 1, 1](primals_11, primals_14, primals_17,
primals_20, primals_23, primals_26, primals_29, primals_32,
primals_35, buf28, buf29, buf30, buf31, buf32, buf33, buf34,
buf35, buf36, num_warps=8, num_stages=1)
del primals_11
del primals_14
del primals_17
del primals_20
del primals_23
del primals_26
del primals_29
del primals_32
del primals_35
buf38 = empty_strided_cuda((9, 1), (1, 1), torch.float32)
triton_poi_fused_sigmoid_5[grid(9)](buf37, buf38, 9, XBLOCK=16,
num_warps=1, num_stages=1)
del buf28
del buf29
del buf30
del buf31
del buf32
del buf33
del buf34
del buf35
del buf36
del buf37
buf39 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf40 = buf39
del buf39
buf117 = reinterpret_tensor(buf23, (1,), (1,), 0)
del buf23
buf118 = empty_strided_cuda((1,), (1,), torch.float32)
buf119 = empty_strided_cuda((1,), (1,), torch.float32)
buf122 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_log_mul_pow_rsub_sum_6[grid(1)](primals_36,
buf38, primals_37, buf117, buf118, buf119, buf122, 1, 4, XBLOCK
=1, num_warps=2, num_stages=1)
buf42 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_7[grid(1)](primals_13, buf42, 1, 3, XBLOCK
=1, num_warps=2, num_stages=1)
buf47 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_7[grid(1)](primals_16, buf47, 1, 3, XBLOCK
=1, num_warps=2, num_stages=1)
buf50 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_8[grid(1)](primals_18, buf50, 1, 24,
XBLOCK=1, num_warps=2, num_stages=1)
buf51 = empty_strided_cuda((), (), torch.float32)
buf126 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_mul_pow_rsub_sum_9[grid(1)](primals_19,
buf50, buf38, buf51, buf126, 1, 6, XBLOCK=1, num_warps=2,
num_stages=1)
buf56 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_7[grid(1)](primals_22, buf56, 1, 3, XBLOCK
=1, num_warps=2, num_stages=1)
buf60 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_7[grid(1)](primals_25, buf60, 1, 3, XBLOCK
=1, num_warps=2, num_stages=1)
buf64 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_8[grid(1)](primals_27, buf64, 1, 24,
XBLOCK=1, num_warps=2, num_stages=1)
buf65 = empty_strided_cuda((), (), torch.float32)
buf123 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_mul_pow_rsub_sum_10[grid(1)](primals_28,
buf64, buf38, buf65, buf123, 1, 6, XBLOCK=1, num_warps=2,
num_stages=1)
buf68 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_11[grid(1)](primals_30, buf68, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf72 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_pow_sum_11[grid(1)](primals_33, buf72, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf121 = empty_strided_cuda((), (), torch.float32)
buf124 = empty_strided_cuda((1,), (1,), torch.float32)
buf125 = empty_strided_cuda((1,), (1,), torch.float32)
buf127 = empty_strided_cuda((1,), (1,), torch.float32)
buf128 = empty_strided_cuda((1,), (1,), torch.float32)
triton_per_fused_add_div_mul_pow_rsub_sum_12[grid(1)](primals_21,
primals_15, primals_24, primals_12, buf42, buf38, buf27, buf18,
buf9, buf50, buf51, buf47, buf60, buf56, buf68, primals_31,
buf64, buf65, buf119, buf72, primals_34, buf121, buf124, buf125,
buf127, buf128, 1, 12, XBLOCK=1, num_warps=2, num_stages=1)
del buf119
del buf18
del buf27
del buf42
del buf47
del buf50
del buf51
del buf56
del buf60
del buf64
del buf65
del buf68
del buf72
del buf9
buf44 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf45 = buf44
del buf44
buf48 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf49 = buf48
del buf48
buf53 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf54 = buf53
del buf53
buf57 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf58 = buf57
del buf57
buf62 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf63 = buf62
del buf62
buf66 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf67 = buf66
del buf66
buf70 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf71 = buf70
del buf70
buf73 = torch.ops.aten.rand.default([4, 4, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf74 = buf73
del buf73
buf84 = empty_strided_cuda((36, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf75 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 0)
triton_poi_fused_add_div_log_rsub_sub_13[grid(256)](buf38, buf40,
buf75, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf76 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 256)
triton_poi_fused_add_div_log_rsub_sub_14[grid(256)](buf38, buf45,
buf76, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf77 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 512)
triton_poi_fused_add_div_log_rsub_sub_15[grid(256)](buf38, buf49,
buf77, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf78 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 768)
triton_poi_fused_add_div_log_rsub_sub_16[grid(256)](buf38, buf54,
buf78, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf79 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1024)
triton_poi_fused_add_div_log_rsub_sub_17[grid(256)](buf38, buf58,
buf79, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf80 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1280)
triton_poi_fused_add_div_log_rsub_sub_18[grid(256)](buf38, buf63,
buf80, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf81 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1536)
triton_poi_fused_add_div_log_rsub_sub_19[grid(256)](buf38, buf67,
buf81, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf82 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 1792)
triton_poi_fused_add_div_log_rsub_sub_20[grid(256)](buf38, buf71,
buf82, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf83 = reinterpret_tensor(buf84, (4, 4, 4, 4), (64, 16, 4, 1), 2048)
triton_poi_fused_add_div_log_rsub_sub_21[grid(256)](buf38, buf74,
buf83, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf85 = empty_strided_cuda((9, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_poi_fused_sigmoid_22[grid(2304)](buf84, buf85, 2304, XBLOCK=
128, num_warps=4, num_stages=1)
del buf75
del buf76
del buf77
del buf78
del buf79
del buf80
del buf81
del buf82
del buf83
del buf84
buf86 = buf74
del buf74
buf87 = buf71
del buf71
buf90 = buf67
del buf67
buf91 = buf63
del buf63
buf94 = buf58
del buf58
buf95 = buf54
del buf54
buf97 = buf49
del buf49
buf98 = buf45
del buf45
buf100 = buf40
del buf40
buf101 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf103 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf104 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf106 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf107 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf109 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf110 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf112 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
buf113 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32
)
triton_poi_fused_div_mul_rsub_tanh_23[grid(256)](buf85, buf22,
buf38, buf86, buf87, buf90, buf91, buf94, buf95, buf97, buf98,
buf100, buf101, buf103, buf104, buf106, buf107, buf109, buf110,
buf112, buf113, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf89 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_34, reinterpret_tensor(buf87, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_33, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf89)
buf93 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_31, reinterpret_tensor(buf91, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_30, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf93)
buf96 = empty_strided_cuda((64, 6), (6, 1), torch.float32)
extern_kernels.addmm(primals_28, reinterpret_tensor(buf95, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_27, (4, 6), (1, 4), 0),
alpha=1, beta=1, out=buf96)
buf99 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
extern_kernels.addmm(primals_25, reinterpret_tensor(buf98, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_24, (4, 3), (1, 4), 0),
alpha=1, beta=1, out=buf99)
buf102 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
extern_kernels.addmm(primals_22, reinterpret_tensor(buf101, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_21, (4, 3), (1, 4), 0),
alpha=1, beta=1, out=buf102)
buf105 = empty_strided_cuda((64, 6), (6, 1), torch.float32)
extern_kernels.addmm(primals_19, reinterpret_tensor(buf104, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_18, (4, 6), (1, 4), 0),
alpha=1, beta=1, out=buf105)
buf108 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
extern_kernels.addmm(primals_16, reinterpret_tensor(buf107, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_15, (4, 3), (1, 4), 0),
alpha=1, beta=1, out=buf108)
buf111 = empty_strided_cuda((64, 3), (3, 1), torch.float32)
extern_kernels.addmm(primals_13, reinterpret_tensor(buf110, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_12, (4, 3), (1, 4), 0),
alpha=1, beta=1, out=buf111)
buf115 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_37, reinterpret_tensor(buf113, (64, 4),
(4, 1), 0), reinterpret_tensor(primals_36, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf115)
return (reinterpret_tensor(buf111, (4, 4, 4, 3), (48, 12, 3, 1), 0),
reinterpret_tensor(buf108, (4, 4, 4, 3), (48, 12, 3, 1), 0),
reinterpret_tensor(buf105, (4, 4, 4, 6), (96, 24, 6, 1), 0),
reinterpret_tensor(buf102, (4, 4, 4, 3), (48, 12, 3, 1), 0),
reinterpret_tensor(buf99, (4, 4, 4, 3), (48, 12, 3, 1), 0),
reinterpret_tensor(buf96, (4, 4, 4, 6), (96, 24, 6, 1), 0),
reinterpret_tensor(buf93, (4, 4, 4, 1), (16, 4, 1, 1), 0),
reinterpret_tensor(buf89, (4, 4, 4, 1), (16, 4, 1, 1), 0),
reinterpret_tensor(buf115, (4, 4, 4, 1), (16, 4, 1, 1), 0), buf121,
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_12, primals_13,
primals_15, primals_16, primals_18, primals_19, primals_21,
primals_22, primals_24, primals_25, primals_27, primals_28,
primals_30, primals_31, primals_33, primals_34, primals_36,
primals_37, buf2, buf3, buf4, buf7, buf8, buf11, buf12, buf13,
buf16, buf17, buf20, buf21, buf22, buf25, buf26, buf38,
reinterpret_tensor(buf38, (1,), (1,), 0), reinterpret_tensor(buf38,
(1,), (1,), 1), reinterpret_tensor(buf38, (1,), (1,), 2),
reinterpret_tensor(buf38, (1,), (1,), 3), reinterpret_tensor(buf38,
(1,), (1,), 4), reinterpret_tensor(buf38, (1,), (1,), 5),
reinterpret_tensor(buf38, (1,), (1,), 6), reinterpret_tensor(buf38,
(1,), (1,), 7), reinterpret_tensor(buf38, (1,), (1,), 8), buf85,
buf86, buf87, buf90, buf91, buf94, buf95, buf97, buf98, buf100,
buf101, buf103, buf104, buf106, buf107, buf109, buf110, buf112,
buf113, buf117, buf118, buf122, buf123, buf124, buf125, buf126,
buf127, buf128, buf129, buf130, buf131)
class ConcreteDropout(nn.Module):
def __init__(self, weight_regularizer=1e-06, dropout_regularizer=1e-05,
init_min=0.1, init_max=0.1):
super(ConcreteDropout, self).__init__()
self.weight_regularizer = weight_regularizer
self.dropout_regularizer = dropout_regularizer
init_min = np.log(init_min) - np.log(1.0 - init_min)
init_max = np.log(init_max) - np.log(1.0 - init_max)
self.p_logit = nn.Parameter(torch.empty(1).uniform_(init_min, init_max)
)
def forward(self, x, layer):
p = torch.sigmoid(self.p_logit)
out = layer(self._concrete_dropout(x, p))
sum_of_square = 0
for param in layer.parameters():
sum_of_square += torch.sum(torch.pow(param, 2))
weights_regularizer = self.weight_regularizer * sum_of_square / (1 - p)
dropout_regularizer = p * torch.log(p)
dropout_regularizer += (1.0 - p) * torch.log(1.0 - p)
input_dimensionality = x[0].numel()
dropout_regularizer *= self.dropout_regularizer * input_dimensionality
regularization = weights_regularizer + dropout_regularizer
return out, regularization
def _concrete_dropout(self, x, p):
eps = 1e-07
temp = 0.1
unif_noise = torch.rand_like(x)
drop_prob = torch.log(p + eps) - torch.log(1 - p + eps) + torch.log(
unif_noise + eps) - torch.log(1 - unif_noise + eps)
drop_prob = torch.sigmoid(drop_prob / temp)
random_tensor = 1 - drop_prob
retain_prob = 1 - p
x = torch.mul(x, random_tensor)
x /= retain_prob
return x
class ConcreteDenseMixtureNew(nn.Module):
def __init__(self, X_dim, Y_dim, nb_features, weight_regularizer,
dropout_regularizer, verbose=True):
super(ConcreteDenseMixtureNew, self).__init__()
self.verbose = verbose
self.rank = 2
self.linear1 = nn.Linear(X_dim, nb_features)
self.linear2 = nn.Linear(nb_features, nb_features)
self.linear3 = nn.Linear(nb_features, nb_features)
self.linear4_mu = nn.Linear(nb_features, Y_dim - 1)
self.linear4_logvar = nn.Linear(nb_features, Y_dim - 1)
self.linear4_F = nn.Linear(nb_features, (Y_dim - 1) * self.rank)
self.linear4_mu2 = nn.Linear(nb_features, Y_dim - 1)
self.linear4_logvar2 = nn.Linear(nb_features, Y_dim - 1)
self.linear4_F2 = nn.Linear(nb_features, (Y_dim - 1) * self.rank)
self.linear4_alpha = nn.Linear(nb_features, 1)
self.linear4_mu_classifier = nn.Linear(nb_features, 1)
self.linear4_logvar_classifier = nn.Linear(nb_features, 1)
self.conc_drop1 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop3 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_F = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_F2 = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_alpha = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_mu_classifier = ConcreteDropout(weight_regularizer=
weight_regularizer, dropout_regularizer=dropout_regularizer)
self.conc_drop_logvar_classifier = ConcreteDropout(weight_regularizer
=weight_regularizer, dropout_regularizer=dropout_regularizer)
self.tanh = nn.Tanh()
def forward(self, input_0):
primals_3 = self.linear1.weight
primals_4 = self.linear1.bias
primals_6 = self.linear2.weight
primals_7 = self.linear2.bias
primals_9 = self.linear3.weight
primals_10 = self.linear3.bias
primals_12 = self.linear4_mu.weight
primals_13 = self.linear4_mu.bias
primals_15 = self.linear4_logvar.weight
primals_16 = self.linear4_logvar.bias
primals_18 = self.linear4_F.weight
primals_19 = self.linear4_F.bias
primals_21 = self.linear4_mu2.weight
primals_22 = self.linear4_mu2.bias
primals_24 = self.linear4_logvar2.weight
primals_25 = self.linear4_logvar2.bias
primals_27 = self.linear4_F2.weight
primals_28 = self.linear4_F2.bias
primals_30 = self.linear4_alpha.weight
primals_2 = self.linear4_alpha.bias
primals_33 = self.linear4_mu_classifier.weight
primals_5 = self.linear4_mu_classifier.bias
primals_36 = self.linear4_logvar_classifier.weight
primals_8 = self.linear4_logvar_classifier.bias
primals_11 = self.conc_drop1.p_logit
primals_14 = self.conc_drop2.p_logit
primals_17 = self.conc_drop3.p_logit
primals_20 = self.conc_drop_mu.p_logit
primals_23 = self.conc_drop_logvar.p_logit
primals_26 = self.conc_drop_F.p_logit
primals_29 = self.conc_drop_mu2.p_logit
primals_31 = self.conc_drop_logvar2.p_logit
primals_32 = self.conc_drop_F2.p_logit
primals_34 = self.conc_drop_alpha.p_logit
primals_35 = self.conc_drop_mu_classifier.p_logit
primals_37 = self.conc_drop_logvar_classifier.p_logit
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22, primals_23, primals_24,
primals_25, primals_26, primals_27, primals_28, primals_29,
primals_30, primals_31, primals_32, primals_33, primals_34,
primals_35, primals_36, primals_37])
return output[0], output[1], output[2], output[3], output[4], output[5
], output[6], output[7], output[8], output[9]
|
jiwoncpark/fast-forward
|
ConcreteDenseMixture
| false | 10,448 |
[
"MIT"
] | 0 |
640a521241a8756be2a0d42282e88d56a2290fca
|
https://github.com/jiwoncpark/fast-forward/tree/640a521241a8756be2a0d42282e88d56a2290fca
|
FirstBlock
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ro/crodqpemvzed7dwj6ikboerwmi47yfap7jg5xgrgodxuptzllrip.py
# Topologically Sorted Source Nodes: [mul, leaky_relu], Original ATen: [aten.mul, aten.leaky_relu, aten.leaky_relu_backward]
# Source node to ATen node mapping:
# leaky_relu => gt, mul_1, where
# mul => mul
# Graph fragment:
# %mul : [num_users=3] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 1.0), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%mul, 0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, 0.2), kwargs = {})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul, %mul_1), kwargs = {})
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%where, 0), kwargs = {})
triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0 = async_compile.triton('triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0(in_out_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = tmp7 > tmp3
tl.store(in_out_ptr0 + (x0), tmp7, xmask)
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_2, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [mul, leaky_relu], Original ATen: [aten.mul, aten.leaky_relu, aten.leaky_relu_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0.run(buf1, buf2, 256, grid=grid(256), stream=stream0)
return (buf1, primals_1, primals_2, buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import numpy as np
import torch.nn as nn
class BatchNormLayer(nn.Module):
"""Implements batch normalization layer."""
def __init__(self, channels, gamma=False, beta=True, decay=0.9, epsilon
=1e-05):
"""Initializes with basic settings.
Args:
channels: Number of channels of the input tensor.
gamma: Whether the scale (weight) of the affine mapping is learnable.
beta: Whether the center (bias) of the affine mapping is learnable.
decay: Decay factor for moving average operations in this layer.
epsilon: A value added to the denominator for numerical stability.
"""
super().__init__()
self.bn = nn.BatchNorm2d(num_features=channels, affine=True,
track_running_stats=True, momentum=1 - decay, eps=epsilon)
self.bn.weight.requires_grad = gamma
self.bn.bias.requires_grad = beta
def forward(self, x):
return self.bn(x)
class FirstBlock(nn.Module):
"""Implements the first block, which is a convolutional block."""
def __init__(self, in_channels, out_channels, use_wscale=False,
wscale_gain=np.sqrt(2.0), use_bn=False, activation_type='lrelu'):
super().__init__()
self.conv = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=3, stride=1, padding=1, bias=False)
self.scale = wscale_gain / np.sqrt(in_channels * 3 * 3
) if use_wscale else 1.0
self.bn = BatchNormLayer(channels=out_channels
) if use_bn else nn.Identity()
if activation_type == 'linear':
self.activate = nn.Identity()
elif activation_type == 'lrelu':
self.activate = nn.LeakyReLU(negative_slope=0.2, inplace=True)
else:
raise NotImplementedError(
f'Not implemented activation function: {activation_type}!')
def forward(self, x):
return self.activate(self.bn(self.conv(x) * self.scale))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import numpy as np
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0(in_out_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = tmp7 > tmp3
tl.store(in_out_ptr0 + x0, tmp7, xmask)
tl.store(out_ptr0 + x0, tmp8, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_2, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_leaky_relu_leaky_relu_backward_mul_0[grid(256)](buf1,
buf2, 256, XBLOCK=256, num_warps=4, num_stages=1)
return buf1, primals_1, primals_2, buf2
class BatchNormLayer(nn.Module):
"""Implements batch normalization layer."""
def __init__(self, channels, gamma=False, beta=True, decay=0.9, epsilon
=1e-05):
"""Initializes with basic settings.
Args:
channels: Number of channels of the input tensor.
gamma: Whether the scale (weight) of the affine mapping is learnable.
beta: Whether the center (bias) of the affine mapping is learnable.
decay: Decay factor for moving average operations in this layer.
epsilon: A value added to the denominator for numerical stability.
"""
super().__init__()
self.bn = nn.BatchNorm2d(num_features=channels, affine=True,
track_running_stats=True, momentum=1 - decay, eps=epsilon)
self.bn.weight.requires_grad = gamma
self.bn.bias.requires_grad = beta
def forward(self, x):
return self.bn(x)
class FirstBlockNew(nn.Module):
"""Implements the first block, which is a convolutional block."""
def __init__(self, in_channels, out_channels, use_wscale=False,
wscale_gain=np.sqrt(2.0), use_bn=False, activation_type='lrelu'):
super().__init__()
self.conv = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=3, stride=1, padding=1, bias=False)
self.scale = wscale_gain / np.sqrt(in_channels * 3 * 3
) if use_wscale else 1.0
self.bn = BatchNormLayer(channels=out_channels
) if use_bn else nn.Identity()
if activation_type == 'linear':
self.activate = nn.Identity()
elif activation_type == 'lrelu':
self.activate = nn.LeakyReLU(negative_slope=0.2, inplace=True)
else:
raise NotImplementedError(
f'Not implemented activation function: {activation_type}!')
def forward(self, input_0):
primals_1 = self.conv.weight
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
|
lelechen63/idinvert_pytorch
|
FirstBlock
| false | 10,449 |
[
"MIT"
] | 0 |
0469e1e5460ee4dd626c05bd35a83d52f9dc2cac
|
https://github.com/lelechen63/idinvert_pytorch/tree/0469e1e5460ee4dd626c05bd35a83d52f9dc2cac
|
LastBlock
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ey/cey4yc74rqvkmtcxtse2vt3dw6pfdi3zwtwezx7cdzkykhz4kzp7.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# x_1 => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mm, 1.0), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.mm]
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(buf1, 16, grid=grid(16), stream=stream0)
return (buf1, primals_1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import numpy as np
import torch.nn as nn
class BatchNormLayer(nn.Module):
"""Implements batch normalization layer."""
def __init__(self, channels, gamma=False, beta=True, decay=0.9, epsilon
=1e-05):
"""Initializes with basic settings.
Args:
channels: Number of channels of the input tensor.
gamma: Whether the scale (weight) of the affine mapping is learnable.
beta: Whether the center (bias) of the affine mapping is learnable.
decay: Decay factor for moving average operations in this layer.
epsilon: A value added to the denominator for numerical stability.
"""
super().__init__()
self.bn = nn.BatchNorm2d(num_features=channels, affine=True,
track_running_stats=True, momentum=1 - decay, eps=epsilon)
self.bn.weight.requires_grad = gamma
self.bn.bias.requires_grad = beta
def forward(self, x):
return self.bn(x)
class LastBlock(nn.Module):
"""Implements the last block, which is a dense block."""
def __init__(self, in_channels, out_channels, use_wscale=False,
wscale_gain=1.0, use_bn=False):
super().__init__()
self.fc = nn.Linear(in_features=in_channels, out_features=
out_channels, bias=False)
self.scale = wscale_gain / np.sqrt(in_channels) if use_wscale else 1.0
self.bn = BatchNormLayer(channels=out_channels
) if use_bn else nn.Identity()
def forward(self, x):
x = x.view(x.shape[0], -1)
x = self.fc(x) * self.scale
x = x.view(x.shape[0], x.shape[1], 1, 1)
return self.bn(x).view(x.shape[0], x.shape[1])
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import numpy as np
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_0(in_out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(in_out_ptr0 + x0, tmp2, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (4, 4),
(1, 4), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_mul_0[grid(16)](buf1, 16, XBLOCK=16, num_warps=1,
num_stages=1)
return buf1, primals_1
class BatchNormLayer(nn.Module):
"""Implements batch normalization layer."""
def __init__(self, channels, gamma=False, beta=True, decay=0.9, epsilon
=1e-05):
"""Initializes with basic settings.
Args:
channels: Number of channels of the input tensor.
gamma: Whether the scale (weight) of the affine mapping is learnable.
beta: Whether the center (bias) of the affine mapping is learnable.
decay: Decay factor for moving average operations in this layer.
epsilon: A value added to the denominator for numerical stability.
"""
super().__init__()
self.bn = nn.BatchNorm2d(num_features=channels, affine=True,
track_running_stats=True, momentum=1 - decay, eps=epsilon)
self.bn.weight.requires_grad = gamma
self.bn.bias.requires_grad = beta
def forward(self, x):
return self.bn(x)
class LastBlockNew(nn.Module):
"""Implements the last block, which is a dense block."""
def __init__(self, in_channels, out_channels, use_wscale=False,
wscale_gain=1.0, use_bn=False):
super().__init__()
self.fc = nn.Linear(in_features=in_channels, out_features=
out_channels, bias=False)
self.scale = wscale_gain / np.sqrt(in_channels) if use_wscale else 1.0
self.bn = BatchNormLayer(channels=out_channels
) if use_bn else nn.Identity()
def forward(self, input_0):
primals_1 = self.fc.weight
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
|
lelechen63/idinvert_pytorch
|
LastBlock
| false | 10,450 |
[
"MIT"
] | 0 |
0469e1e5460ee4dd626c05bd35a83d52f9dc2cac
|
https://github.com/lelechen63/idinvert_pytorch/tree/0469e1e5460ee4dd626c05bd35a83d52f9dc2cac
|
MLP
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/w3/cw3egt7ajdde7mbqzrdxs4mdcaxj75b4l3brz5gbsf4yd73gbids.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_1 => relu
# Graph fragment:
# %add_tensor_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_2, %primals_3), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_2,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2048
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 512
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/y2/cy2lwgz7dq2q2z4ifepdde4l7vyyvrwcx4zjn2ezmtzcanvhv374.py
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_2 => relu_1
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_5), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_1 = async_compile.triton('triton_poi_fused_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/dm/cdmk5acvohhdeyjsc6lsgsurbqis3xmgjftzfmanusy52dkygd23.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_3 => relu_2
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_7), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_2 = async_compile.triton('triton_poi_fused_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (512, 784), (784, 1))
assert_size_stride(primals_3, (512, ), (1, ))
assert_size_stride(primals_4, (256, 512), (512, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (64, 256), (256, 1))
assert_size_stride(primals_7, (64, ), (1, ))
assert_size_stride(primals_8, (10, 64), (64, 1))
assert_size_stride(primals_9, (10, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 512), (512, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784, 512), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 2048, grid=grid(2048), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 256), (256, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (512, 256), (1, 512), 0), out=buf2)
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf3, primals_5, 1024, grid=grid(1024), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 64), (64, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf3, reinterpret_tensor(primals_6, (256, 64), (1, 256), 0), out=buf4)
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf5, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
buf6 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, buf5, reinterpret_tensor(primals_8, (64, 10), (1, 64), 0), alpha=1, beta=1, out=buf6)
del primals_9
return (buf6, primals_1, buf1, buf3, buf5, primals_8, primals_6, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((512, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((512, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((256, 512), (512, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((64, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((10, 64), (64, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((10, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class MLP(nn.Module):
def __init__(self):
super(MLP, self).__init__()
self.fc1 = nn.Linear(28 * 28, 512)
self.fc2 = nn.Linear(512, 256)
self.fc3 = nn.Linear(256, 64)
self.fc4 = nn.Linear(64, 10)
self.dropout = nn.Dropout(0.1)
def forward(self, x):
x = x.view(-1, 28 * 28)
x = F.relu(self.fc1(x))
self.dropout(x)
x = F.relu(self.fc2(x))
self.dropout(x)
x = F.relu(self.fc3(x))
self.dropout(x)
x = self.fc4(x)
return x
def get_inputs():
return [torch.rand([4, 784])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 512
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_relu_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (512, 784), (784, 1))
assert_size_stride(primals_3, (512,), (1,))
assert_size_stride(primals_4, (256, 512), (512, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (64, 256), (256, 1))
assert_size_stride(primals_7, (64,), (1,))
assert_size_stride(primals_8, (10, 64), (64, 1))
assert_size_stride(primals_9, (10,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 512), (512, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784,
512), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(2048)](buf1, primals_3, 2048, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 256), (256, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (512, 256), (
1, 512), 0), out=buf2)
buf3 = buf2
del buf2
triton_poi_fused_relu_1[grid(1024)](buf3, primals_5, 1024, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 64), (64, 1), torch.float32)
extern_kernels.mm(buf3, reinterpret_tensor(primals_6, (256, 64), (1,
256), 0), out=buf4)
buf5 = buf4
del buf4
triton_poi_fused_relu_2[grid(256)](buf5, primals_7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_7
buf6 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
extern_kernels.addmm(primals_9, buf5, reinterpret_tensor(primals_8,
(64, 10), (1, 64), 0), alpha=1, beta=1, out=buf6)
del primals_9
return buf6, primals_1, buf1, buf3, buf5, primals_8, primals_6, primals_4
class MLPNew(nn.Module):
def __init__(self):
super(MLPNew, self).__init__()
self.fc1 = nn.Linear(28 * 28, 512)
self.fc2 = nn.Linear(512, 256)
self.fc3 = nn.Linear(256, 64)
self.fc4 = nn.Linear(64, 10)
self.dropout = nn.Dropout(0.1)
def forward(self, input_0):
primals_2 = self.fc1.weight
primals_3 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.fc3.weight
primals_7 = self.fc3.bias
primals_8 = self.fc4.weight
primals_9 = self.fc4.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
|
lynfi/Classification
|
MLP
| false | 10,451 |
[
"MIT"
] | 0 |
691731629c6577432c8c9eee70b67911011a07b7
|
https://github.com/lynfi/Classification/tree/691731629c6577432c8c9eee70b67911011a07b7
|
CausalConv1d
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/fk/cfkrz7kxwztwd7brbtic3lrvcvxemi6b7x3jqgbfiyfesqxvlh2f.py
# Topologically Sorted Source Nodes: [pad, x], Original ATen: [aten.constant_pad_nd, aten.squeeze]
# Source node to ATen node mapping:
# pad => constant_pad_nd
# x => squeeze
# Graph fragment:
# %constant_pad_nd : [num_users=1] = call_function[target=torch.ops.aten.constant_pad_nd.default](args = (%unsqueeze, [3, 0, 0, 0], 0.0), kwargs = {})
# %squeeze : [num_users=2] = call_function[target=torch.ops.aten.squeeze.dim](args = (%constant_pad_nd, 2), kwargs = {})
triton_poi_fused_constant_pad_nd_squeeze_0 = async_compile.triton('triton_poi_fused_constant_pad_nd_squeeze_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_constant_pad_nd_squeeze_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_constant_pad_nd_squeeze_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = (-3) + x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.load(in_ptr0 + (x1), tmp2 & xmask, eviction_policy='evict_last', other=0.0)
tl.store(out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tc/ctcagp37ljugm52zu6ckorigrppqo67voefe2f2odg5r6hyllhyu.py
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv1d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%squeeze, %primals_2, %primals_3, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pad, x], Original ATen: [aten.constant_pad_nd, aten.squeeze]
stream0 = get_raw_stream(0)
triton_poi_fused_constant_pad_nd_squeeze_0.run(primals_1, buf0, 64, grid=grid(64), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1), (4, 1, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv1d], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 16, grid=grid(16), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class CausalConv1d(nn.Conv1d):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
dilation=1, groups=1, bias=True):
super(CausalConv1d, self).__init__(in_channels, out_channels,
kernel_size, stride=stride, padding=0, dilation=dilation,
groups=groups, bias=bias)
self.left_padding = dilation * (kernel_size - 1)
def forward(self, input):
x = F.pad(input.unsqueeze(2), (self.left_padding, 0, 0, 0)).squeeze(2)
return super(CausalConv1d, self).forward(x)
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_constant_pad_nd_squeeze_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = -3 + x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.load(in_ptr0 + x1, tmp2 & xmask, eviction_policy='evict_last',
other=0.0)
tl.store(out_ptr0 + x2, tmp3, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_constant_pad_nd_squeeze_0[grid(64)](primals_1,
buf0, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1), (4, 1, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(16)](buf2, primals_3, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_3
return buf2, primals_2, buf0
class CausalConv1dNew(nn.Conv1d):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
dilation=1, groups=1, bias=True):
super(CausalConv1dNew, self).__init__(in_channels, out_channels,
kernel_size, stride=stride, padding=0, dilation=dilation,
groups=groups, bias=bias)
self.left_padding = dilation * (kernel_size - 1)
def forward(self, input_0):
primals_2 = self.weight
primals_3 = self.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
marc-moreaux/pytorch_text_generator
|
CausalConv1d
| false | 10,452 |
[
"MIT"
] | 0 |
99dd11c67d89f8a09faa28b7032fcc66f90672c0
|
https://github.com/marc-moreaux/pytorch_text_generator/tree/99dd11c67d89f8a09faa28b7032fcc66f90672c0
|
PositionwiseFeedForward
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/r3/cr3febcwm3t44fuoitsx3ou2p6xg4sk4f7unagmmrvffasxf47te.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ji/cji7mw45fbdoanjc5e6qu3e2bf5d6jnnjabskl6onjlk7uv7oqud.py
# Topologically Sorted Source Nodes: [add, output_2], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# output_2 => var_mean
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %primals_1), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add, [3]), kwargs = {correction: 0, keepdim: True})
triton_poi_fused_add_native_layer_norm_1 = async_compile.triton('triton_poi_fused_add_native_layer_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_1(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + (x0), tmp16, xmask)
tl.store(out_ptr1 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xy/cxyvzp6lij7d3yqq2ut3vi6guk7xnzb7qwqb66dthlly44r65vfk.py
# Topologically Sorted Source Nodes: [add, output_2], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# output_2 => add_1, add_2, mul, mul_1, rsqrt, sub
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %primals_1), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %getitem_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %primals_6), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_7), kwargs = {})
triton_poi_fused_add_native_layer_norm_2 = async_compile.triton('triton_poi_fused_add_native_layer_norm_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + (x2), tmp13, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_3, buf6, 256, grid=grid(256), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [add, output_2], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_1.run(buf2, primals_1, buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, output_2], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_2.run(buf2, primals_1, buf3, buf4, primals_6, primals_7, buf5, 256, grid=grid(256), stream=stream0)
del buf3
del buf4
del primals_7
return (buf5, primals_1, primals_6, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), buf2, primals_4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.functional import F
from torch.nn import functional as F
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
class PositionwiseFeedForward(nn.Module):
"""Implements position-wise feedforward sublayer.
FFN(x) = max(0, xW1 + b1)W2 + b2
"""
def __init__(self, d_model, d_ff, dropout=0.1):
super(PositionwiseFeedForward, self).__init__()
self.w_1 = nn.Linear(d_model, d_ff)
self.w_2 = nn.Linear(d_ff, d_model)
self.dropout = nn.Dropout(dropout)
self.layer_norm = nn.LayerNorm(d_model)
def forward(self, x):
residual = x
output = self.w_2(F.relu(self.w_1(x)))
output = self.dropout(output)
output = self.layer_norm(output + residual)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'d_model': 4, 'd_ff': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_1(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + x0, tmp16, xmask)
tl.store(out_ptr1 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + x2, tmp13, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(256)](buf1,
primals_3, buf6, 256, XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused_add_native_layer_norm_1[grid(64)](buf2, primals_1,
buf3, buf4, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_native_layer_norm_2[grid(256)](buf2, primals_1,
buf3, buf4, primals_6, primals_7, buf5, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf3
del buf4
del primals_7
return buf5, primals_1, primals_6, reinterpret_tensor(buf1, (64, 4), (4,
1), 0), buf2, primals_4, buf6
class PositionwiseFeedForwardNew(nn.Module):
"""Implements position-wise feedforward sublayer.
FFN(x) = max(0, xW1 + b1)W2 + b2
"""
def __init__(self, d_model, d_ff, dropout=0.1):
super(PositionwiseFeedForwardNew, self).__init__()
self.w_1 = nn.Linear(d_model, d_ff)
self.w_2 = nn.Linear(d_ff, d_model)
self.dropout = nn.Dropout(dropout)
self.layer_norm = nn.LayerNorm(d_model)
def forward(self, input_0):
primals_2 = self.w_1.weight
primals_3 = self.w_1.bias
primals_4 = self.w_2.weight
primals_5 = self.w_2.bias
primals_6 = self.layer_norm.weight
primals_7 = self.layer_norm.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
greenstar1151/pytorch-benchmark
|
PositionwiseFeedForward
| false | 10,453 |
[
"BSD-3-Clause"
] | 0 |
8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
https://github.com/greenstar1151/pytorch-benchmark/tree/8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
Decoder
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/se/cse7sfgf5sgpbvm22egbkgr6e26lmuzgp3epjbv2y2kstod536nr.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_2), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
# %le_3 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/rj/crjqubz53turnzr3eytn5bygdu3aanhnbasxm5kv2343ef6srmqz.py
# Topologically Sorted Source Nodes: [conv_transpose2d, x_2], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv_transpose2d => convolution
# x_2 => relu_1
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%unsqueeze_1, %primals_4, %primals_5, [2, 2], [0, 0], [1, 1], True, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_1 = async_compile.triton('triton_poi_fused_convolution_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 25) % 128
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7j/c7jlcpfayfgm4ionvsotdwcpvr26emperi3u36tmlugpw3iltf26.py
# Topologically Sorted Source Nodes: [conv_transpose2d_1, x_3], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv_transpose2d_1 => convolution_1
# x_3 => relu_2
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_1, %primals_6, %primals_7, [2, 2], [0, 0], [1, 1], True, [0, 0], 1), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_1,), kwargs = {})
triton_poi_fused_convolution_relu_2 = async_compile.triton('triton_poi_fused_convolution_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 43264
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 169) % 64
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/d6/cd6lk7yvfcgve23ubphsppwnxecscgcqsmgwcnmcf7e736rcj3do.py
# Topologically Sorted Source Nodes: [conv_transpose2d_2, x_4], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv_transpose2d_2 => convolution_2
# x_4 => relu_3
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_2, %primals_8, %primals_9, [2, 2], [0, 0], [1, 1], True, [0, 0], 1), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_2,), kwargs = {})
triton_poi_fused_convolution_relu_3 = async_compile.triton('triton_poi_fused_convolution_relu_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[131072],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_3', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_3(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 115200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 900) % 32
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hs/chsgbq3hqknrxoqdmknsp4fh57w5pvfawifuk6rnybsj7hi324s7.py
# Topologically Sorted Source Nodes: [conv_transpose2d_3, reconstr], Original ATen: [aten.convolution, aten.sigmoid]
# Source node to ATen node mapping:
# conv_transpose2d_3 => convolution_3
# reconstr => sigmoid
# Graph fragment:
# %convolution_3 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_3, %primals_10, %primals_11, [2, 2], [0, 0], [1, 1], True, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_3,), kwargs = {})
triton_poi_fused_convolution_sigmoid_4 = async_compile.triton('triton_poi_fused_convolution_sigmoid_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_sigmoid_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_sigmoid_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), None)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + (x0), tmp4, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 128, 5, 5), (3200, 25, 5, 1))
assert_size_stride(primals_5, (128, ), (1, ))
assert_size_stride(primals_6, (128, 64, 5, 5), (1600, 25, 5, 1))
assert_size_stride(primals_7, (64, ), (1, ))
assert_size_stride(primals_8, (64, 32, 6, 6), (1152, 36, 6, 1))
assert_size_stride(primals_9, (32, ), (1, ))
assert_size_stride(primals_10, (32, 1, 6, 6), (36, 36, 6, 1))
assert_size_stride(primals_11, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = buf0; del buf0 # reuse
buf10 = empty_strided_cuda((4, 4), (4, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf10, 16, grid=grid(16), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [conv_transpose2d], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(reinterpret_tensor(buf1, (4, 4, 1, 1), (4, 1, 0, 0), 0), primals_4, stride=(2, 2), padding=(0, 0), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 128, 5, 5), (3200, 25, 5, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv_transpose2d, x_2], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_1.run(buf3, primals_5, 12800, grid=grid(12800), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [conv_transpose2d_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(2, 2), padding=(0, 0), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 64, 13, 13), (10816, 169, 13, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [conv_transpose2d_1, x_3], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_2.run(buf5, primals_7, 43264, grid=grid(43264), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [conv_transpose2d_2], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(2, 2), padding=(0, 0), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 32, 30, 30), (28800, 900, 30, 1))
buf7 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [conv_transpose2d_2, x_4], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_3.run(buf7, primals_9, 115200, grid=grid(115200), stream=stream0)
del primals_9
# Topologically Sorted Source Nodes: [conv_transpose2d_3], Original ATen: [aten.convolution]
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(2, 2), padding=(0, 0), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 1, 64, 64), (4096, 4096, 64, 1))
buf9 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [conv_transpose2d_3, reconstr], Original ATen: [aten.convolution, aten.sigmoid]
triton_poi_fused_convolution_sigmoid_4.run(buf9, primals_11, 16384, grid=grid(16384), stream=stream0)
del primals_11
return (buf9, primals_3, primals_4, primals_6, primals_8, primals_10, reinterpret_tensor(buf1, (4, 4, 1, 1), (4, 1, 1, 1), 0), buf3, buf5, buf7, buf9, buf10, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 128, 5, 5), (3200, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((128, 64, 5, 5), (1600, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((64, 32, 6, 6), (1152, 36, 6, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((32, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((32, 1, 6, 6), (36, 36, 6, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class Decoder(nn.Module):
def __init__(self, latent_size, m):
super(Decoder, self).__init__()
self.latent_size = latent_size
self.fc = nn.Linear(latent_size, m)
self.deconv1 = nn.ConvTranspose2d(m, 128, 5, stride=2)
self.deconv2 = nn.ConvTranspose2d(128, 64, 5, stride=2)
self.deconv3 = nn.ConvTranspose2d(64, 32, 6, stride=2)
self.deconv4 = nn.ConvTranspose2d(32, 1, 6, stride=2)
def forward(self, x):
x = torch.relu(self.fc(x))
x = x.unsqueeze(-1).unsqueeze(-1)
x = torch.relu(self.deconv1(x))
x = torch.relu(self.deconv2(x))
x = torch.relu(self.deconv3(x))
reconstr = torch.sigmoid(self.deconv4(x))
return reconstr
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'latent_size': 4, 'm': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 25 % 128
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 43264
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 169 % 64
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_3(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 115200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 900 % 32
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_sigmoid_4(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, None)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + x0, tmp4, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 128, 5, 5), (3200, 25, 5, 1))
assert_size_stride(primals_5, (128,), (1,))
assert_size_stride(primals_6, (128, 64, 5, 5), (1600, 25, 5, 1))
assert_size_stride(primals_7, (64,), (1,))
assert_size_stride(primals_8, (64, 32, 6, 6), (1152, 36, 6, 1))
assert_size_stride(primals_9, (32,), (1,))
assert_size_stride(primals_10, (32, 1, 6, 6), (36, 36, 6, 1))
assert_size_stride(primals_11, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 4),
(1, 4), 0), out=buf0)
del primals_1
buf1 = buf0
del buf0
buf10 = empty_strided_cuda((4, 4), (4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(16)](buf1,
primals_2, buf10, 16, XBLOCK=16, num_warps=1, num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(reinterpret_tensor(buf1, (4, 4, 1,
1), (4, 1, 0, 0), 0), primals_4, stride=(2, 2), padding=(0, 0),
dilation=(1, 1), transposed=True, output_padding=(0, 0), groups
=1, bias=None)
assert_size_stride(buf2, (4, 128, 5, 5), (3200, 25, 5, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_relu_1[grid(12800)](buf3, primals_5,
12800, XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(2, 2),
padding=(0, 0), dilation=(1, 1), transposed=True,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 64, 13, 13), (10816, 169, 13, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_relu_2[grid(43264)](buf5, primals_7,
43264, XBLOCK=512, num_warps=4, num_stages=1)
del primals_7
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(2, 2),
padding=(0, 0), dilation=(1, 1), transposed=True,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 32, 30, 30), (28800, 900, 30, 1))
buf7 = buf6
del buf6
triton_poi_fused_convolution_relu_3[grid(115200)](buf7, primals_9,
115200, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_9
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(2, 2),
padding=(0, 0), dilation=(1, 1), transposed=True,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 1, 64, 64), (4096, 4096, 64, 1))
buf9 = buf8
del buf8
triton_poi_fused_convolution_sigmoid_4[grid(16384)](buf9,
primals_11, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del primals_11
return (buf9, primals_3, primals_4, primals_6, primals_8, primals_10,
reinterpret_tensor(buf1, (4, 4, 1, 1), (4, 1, 1, 1), 0), buf3, buf5,
buf7, buf9, buf10)
class DecoderNew(nn.Module):
def __init__(self, latent_size, m):
super(DecoderNew, self).__init__()
self.latent_size = latent_size
self.fc = nn.Linear(latent_size, m)
self.deconv1 = nn.ConvTranspose2d(m, 128, 5, stride=2)
self.deconv2 = nn.ConvTranspose2d(128, 64, 5, stride=2)
self.deconv3 = nn.ConvTranspose2d(64, 32, 6, stride=2)
self.deconv4 = nn.ConvTranspose2d(32, 1, 6, stride=2)
def forward(self, input_0):
primals_1 = self.fc.weight
primals_2 = self.fc.bias
primals_4 = self.deconv1.weight
primals_5 = self.deconv1.bias
primals_6 = self.deconv2.weight
primals_7 = self.deconv2.bias
primals_8 = self.deconv3.weight
primals_9 = self.deconv3.bias
primals_10 = self.deconv4.weight
primals_11 = self.deconv4.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
|
lshoek/creative-evo-controller
|
Decoder
| false | 10,454 |
[
"MIT"
] | 0 |
a5f1742c172255cca2338b76ae1c5b4db277bb0d
|
https://github.com/lshoek/creative-evo-controller/tree/a5f1742c172255cca2338b76ae1c5b4db277bb0d
|
TemporalConv
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/6s/c6saluonoqfxerrd3p7h7p6inhepj24zjy54kxuetr27abkw2rto.py
# Topologically Sorted Source Nodes: [P, conv2d_1, conv2d_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# P => convolution
# conv2d_1 => convolution_1
# conv2d_2 => convolution_2
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 16], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_ptr0, out_ptr0, out_ptr1, out_ptr2, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex % 4
x3 = (xindex // 4)
y0 = yindex % 4
y1 = (yindex // 4)
x5 = xindex
y4 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x3) + (16*x2) + (64*y1)), xmask & ymask)
tl.store(out_ptr0 + (x5 + (16*y4)), tmp0, xmask & ymask)
tl.store(out_ptr1 + (x5 + (16*y4)), tmp0, xmask & ymask)
tl.store(out_ptr2 + (x5 + (16*y4)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ba/cbaejeje4ctf6bizvw6ak33yknc5bygfyu5seecmofky7uxsvxmu.py
# Topologically Sorted Source Nodes: [P, conv2d_1, Q, PQ, conv2d_2, add_1, out], Original ATen: [aten.convolution, aten.sigmoid, aten.add, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# P => convolution
# PQ => add
# Q => sigmoid
# add_1 => add_1
# conv2d_1 => convolution_1
# conv2d_2 => convolution_2
# out => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_1,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convolution, %sigmoid), kwargs = {})
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%permute, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %convolution_2), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1 = async_compile.triton('triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*i1', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 8) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + (x3), xmask)
tmp4 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr2 + (x3), xmask)
tmp9 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tl.sigmoid(tmp2)
tmp7 = tmp5 + tmp6
tmp10 = tmp8 + tmp9
tmp11 = tmp7 + tmp10
tmp12 = tl.full([1], 0, tl.int32)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tmp14 = 0.0
tmp15 = tmp13 <= tmp14
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
tl.store(in_out_ptr1 + (x3), tmp13, xmask)
tl.store(out_ptr0 + (x3), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [P, conv2d_1, conv2d_2], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(primals_1, buf0, buf2, buf5, 16, 16, grid=grid(16, 16), stream=stream0)
# Topologically Sorted Source Nodes: [P], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 2), (32, 8, 2, 1))
del buf0
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4, 2), (32, 8, 2, 1))
del buf2
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 2), (32, 8, 2, 1))
del buf5
buf4 = buf3; del buf3 # reuse
buf7 = buf1; del buf1 # reuse
buf8 = empty_strided_cuda((4, 4, 4, 2), (32, 8, 2, 1), torch.bool)
# Topologically Sorted Source Nodes: [P, conv2d_1, Q, PQ, conv2d_2, add_1, out], Original ATen: [aten.convolution, aten.sigmoid, aten.add, aten.relu, aten.threshold_backward]
triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1.run(buf4, buf7, primals_5, primals_3, buf6, primals_7, buf8, 128, grid=grid(128), stream=stream0)
del buf6
del primals_3
del primals_5
del primals_7
return (reinterpret_tensor(buf7, (4, 2, 4, 4), (32, 1, 2, 8), 0), primals_2, primals_4, primals_6, reinterpret_tensor(primals_1, (4, 4, 4, 4), (64, 1, 4, 16), 0), buf4, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 1, 3), (12, 3, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 1, 3), (12, 3, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 3), (12, 3, 3, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
import torch.nn as nn
class TemporalConv(nn.Module):
"""Temporal convolution block applied to nodes in the STGCN Layer
For details see: `"Spatio-Temporal Graph Convolutional Networks:
A Deep Learning Framework for Traffic Forecasting"
<https://arxiv.org/abs/1709.04875>`_
Based off the temporal convolution introduced in "Convolutional
Sequence to Sequence Learning" <https://arxiv.org/abs/1709.04875>`_
NB. Given an input sequence of length m and a kernel size of k
the output sequence will have length m-(k-1)
Args:
in_channels (int): Number of input features.
out_channels (int): Number of output features.
kernel_size (int): Convolutional kernel size.
"""
def __init__(self, in_channels, out_channels, kernel_size=3):
super(TemporalConv, self).__init__()
self.conv1 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
self.conv2 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
self.conv3 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
def forward(self, X):
"""Forward pass through temporal convolution block
Args:
X (torch.Tensor): Input data of shape
(batch_size, input_time_steps, num_nodes, in_channels)
"""
X = X.permute(0, 3, 2, 1)
P = self.conv1(X)
Q = torch.sigmoid(self.conv2(X))
PQ = P + Q
out = F.relu(PQ + self.conv3(X))
out = out.permute(0, 3, 2, 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_ptr0, out_ptr0, out_ptr1, out_ptr2,
ynumel, xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex % 4
x3 = xindex // 4
y0 = yindex % 4
y1 = yindex // 4
x5 = xindex
y4 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x3 + 16 * x2 + 64 * y1), xmask & ymask)
tl.store(out_ptr0 + (x5 + 16 * y4), tmp0, xmask & ymask)
tl.store(out_ptr1 + (x5 + 16 * y4), tmp0, xmask & ymask)
tl.store(out_ptr2 + (x5 + 16 * y4), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1(
in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 8 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + x3, xmask)
tmp4 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr2 + x3, xmask)
tmp9 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tl.sigmoid(tmp2)
tmp7 = tmp5 + tmp6
tmp10 = tmp8 + tmp9
tmp11 = tmp7 + tmp10
tmp12 = tl.full([1], 0, tl.int32)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tmp14 = 0.0
tmp15 = tmp13 <= tmp14
tl.store(in_out_ptr0 + x3, tmp2, xmask)
tl.store(in_out_ptr1 + x3, tmp13, xmask)
tl.store(out_ptr0 + x3, tmp15, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 3), (12, 3, 3, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(16, 16)](primals_1, buf0, buf2,
buf5, 16, 16, XBLOCK=16, YBLOCK=16, num_warps=4, num_stages=1)
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 2), (32, 8, 2, 1))
del buf0
buf3 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4, 2), (32, 8, 2, 1))
del buf2
buf6 = extern_kernels.convolution(buf5, primals_6, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 2), (32, 8, 2, 1))
del buf5
buf4 = buf3
del buf3
buf7 = buf1
del buf1
buf8 = empty_strided_cuda((4, 4, 4, 2), (32, 8, 2, 1), torch.bool)
triton_poi_fused_add_convolution_relu_sigmoid_threshold_backward_1[grid
(128)](buf4, buf7, primals_5, primals_3, buf6, primals_7, buf8,
128, XBLOCK=128, num_warps=4, num_stages=1)
del buf6
del primals_3
del primals_5
del primals_7
return reinterpret_tensor(buf7, (4, 2, 4, 4), (32, 1, 2, 8), 0
), primals_2, primals_4, primals_6, reinterpret_tensor(primals_1, (
4, 4, 4, 4), (64, 1, 4, 16), 0), buf4, buf8
class TemporalConvNew(nn.Module):
"""Temporal convolution block applied to nodes in the STGCN Layer
For details see: `"Spatio-Temporal Graph Convolutional Networks:
A Deep Learning Framework for Traffic Forecasting"
<https://arxiv.org/abs/1709.04875>`_
Based off the temporal convolution introduced in "Convolutional
Sequence to Sequence Learning" <https://arxiv.org/abs/1709.04875>`_
NB. Given an input sequence of length m and a kernel size of k
the output sequence will have length m-(k-1)
Args:
in_channels (int): Number of input features.
out_channels (int): Number of output features.
kernel_size (int): Convolutional kernel size.
"""
def __init__(self, in_channels, out_channels, kernel_size=3):
super(TemporalConvNew, self).__init__()
self.conv1 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
self.conv2 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
self.conv3 = nn.Conv2d(in_channels, out_channels, (1, kernel_size))
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_3 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_6 = self.conv3.weight
primals_7 = self.conv3.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
marcdemers/pytorch_geometric_temporal
|
TemporalConv
| false | 10,455 |
[
"MIT"
] | 0 |
446aadcd890158bade2e9974f9840ed5a7bba827
|
https://github.com/marcdemers/pytorch_geometric_temporal/tree/446aadcd890158bade2e9974f9840ed5a7bba827
|
ResBlock
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/fp/cfp5jrxxyxrvhcpoq5tio3p5tkhj5ugdrpyur3x4v6meatzih7jn.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_0 = async_compile.triton('triton_poi_fused_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 4
y1 = (yindex // 4)
tmp0 = tl.load(in_ptr0 + (x2 + (16*y3)), xmask & ymask)
tl.store(out_ptr0 + (y0 + (4*x2) + (64*y1)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/lf/clf45mspfcg7t5x4om2snxq42eoe4jywsisc72sbpggbkipki6jb.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_1 = async_compile.triton('triton_poi_fused_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16384
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = (yindex // 128)
tmp0 = tl.load(in_ptr0 + (x2 + (9*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (128*x2) + (1152*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/j4/cj4ubfzqedljpxnezgsaqlcbmis75oaa7j6dddwnyaytki3in5gp.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# out => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_relu_2 = async_compile.triton('triton_poi_fused_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + (x0), tmp2, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/qo/cqofti6gsrehqlmzbkmbxoumphvdsb6r5tb6hgxmdqpgruph2ym4.py
# Topologically Sorted Source Nodes: [out_2, out_3, out_4], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# out_2 => relu_2
# out_3 => relu_3
# out_4 => add
# Graph fragment:
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_2,), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_3,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%relu_2, %relu_3), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_3, 0), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_2, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_3 = async_compile.triton('triton_poi_fused_add_relu_threshold_backward_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64, 256], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: '*i1', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_relu_threshold_backward_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, out_ptr2, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 256
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 16
y1 = (yindex // 16)
tmp0 = tl.load(in_ptr0 + (x2 + (256*y3)), xmask & ymask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2 + (256*y3)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.full([1, 1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp5 = tmp2 + tmp4
tmp6 = 0.0
tmp7 = tmp4 <= tmp6
tmp8 = tmp2 <= tmp6
tl.store(out_ptr0 + (y0 + (16*x2) + (4096*y1)), tmp5, xmask & ymask)
tl.store(out_ptr1 + (x2 + (256*y3)), tmp7, xmask & ymask)
tl.store(out_ptr2 + (x2 + (256*y3)), tmp8, xmask & ymask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (128, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_4, (256, 128, 1, 1), (128, 1, 1, 1))
assert_size_stride(primals_5, (256, 4, 1, 1), (4, 1, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 1, 16, 4), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
stream0 = get_raw_stream(0)
triton_poi_fused_0.run(primals_2, buf0, 16, 16, grid=grid(16, 16), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(primals_3, buf1, 16384, 9, grid=grid(16384, 9), stream=stream0)
del primals_3
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf0, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 128, 4, 4), (2048, 1, 512, 128))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf3, 8192, grid=grid(8192), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, buf1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 128, 4, 4), (2048, 1, 512, 128))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf5, 8192, grid=grid(8192), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 1, 1024, 256))
# Topologically Sorted Source Nodes: [conv2d_3], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(buf0, primals_5, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 256, 4, 4), (4096, 1, 1024, 256))
buf8 = empty_strided_cuda((4, 256, 4, 4), (4096, 16, 4, 1), torch.float32)
buf9 = empty_strided_cuda((4, 256, 4, 4), (4096, 1, 1024, 256), torch.bool)
buf10 = empty_strided_cuda((4, 256, 4, 4), (4096, 1, 1024, 256), torch.bool)
# Topologically Sorted Source Nodes: [out_2, out_3, out_4], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_3.run(buf6, buf7, buf8, buf9, buf10, 64, 256, grid=grid(64, 256), stream=stream0)
del buf6
del buf7
return (buf8, primals_1, buf0, buf1, primals_4, primals_5, buf3, buf5, buf9, buf10, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((128, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((256, 128, 1, 1), (128, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import init as init
class conv_relu(nn.Module):
"""docstring for conv_relu"""
def __init__(self, in_channels, out_channels, **kwargs):
super(conv_relu, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)
def forward(self, x):
out = F.relu(self.conv(x), inplace=True)
return out
class ResBlock(nn.Module):
"""docstring for ResBlock"""
def __init__(self, in_channels):
super(ResBlock, self).__init__()
self.res1a = conv_relu(in_channels, 128, kernel_size=1)
self.res1b = conv_relu(128, 128, kernel_size=3, padding=1)
self.res1c = conv_relu(128, 256, kernel_size=1)
self.res2a = conv_relu(in_channels, 256, kernel_size=1)
def forward(self, x):
out1 = self.res1a(x)
out1 = self.res1b(out1)
out1 = self.res1c(out1)
out2 = self.res2a(x)
out = out1 + out2
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import init as init
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 4
y1 = yindex // 4
tmp0 = tl.load(in_ptr0 + (x2 + 16 * y3), xmask & ymask)
tl.store(out_ptr0 + (y0 + 4 * x2 + 64 * y1), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = yindex // 128
tmp0 = tl.load(in_ptr0 + (x2 + 9 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 128 * x2 + 1152 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_relu_2(in_out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, None)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + x0, tmp2, None)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_3(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, out_ptr2, ynumel, xnumel, YBLOCK: tl.constexpr,
XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 256
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 16
y1 = yindex // 16
tmp0 = tl.load(in_ptr0 + (x2 + 256 * y3), xmask & ymask,
eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2 + 256 * y3), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.full([1, 1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp5 = tmp2 + tmp4
tmp6 = 0.0
tmp7 = tmp4 <= tmp6
tmp8 = tmp2 <= tmp6
tl.store(out_ptr0 + (y0 + 16 * x2 + 4096 * y1), tmp5, xmask & ymask)
tl.store(out_ptr1 + (x2 + 256 * y3), tmp7, xmask & ymask)
tl.store(out_ptr2 + (x2 + 256 * y3), tmp8, xmask & ymask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (128, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_4, (256, 128, 1, 1), (128, 1, 1, 1))
assert_size_stride(primals_5, (256, 4, 1, 1), (4, 1, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 1, 16, 4), torch.float32)
get_raw_stream(0)
triton_poi_fused_0[grid(16, 16)](primals_2, buf0, 16, 16, XBLOCK=16,
YBLOCK=16, num_warps=4, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_1[grid(16384, 9)](primals_3, buf1, 16384, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_3
buf2 = extern_kernels.convolution(buf0, primals_1, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 128, 4, 4), (2048, 1, 512, 128))
buf3 = buf2
del buf2
triton_poi_fused_relu_2[grid(8192)](buf3, 8192, XBLOCK=256,
num_warps=4, num_stages=1)
buf4 = extern_kernels.convolution(buf3, buf1, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 128, 4, 4), (2048, 1, 512, 128))
buf5 = buf4
del buf4
triton_poi_fused_relu_2[grid(8192)](buf5, 8192, XBLOCK=256,
num_warps=4, num_stages=1)
buf6 = extern_kernels.convolution(buf5, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 1, 1024, 256))
buf7 = extern_kernels.convolution(buf0, primals_5, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 256, 4, 4), (4096, 1, 1024, 256))
buf8 = empty_strided_cuda((4, 256, 4, 4), (4096, 16, 4, 1), torch.
float32)
buf9 = empty_strided_cuda((4, 256, 4, 4), (4096, 1, 1024, 256),
torch.bool)
buf10 = empty_strided_cuda((4, 256, 4, 4), (4096, 1, 1024, 256),
torch.bool)
triton_poi_fused_add_relu_threshold_backward_3[grid(64, 256)](buf6,
buf7, buf8, buf9, buf10, 64, 256, XBLOCK=32, YBLOCK=32,
num_warps=4, num_stages=1)
del buf6
del buf7
return (buf8, primals_1, buf0, buf1, primals_4, primals_5, buf3, buf5,
buf9, buf10)
class conv_relu(nn.Module):
"""docstring for conv_relu"""
def __init__(self, in_channels, out_channels, **kwargs):
super(conv_relu, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)
def forward(self, x):
out = F.relu(self.conv(x), inplace=True)
return out
class ResBlockNew(nn.Module):
"""docstring for ResBlock"""
def __init__(self, in_channels):
super(ResBlockNew, self).__init__()
self.res1a = conv_relu(in_channels, 128, kernel_size=1)
self.res1b = conv_relu(128, 128, kernel_size=3, padding=1)
self.res1c = conv_relu(128, 256, kernel_size=1)
self.res2a = conv_relu(in_channels, 256, kernel_size=1)
def forward(self, input_0):
primals_1 = self.res1a.conv.weight
primals_3 = self.res1b.conv.weight
primals_4 = self.res1c.conv.weight
primals_5 = self.res2a.conv.weight
primals_2 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
llpspark/PytorchToCaffe
|
ResBlock
| false | 10,456 |
[
"MIT"
] | 0 |
01f6fb2cfd42e2c06ae5d46a7a91f7fd6d40d5d1
|
https://github.com/llpspark/PytorchToCaffe/tree/01f6fb2cfd42e2c06ae5d46a7a91f7fd6d40d5d1
|
PairwiseRankingLoss
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/eq/ceqhkwm2d735ddv4kryb45afnds5mdj2jfkqyfgb53xuu6ykceh3.py
# Topologically Sorted Source Nodes: [sub, add, clamp, cost_sent, sub_1, add_1, clamp_1, cost_img, loss], Original ATen: [aten.rsub, aten.add, aten.clamp, aten.sum]
# Source node to ATen node mapping:
# add => add
# add_1 => add_1
# clamp => clamp_min
# clamp_1 => clamp_min_1
# cost_img => sum_2
# cost_sent => sum_1
# loss => add_2
# sub => sub
# sub_1 => sub_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (4, %arg0_1), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub, %arg1_1), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%add, 0.0), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%clamp_min,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (4, %arg2_1), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_1, %arg3_1), kwargs = {})
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%add_1, 0.0), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%clamp_min_1,), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, %sum_2), kwargs = {})
triton_per_fused_add_clamp_rsub_sum_0 = async_compile.triton('triton_per_fused_add_clamp_rsub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 6), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_clamp_rsub_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 4, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_clamp_rsub_sum_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp3 = tl.load(in_ptr1 + (r0), None)
tmp10 = tl.load(in_ptr2 + (r0), None)
tmp12 = tl.load(in_ptr3 + (r0), None)
tmp1 = 4.0
tmp2 = tmp1 - tmp0
tmp4 = tmp2 + tmp3
tmp5 = 0.0
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp7 = tl.broadcast_to(tmp6, [RBLOCK])
tmp9 = triton_helpers.promote_to_tensor(tl.sum(tmp7, 0))
tmp11 = tmp1 - tmp10
tmp13 = tmp11 + tmp12
tmp14 = triton_helpers.maximum(tmp13, tmp5)
tmp15 = tl.broadcast_to(tmp14, [RBLOCK])
tmp17 = triton_helpers.promote_to_tensor(tl.sum(tmp15, 0))
tmp18 = tmp9 + tmp17
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp18, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1, arg3_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg3_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [sub, add, clamp, cost_sent, sub_1, add_1, clamp_1, cost_img, loss], Original ATen: [aten.rsub, aten.add, aten.clamp, aten.sum]
stream0 = get_raw_stream(0)
triton_per_fused_add_clamp_rsub_sum_0.run(buf2, arg0_1, arg1_1, arg2_1, arg3_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
del arg3_1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg3_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1, arg3_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class PairwiseRankingLoss(nn.Module):
"""
Pairwise ranking loss
"""
def __init__(self, margin):
super(PairwiseRankingLoss, self).__init__()
self.margin = margin
def forward(self, anchor1, anchor2, img_sentc, sent_imgc):
cost_sent = torch.clamp(self.margin - anchor1 + img_sentc, min=0.0
).sum()
cost_img = torch.clamp(self.margin - anchor2 + sent_imgc, min=0.0).sum(
)
loss = cost_sent + cost_img
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'margin': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_add_clamp_rsub_sum_0(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, in_ptr3, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp3 = tl.load(in_ptr1 + r0, None)
tmp10 = tl.load(in_ptr2 + r0, None)
tmp12 = tl.load(in_ptr3 + r0, None)
tmp1 = 4.0
tmp2 = tmp1 - tmp0
tmp4 = tmp2 + tmp3
tmp5 = 0.0
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp7 = tl.broadcast_to(tmp6, [RBLOCK])
tmp9 = triton_helpers.promote_to_tensor(tl.sum(tmp7, 0))
tmp11 = tmp1 - tmp10
tmp13 = tmp11 + tmp12
tmp14 = triton_helpers.maximum(tmp13, tmp5)
tmp15 = tl.broadcast_to(tmp14, [RBLOCK])
tmp17 = triton_helpers.promote_to_tensor(tl.sum(tmp15, 0))
tmp18 = tmp9 + tmp17
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp18, None)
def call(args):
arg0_1, arg1_1, arg2_1, arg3_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg3_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_add_clamp_rsub_sum_0[grid(1)](buf2, arg0_1, arg1_1,
arg2_1, arg3_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
del arg3_1
return buf2,
class PairwiseRankingLossNew(nn.Module):
"""
Pairwise ranking loss
"""
def __init__(self, margin):
super(PairwiseRankingLossNew, self).__init__()
self.margin = margin
def forward(self, input_0, input_1, input_2, input_3):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
arg3_1 = input_3
output = call([arg0_1, arg1_1, arg2_1, arg3_1])
return output[0]
|
maksimovVva/SentEval
|
PairwiseRankingLoss
| false | 10,457 |
[
"BSD-3-Clause"
] | 0 |
d3aa5f24dd84b48ea476e73f4b59a4e1ace7775c
|
https://github.com/maksimovVva/SentEval/tree/d3aa5f24dd84b48ea476e73f4b59a4e1ace7775c
|
DiscreteNet
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/y7/cy7463cf27qpra2f6ndigmu6ve4q6o3cbvgetgqehejtevv6yfa5.py
# Topologically Sorted Source Nodes: [pi1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# pi1 => clamp_max, clamp_min
# Graph fragment:
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_1, 0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 6), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_1, 0), kwargs = {})
# %ge_1 : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%view_1, 6), kwargs = {})
# %bitwise_or_1 : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le_1, %ge_1), kwargs = {})
triton_poi_fused_hardtanh_hardtanh_backward_0 = async_compile.triton('triton_poi_fused_hardtanh_hardtanh_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_hardtanh_hardtanh_backward_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3h/c3hkfz2cekpaaeiqcpxcl3mxsxlmofvggknydsueowa7yt644s4y.py
# Topologically Sorted Source Nodes: [v1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# v1 => clamp_max_1, clamp_min_1
# Graph fragment:
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_5, 0), kwargs = {})
# %clamp_max_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min_1, 6), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_5, 0), kwargs = {})
# %ge : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%view_5, 6), kwargs = {})
# %bitwise_or : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le, %ge), kwargs = {})
triton_poi_fused_hardtanh_hardtanh_backward_1 = async_compile.triton('triton_poi_fused_hardtanh_hardtanh_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_hardtanh_hardtanh_backward_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_1(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x3 = (xindex // 1600)
x5 = xindex % 1600
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x4), tmp6, xmask)
tl.store(out_ptr1 + (x5 + (1664*x3)), tmp9, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (200, 4), (4, 1))
assert_size_stride(primals_2, (200, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 200), (200, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (100, 4), (4, 1))
assert_size_stride(primals_7, (100, ), (1, ))
assert_size_stride(primals_8, (1, 100), (100, 1))
assert_size_stride(primals_9, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 200), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1), torch.float32)
buf8 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1), torch.bool)
# Topologically Sorted Source Nodes: [pi1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_hardtanh_hardtanh_backward_0.run(buf0, primals_2, buf1, buf8, 12800, grid=grid(12800), stream=stream0)
del buf0
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [logits], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 200), (200, 1), 0), reinterpret_tensor(primals_4, (200, 4), (1, 200), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 100), (1, 4), 0), out=buf3)
del primals_6
buf4 = empty_strided_cuda((4, 4, 4, 100), (1600, 400, 100, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1), torch.bool)
# Topologically Sorted Source Nodes: [v1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
triton_poi_fused_hardtanh_hardtanh_backward_1.run(buf3, primals_7, buf4, buf7, 6400, grid=grid(6400), stream=stream0)
del buf3
del primals_7
buf6 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [values], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf4, (64, 100), (100, 1), 0), reinterpret_tensor(primals_8, (100, 1), (1, 100), 0), alpha=1, beta=1, out=buf6)
del primals_9
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf6, (4, 4, 4, 1), (16, 4, 1, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 200), (200, 1), 0), reinterpret_tensor(buf4, (64, 100), (100, 1), 0), primals_8, buf7, primals_4, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((200, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((200, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 200), (200, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((100, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((100, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 100), (100, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class DiscreteNet(nn.Module):
def __init__(self, s_dim, a_dim):
super(DiscreteNet, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.pi1 = nn.Linear(s_dim, 200)
self.pi2 = nn.Linear(200, a_dim)
self.v1 = nn.Linear(s_dim, 100)
self.v2 = nn.Linear(100, 1)
set_init([self.pi1, self.pi2, self.v1, self.v2])
self.distribution = torch.distributions.Categorical
def forward(self, x):
pi1 = F.relu6(self.pi1(x))
logits = self.pi2(pi1)
v1 = F.relu6(self.v1(x))
values = self.v2(v1)
return logits, values
def choose_action(self, s):
self.eval()
logits, _ = self.forward(s)
prob = F.softmax(logits, dim=1).data
m = self.distribution(prob)
return m.sample().numpy()[0]
def loss_func(self, s, a, v_t):
self.train()
logits, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
probs = F.softmax(logits, dim=1)
m = self.distribution(probs)
exp_v = m.log_prob(a) * td.detach().squeeze()
a_loss = -exp_v
total_loss = (c_loss + a_loss).mean()
return total_loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'s_dim': 4, 'a_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_1(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x3 = xindex // 1600
x5 = xindex % 1600
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x4, tmp6, xmask)
tl.store(out_ptr1 + (x5 + 1664 * x3), tmp9, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (200, 4), (4, 1))
assert_size_stride(primals_2, (200,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 200), (200, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (100, 4), (4, 1))
assert_size_stride(primals_7, (100,), (1,))
assert_size_stride(primals_8, (1, 100), (100, 1))
assert_size_stride(primals_9, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 200), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1),
torch.float32)
buf8 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_hardtanh_hardtanh_backward_0[grid(12800)](buf0,
primals_2, buf1, buf8, 12800, XBLOCK=256, num_warps=4, num_stages=1
)
del buf0
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 200),
(200, 1), 0), reinterpret_tensor(primals_4, (200, 4), (1, 200),
0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 100), (1, 4), 0), out=buf3)
del primals_6
buf4 = empty_strided_cuda((4, 4, 4, 100), (1600, 400, 100, 1),
torch.float32)
buf7 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1),
torch.bool)
triton_poi_fused_hardtanh_hardtanh_backward_1[grid(6400)](buf3,
primals_7, buf4, buf7, 6400, XBLOCK=256, num_warps=4, num_stages=1)
del buf3
del primals_7
buf6 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_9, reinterpret_tensor(buf4, (64, 100),
(100, 1), 0), reinterpret_tensor(primals_8, (100, 1), (1, 100),
0), alpha=1, beta=1, out=buf6)
del primals_9
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(buf6, (4, 4, 4, 1), (16, 4, 1, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 200), (200, 1), 0
), reinterpret_tensor(buf4, (64, 100), (100, 1), 0
), primals_8, buf7, primals_4, buf8
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class DiscreteNetNew(nn.Module):
def __init__(self, s_dim, a_dim):
super(DiscreteNetNew, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.pi1 = nn.Linear(s_dim, 200)
self.pi2 = nn.Linear(200, a_dim)
self.v1 = nn.Linear(s_dim, 100)
self.v2 = nn.Linear(100, 1)
set_init([self.pi1, self.pi2, self.v1, self.v2])
self.distribution = torch.distributions.Categorical
def choose_action(self, s):
self.eval()
logits, _ = self.forward(s)
prob = F.softmax(logits, dim=1).data
m = self.distribution(prob)
return m.sample().numpy()[0]
def loss_func(self, s, a, v_t):
self.train()
logits, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
probs = F.softmax(logits, dim=1)
m = self.distribution(probs)
exp_v = m.log_prob(a) * td.detach().squeeze()
a_loss = -exp_v
total_loss = (c_loss + a_loss).mean()
return total_loss
def forward(self, input_0):
primals_1 = self.pi1.weight
primals_2 = self.pi1.bias
primals_4 = self.pi2.weight
primals_5 = self.pi2.bias
primals_6 = self.v1.weight
primals_7 = self.v1.bias
primals_8 = self.v2.weight
primals_9 = self.v2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0], output[1]
|
lws803/pytorch-A3C
|
DiscreteNet
| false | 10,458 |
[
"MIT"
] | 0 |
944e7f42a8fa54b7d6efbe169d8a3467b20a0f7f
|
https://github.com/lws803/pytorch-A3C/tree/944e7f42a8fa54b7d6efbe169d8a3467b20a0f7f
|
NoiseLayer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/hy/chy2gdvcdlpxio6r2ezu76sbqj6jaum4snkq6izi5zxkdqh3u2yj.py
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.add]
# Source node to ATen node mapping:
# mul => mul
# x => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %randn), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_1, %mul), kwargs = {})
triton_poi_fused_add_mul_0 = async_compile.triton('triton_poi_fused_add_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = tmp0 + tmp3
tl.store(out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [noise], Original ATen: [aten.randn]
buf0 = torch.ops.aten.randn.default([4, 1, 4, 4], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_0.run(primals_1, primals_2, buf1, buf2, 256, grid=grid(256), stream=stream0)
del primals_1
del primals_2
return (buf2, buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class NoiseLayer(nn.Module):
"""adds noise. noise is per pixel (constant over channels) with per-channel weight"""
def __init__(self, channels):
super().__init__()
self.weight = nn.Parameter(torch.zeros(channels))
self.noise = None
def forward(self, x, noise=None):
if noise is None and self.noise is None:
noise = torch.randn(x.size(0), 1, x.size(2), x.size(3), device=
x.device, dtype=x.dtype)
elif noise is None:
noise = self.noise
x = x + self.weight.view(1, -1, 1, 1) * noise
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'channels': 4}]
|
import torch
from torch import device
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_mul_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 * tmp2
tmp4 = tmp0 + tmp3
tl.store(out_ptr0 + x3, tmp4, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = torch.ops.aten.randn.default([4, 1, 4, 4], dtype=torch.
float32, device=device(type='cuda', index=0), pin_memory=False)
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_0[grid(256)](primals_1, primals_2, buf1,
buf2, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_2
return buf2, buf1
class NoiseLayerNew(nn.Module):
"""adds noise. noise is per pixel (constant over channels) with per-channel weight"""
def __init__(self, channels):
super().__init__()
self.weight = nn.Parameter(torch.zeros(channels))
self.noise = None
def forward(self, input_0):
primals_2 = self.weight
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
|
justinpinkney/ganspace
|
NoiseLayer
| false | 10,459 |
[
"Apache-2.0"
] | 0 |
7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
https://github.com/justinpinkney/ganspace/tree/7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
mlp
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/cv/ccvagufpv2m2vfo3w35yjibzs4ygygyoilwi2jrpbsvlftqb6nzr.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_3), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 2048
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (2048, 4), (4, 1))
assert_size_stride(primals_3, (2048, ), (1, ))
assert_size_stride(primals_4, (2048, 2048), (2048, 1))
assert_size_stride(primals_5, (2048, ), (1, ))
assert_size_stride(primals_6, (4, 2048), (2048, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 2048), (2048, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (4, 2048), (1, 4), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 8192, grid=grid(8192), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 2048), (2048, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (2048, 2048), (1, 2048), 0), out=buf2)
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [relu_1], Original ATen: [aten.relu]
triton_poi_fused_relu_0.run(buf3, primals_5, 8192, grid=grid(8192), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf3, reinterpret_tensor(primals_6, (2048, 4), (1, 2048), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (buf4, primals_1, buf1, buf3, primals_6, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((2048, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((2048, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((2048, 2048), (2048, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((2048, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 2048), (2048, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class mlp(nn.Module):
def __init__(self, seq_len):
super(mlp, self).__init__()
self.lin1 = nn.Linear(seq_len, 2048)
self.lin2 = nn.Linear(2048, 2048)
self.lin3 = nn.Linear(2048, seq_len)
self.relu = nn.ReLU()
def forward(self, input_):
input_ = input_.reshape(input_.size(0), -1)
out = self.lin1(input_)
out = self.lin2(self.relu(out))
out = self.lin3(self.relu(out))
return out.view(input_.size(0), -1)
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'seq_len': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 2048
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (2048, 4), (4, 1))
assert_size_stride(primals_3, (2048,), (1,))
assert_size_stride(primals_4, (2048, 2048), (2048, 1))
assert_size_stride(primals_5, (2048,), (1,))
assert_size_stride(primals_6, (4, 2048), (2048, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 2048), (2048, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (4, 2048
), (1, 4), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(8192)](buf1, primals_3, 8192, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 2048), (2048, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (2048, 2048),
(1, 2048), 0), out=buf2)
buf3 = buf2
del buf2
triton_poi_fused_relu_0[grid(8192)](buf3, primals_5, 8192, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, buf3, reinterpret_tensor(primals_6,
(2048, 4), (1, 2048), 0), alpha=1, beta=1, out=buf4)
del primals_7
return buf4, primals_1, buf1, buf3, primals_6, primals_4
class mlpNew(nn.Module):
def __init__(self, seq_len):
super(mlpNew, self).__init__()
self.lin1 = nn.Linear(seq_len, 2048)
self.lin2 = nn.Linear(2048, 2048)
self.lin3 = nn.Linear(2048, seq_len)
self.relu = nn.ReLU()
def forward(self, input_0):
primals_2 = self.lin1.weight
primals_3 = self.lin1.bias
primals_4 = self.lin2.weight
primals_5 = self.lin2.bias
primals_6 = self.lin3.weight
primals_7 = self.lin3.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
liuziyang1106/sodeep
|
mlp
| false | 10,460 |
[
"BSD-3-Clause-Clear"
] | 0 |
47f8a5cbe5b8405624877efc81cb28f104f1e2d7
|
https://github.com/liuziyang1106/sodeep/tree/47f8a5cbe5b8405624877efc81cb28f104f1e2d7
|
GetSegPred
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ws/cwsbiv2j3g4xvieopa3bkoe5ikyusgpeusghiciceorbcngdbj6v.py
# Topologically Sorted Source Nodes: [add, mul, sub, round_1, temp_cloud, setitem], Original ATen: [aten.add, aten.mul, aten.sub, aten.round, aten._to_copy, aten.lift_fresh, aten.index_put]
# Source node to ATen node mapping:
# add => add
# mul => mul
# round_1 => round_1
# setitem => full_default, index_put
# sub => sub
# temp_cloud => convert_element_type
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, 1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add, 0.0), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, 0.501), kwargs = {})
# %round_1 : [num_users=1] = call_function[target=torch.ops.aten.round.default](args = (%sub,), kwargs = {})
# %convert_element_type : [num_users=2] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%round_1, torch.int64), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put : [num_users=12] = call_function[target=torch.ops.aten.index_put_.default](args = (%convert_element_type, [%eq], %full_default), kwargs = {})
triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0 = async_compile.triton('triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 * tmp3
tmp5 = 0.501
tmp6 = tmp4 - tmp5
tmp7 = libdevice.nearbyint(tmp6)
tmp8 = tmp7.to(tl.int64)
tmp9 = tl.full([1], -1, tl.int64)
tmp10 = tmp8 == tmp9
tmp11 = tl.full([1], 0, tl.int64)
tmp12 = tl.where(tmp10, tmp11, tmp8)
tl.store(out_ptr0 + (x0), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wy/cwy5lnusqcz7skukmacggq3zuqyx7pqh6vt43oihdsykv4kmxgg4.py
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# cat => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%unsqueeze, %unsqueeze_1, %unsqueeze_2, %unsqueeze_3],), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = (xindex // 64)
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x3
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x1 + (16*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.full([XBLOCK], 4, tl.int32)
tmp7 = tmp5 + tmp6
tmp8 = tmp5 < 0
tmp9 = tl.where(tmp8, tmp7, tmp5)
tl.device_assert(((0 <= tl.broadcast_to(tmp9, [XBLOCK])) & (tl.broadcast_to(tmp9, [XBLOCK]) < 4)) | ~(tmp4 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp9, [XBLOCK]) < 4")
tmp11 = tl.load(in_ptr0 + (4 + x1 + (16*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp12 = tmp11 + tmp6
tmp13 = tmp11 < 0
tmp14 = tl.where(tmp13, tmp12, tmp11)
tl.device_assert(((0 <= tl.broadcast_to(tmp14, [XBLOCK])) & (tl.broadcast_to(tmp14, [XBLOCK]) < 4)) | ~(tmp4 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp14, [XBLOCK]) < 4")
tmp16 = tl.load(in_ptr0 + (8 + x1 + (16*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp17 = tmp16 + tmp6
tmp18 = tmp16 < 0
tmp19 = tl.where(tmp18, tmp17, tmp16)
tl.device_assert(((0 <= tl.broadcast_to(tmp19, [XBLOCK])) & (tl.broadcast_to(tmp19, [XBLOCK]) < 4)) | ~(tmp4 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp19, [XBLOCK]) < 4")
tmp21 = tl.load(in_ptr1 + (tmp9 + (4*tmp19) + (16*tmp14) + (64*x0)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp22 = tmp0 >= tmp3
tmp23 = tl.full([1], 2, tl.int64)
tmp24 = tmp0 < tmp23
tmp25 = tmp22 & tmp24
tmp26 = tl.load(in_ptr0 + (64 + x1 + (16*x2)), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp27 = tmp26 + tmp6
tmp28 = tmp26 < 0
tmp29 = tl.where(tmp28, tmp27, tmp26)
tl.device_assert(((0 <= tl.broadcast_to(tmp29, [XBLOCK])) & (tl.broadcast_to(tmp29, [XBLOCK]) < 4)) | ~(tmp25 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp29, [XBLOCK]) < 4")
tmp31 = tl.load(in_ptr0 + (68 + x1 + (16*x2)), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp32 = tmp31 + tmp6
tmp33 = tmp31 < 0
tmp34 = tl.where(tmp33, tmp32, tmp31)
tl.device_assert(((0 <= tl.broadcast_to(tmp34, [XBLOCK])) & (tl.broadcast_to(tmp34, [XBLOCK]) < 4)) | ~(tmp25 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp34, [XBLOCK]) < 4")
tmp36 = tl.load(in_ptr0 + (72 + x1 + (16*x2)), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp37 = tmp36 + tmp6
tmp38 = tmp36 < 0
tmp39 = tl.where(tmp38, tmp37, tmp36)
tl.device_assert(((0 <= tl.broadcast_to(tmp39, [XBLOCK])) & (tl.broadcast_to(tmp39, [XBLOCK]) < 4)) | ~(tmp25 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp39, [XBLOCK]) < 4")
tmp41 = tl.load(in_ptr1 + (256 + tmp29 + (4*tmp39) + (16*tmp34) + (64*x0)), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp42 = tmp0 >= tmp23
tmp43 = tl.full([1], 3, tl.int64)
tmp44 = tmp0 < tmp43
tmp45 = tmp42 & tmp44
tmp46 = tl.load(in_ptr0 + (128 + x1 + (16*x2)), tmp45 & xmask, eviction_policy='evict_last', other=0.0)
tmp47 = tmp46 + tmp6
tmp48 = tmp46 < 0
tmp49 = tl.where(tmp48, tmp47, tmp46)
tl.device_assert(((0 <= tl.broadcast_to(tmp49, [XBLOCK])) & (tl.broadcast_to(tmp49, [XBLOCK]) < 4)) | ~(tmp45 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp49, [XBLOCK]) < 4")
tmp51 = tl.load(in_ptr0 + (132 + x1 + (16*x2)), tmp45 & xmask, eviction_policy='evict_last', other=0.0)
tmp52 = tmp51 + tmp6
tmp53 = tmp51 < 0
tmp54 = tl.where(tmp53, tmp52, tmp51)
tl.device_assert(((0 <= tl.broadcast_to(tmp54, [XBLOCK])) & (tl.broadcast_to(tmp54, [XBLOCK]) < 4)) | ~(tmp45 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp54, [XBLOCK]) < 4")
tmp56 = tl.load(in_ptr0 + (136 + x1 + (16*x2)), tmp45 & xmask, eviction_policy='evict_last', other=0.0)
tmp57 = tmp56 + tmp6
tmp58 = tmp56 < 0
tmp59 = tl.where(tmp58, tmp57, tmp56)
tl.device_assert(((0 <= tl.broadcast_to(tmp59, [XBLOCK])) & (tl.broadcast_to(tmp59, [XBLOCK]) < 4)) | ~(tmp45 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp59, [XBLOCK]) < 4")
tmp61 = tl.load(in_ptr1 + (512 + tmp49 + (4*tmp59) + (16*tmp54) + (64*x0)), tmp45 & xmask, eviction_policy='evict_last', other=0.0)
tmp62 = tmp0 >= tmp43
tmp63 = tl.full([1], 4, tl.int64)
tmp64 = tmp0 < tmp63
tmp65 = tl.load(in_ptr0 + (192 + x1 + (16*x2)), tmp62 & xmask, eviction_policy='evict_last', other=0.0)
tmp66 = tmp65 + tmp6
tmp67 = tmp65 < 0
tmp68 = tl.where(tmp67, tmp66, tmp65)
tl.device_assert(((0 <= tl.broadcast_to(tmp68, [XBLOCK])) & (tl.broadcast_to(tmp68, [XBLOCK]) < 4)) | ~(tmp62 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp68, [XBLOCK]) < 4")
tmp70 = tl.load(in_ptr0 + (196 + x1 + (16*x2)), tmp62 & xmask, eviction_policy='evict_last', other=0.0)
tmp71 = tmp70 + tmp6
tmp72 = tmp70 < 0
tmp73 = tl.where(tmp72, tmp71, tmp70)
tl.device_assert(((0 <= tl.broadcast_to(tmp73, [XBLOCK])) & (tl.broadcast_to(tmp73, [XBLOCK]) < 4)) | ~(tmp62 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp73, [XBLOCK]) < 4")
tmp75 = tl.load(in_ptr0 + (200 + x1 + (16*x2)), tmp62 & xmask, eviction_policy='evict_last', other=0.0)
tmp76 = tmp75 + tmp6
tmp77 = tmp75 < 0
tmp78 = tl.where(tmp77, tmp76, tmp75)
tl.device_assert(((0 <= tl.broadcast_to(tmp78, [XBLOCK])) & (tl.broadcast_to(tmp78, [XBLOCK]) < 4)) | ~(tmp62 & xmask), "index out of bounds: 0 <= tl.broadcast_to(tmp78, [XBLOCK]) < 4")
tmp80 = tl.load(in_ptr1 + (768 + tmp68 + (4*tmp78) + (16*tmp73) + (64*x0)), tmp62 & xmask, eviction_policy='evict_last', other=0.0)
tmp81 = tl.where(tmp45, tmp61, tmp80)
tmp82 = tl.where(tmp25, tmp41, tmp81)
tmp83 = tl.where(tmp4, tmp21, tmp82)
tl.store(out_ptr0 + (x5), tmp83, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.int64)
# Topologically Sorted Source Nodes: [add, mul, sub, round_1, temp_cloud, setitem], Original ATen: [aten.add, aten.mul, aten.sub, aten.round, aten._to_copy, aten.lift_fresh, aten.index_put]
stream0 = get_raw_stream(0)
triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf0, arg1_1, buf1, 256, grid=grid(256), stream=stream0)
del arg1_1
del buf0
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.utils.data.dataset
class GetSegPred(torch.nn.Module):
def __init__(self, scale):
super(GetSegPred, self).__init__()
self.scale = scale // 2
def forward(self, segs, ptcloud):
temp_cloud = torch.round((ptcloud + 1) * self.scale - 0.501).long()
temp_cloud[temp_cloud == -1] = 0
segsT = torch.transpose(segs, 1, 4)
preds = []
for i, p in enumerate(temp_cloud):
pred = segsT[i, p[:, 0], p[:, 1], p[:, 2]].unsqueeze(dim=0)
preds.append(pred)
return torch.cat(preds, dim=0).contiguous()
def get_inputs():
return [torch.rand([4, 4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'scale': 1.0}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.utils.data.dataset
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0(in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 * tmp3
tmp5 = 0.501
tmp6 = tmp4 - tmp5
tmp7 = libdevice.nearbyint(tmp6)
tmp8 = tmp7.to(tl.int64)
tmp9 = tl.full([1], -1, tl.int64)
tmp10 = tmp8 == tmp9
tmp11 = tl.full([1], 0, tl.int64)
tmp12 = tl.where(tmp10, tmp11, tmp8)
tl.store(out_ptr0 + x0, tmp12, xmask)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex // 64
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x3
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x1 + 16 * x2), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.full([XBLOCK], 4, tl.int32)
tmp7 = tmp5 + tmp6
tmp8 = tmp5 < 0
tmp9 = tl.where(tmp8, tmp7, tmp5)
tl.device_assert((0 <= tl.broadcast_to(tmp9, [XBLOCK])) & (tl.
broadcast_to(tmp9, [XBLOCK]) < 4) | ~(tmp4 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp9, [XBLOCK]) < 4')
tmp11 = tl.load(in_ptr0 + (4 + x1 + 16 * x2), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp12 = tmp11 + tmp6
tmp13 = tmp11 < 0
tmp14 = tl.where(tmp13, tmp12, tmp11)
tl.device_assert((0 <= tl.broadcast_to(tmp14, [XBLOCK])) & (tl.
broadcast_to(tmp14, [XBLOCK]) < 4) | ~(tmp4 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp14, [XBLOCK]) < 4')
tmp16 = tl.load(in_ptr0 + (8 + x1 + 16 * x2), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp17 = tmp16 + tmp6
tmp18 = tmp16 < 0
tmp19 = tl.where(tmp18, tmp17, tmp16)
tl.device_assert((0 <= tl.broadcast_to(tmp19, [XBLOCK])) & (tl.
broadcast_to(tmp19, [XBLOCK]) < 4) | ~(tmp4 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp19, [XBLOCK]) < 4')
tmp21 = tl.load(in_ptr1 + (tmp9 + 4 * tmp19 + 16 * tmp14 + 64 * x0),
tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp22 = tmp0 >= tmp3
tmp23 = tl.full([1], 2, tl.int64)
tmp24 = tmp0 < tmp23
tmp25 = tmp22 & tmp24
tmp26 = tl.load(in_ptr0 + (64 + x1 + 16 * x2), tmp25 & xmask,
eviction_policy='evict_last', other=0.0)
tmp27 = tmp26 + tmp6
tmp28 = tmp26 < 0
tmp29 = tl.where(tmp28, tmp27, tmp26)
tl.device_assert((0 <= tl.broadcast_to(tmp29, [XBLOCK])) & (tl.
broadcast_to(tmp29, [XBLOCK]) < 4) | ~(tmp25 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp29, [XBLOCK]) < 4')
tmp31 = tl.load(in_ptr0 + (68 + x1 + 16 * x2), tmp25 & xmask,
eviction_policy='evict_last', other=0.0)
tmp32 = tmp31 + tmp6
tmp33 = tmp31 < 0
tmp34 = tl.where(tmp33, tmp32, tmp31)
tl.device_assert((0 <= tl.broadcast_to(tmp34, [XBLOCK])) & (tl.
broadcast_to(tmp34, [XBLOCK]) < 4) | ~(tmp25 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp34, [XBLOCK]) < 4')
tmp36 = tl.load(in_ptr0 + (72 + x1 + 16 * x2), tmp25 & xmask,
eviction_policy='evict_last', other=0.0)
tmp37 = tmp36 + tmp6
tmp38 = tmp36 < 0
tmp39 = tl.where(tmp38, tmp37, tmp36)
tl.device_assert((0 <= tl.broadcast_to(tmp39, [XBLOCK])) & (tl.
broadcast_to(tmp39, [XBLOCK]) < 4) | ~(tmp25 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp39, [XBLOCK]) < 4')
tmp41 = tl.load(in_ptr1 + (256 + tmp29 + 4 * tmp39 + 16 * tmp34 + 64 *
x0), tmp25 & xmask, eviction_policy='evict_last', other=0.0)
tmp42 = tmp0 >= tmp23
tmp43 = tl.full([1], 3, tl.int64)
tmp44 = tmp0 < tmp43
tmp45 = tmp42 & tmp44
tmp46 = tl.load(in_ptr0 + (128 + x1 + 16 * x2), tmp45 & xmask,
eviction_policy='evict_last', other=0.0)
tmp47 = tmp46 + tmp6
tmp48 = tmp46 < 0
tmp49 = tl.where(tmp48, tmp47, tmp46)
tl.device_assert((0 <= tl.broadcast_to(tmp49, [XBLOCK])) & (tl.
broadcast_to(tmp49, [XBLOCK]) < 4) | ~(tmp45 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp49, [XBLOCK]) < 4')
tmp51 = tl.load(in_ptr0 + (132 + x1 + 16 * x2), tmp45 & xmask,
eviction_policy='evict_last', other=0.0)
tmp52 = tmp51 + tmp6
tmp53 = tmp51 < 0
tmp54 = tl.where(tmp53, tmp52, tmp51)
tl.device_assert((0 <= tl.broadcast_to(tmp54, [XBLOCK])) & (tl.
broadcast_to(tmp54, [XBLOCK]) < 4) | ~(tmp45 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp54, [XBLOCK]) < 4')
tmp56 = tl.load(in_ptr0 + (136 + x1 + 16 * x2), tmp45 & xmask,
eviction_policy='evict_last', other=0.0)
tmp57 = tmp56 + tmp6
tmp58 = tmp56 < 0
tmp59 = tl.where(tmp58, tmp57, tmp56)
tl.device_assert((0 <= tl.broadcast_to(tmp59, [XBLOCK])) & (tl.
broadcast_to(tmp59, [XBLOCK]) < 4) | ~(tmp45 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp59, [XBLOCK]) < 4')
tmp61 = tl.load(in_ptr1 + (512 + tmp49 + 4 * tmp59 + 16 * tmp54 + 64 *
x0), tmp45 & xmask, eviction_policy='evict_last', other=0.0)
tmp62 = tmp0 >= tmp43
tl.full([1], 4, tl.int64)
tmp65 = tl.load(in_ptr0 + (192 + x1 + 16 * x2), tmp62 & xmask,
eviction_policy='evict_last', other=0.0)
tmp66 = tmp65 + tmp6
tmp67 = tmp65 < 0
tmp68 = tl.where(tmp67, tmp66, tmp65)
tl.device_assert((0 <= tl.broadcast_to(tmp68, [XBLOCK])) & (tl.
broadcast_to(tmp68, [XBLOCK]) < 4) | ~(tmp62 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp68, [XBLOCK]) < 4')
tmp70 = tl.load(in_ptr0 + (196 + x1 + 16 * x2), tmp62 & xmask,
eviction_policy='evict_last', other=0.0)
tmp71 = tmp70 + tmp6
tmp72 = tmp70 < 0
tmp73 = tl.where(tmp72, tmp71, tmp70)
tl.device_assert((0 <= tl.broadcast_to(tmp73, [XBLOCK])) & (tl.
broadcast_to(tmp73, [XBLOCK]) < 4) | ~(tmp62 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp73, [XBLOCK]) < 4')
tmp75 = tl.load(in_ptr0 + (200 + x1 + 16 * x2), tmp62 & xmask,
eviction_policy='evict_last', other=0.0)
tmp76 = tmp75 + tmp6
tmp77 = tmp75 < 0
tmp78 = tl.where(tmp77, tmp76, tmp75)
tl.device_assert((0 <= tl.broadcast_to(tmp78, [XBLOCK])) & (tl.
broadcast_to(tmp78, [XBLOCK]) < 4) | ~(tmp62 & xmask),
'index out of bounds: 0 <= tl.broadcast_to(tmp78, [XBLOCK]) < 4')
tmp80 = tl.load(in_ptr1 + (768 + tmp68 + 4 * tmp78 + 16 * tmp73 + 64 *
x0), tmp62 & xmask, eviction_policy='evict_last', other=0.0)
tmp81 = tl.where(tmp45, tmp61, tmp80)
tmp82 = tl.where(tmp25, tmp41, tmp81)
tmp83 = tl.where(tmp4, tmp21, tmp82)
tl.store(out_ptr0 + x5, tmp83, xmask)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.int64)
get_raw_stream(0)
triton_poi_fused__to_copy_add_index_put_lift_fresh_mul_round_sub_0[grid
(256)](arg0_1, buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_cat_1[grid(256)](buf0, arg1_1, buf1, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg1_1
del buf0
return buf1,
class GetSegPredNew(torch.nn.Module):
def __init__(self, scale):
super(GetSegPredNew, self).__init__()
self.scale = scale // 2
def forward(self, input_0, input_1):
arg1_1 = input_0
arg0_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
melisataspinar/Concurrent-Completion-and-Part-Segmentation-for-3D-Missing-Point-Clouds-viaSynergistic-Feature-Mappi
|
GetSegPred
| false | 10,461 |
[
"MIT"
] | 0 |
3b03f3c167d9927a660d798ffcd8ecc0f5cbaf89
|
https://github.com/melisataspinar/Concurrent-Completion-and-Part-Segmentation-for-3D-Missing-Point-Clouds-viaSynergistic-Feature-Mappi/tree/3b03f3c167d9927a660d798ffcd8ecc0f5cbaf89
|
StyleMod
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yg/cygzi6evso6kefobgrwgjcxh5qgbn7zouyf7to742xeh6bsflplb.py
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, 0.5), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2n/c2npkiljqvyo5boces3kqpxvha7qfatrulgz7qm2oitxblzqbabd.py
# Topologically Sorted Source Nodes: [add, mul_2, x], Original ATen: [aten.add, aten.mul]
# Source node to ATen node mapping:
# add => add
# mul_2 => mul_2
# x => add_1
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%select, 1.0), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_4, %add), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %select_1), kwargs = {})
triton_poi_fused_add_mul_1 = async_compile.triton('triton_poi_fused_add_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4096
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 16) % 4
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr1 + (x1 + (8*x2)), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x1), None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (4 + x1 + (8*x2)), None, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr2 + (4 + x1), None, eviction_policy='evict_last')
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp5 = tmp1 + tmp4
tmp6 = tmp5 + tmp3
tmp7 = tmp0 * tmp6
tmp10 = tmp9 * tmp3
tmp11 = tmp8 + tmp10
tmp12 = tmp7 + tmp11
tl.store(out_ptr0 + (x3), tmp12, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (8, ), (1, ))
assert_size_stride(primals_2, (8, 4), (4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (64, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((8, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_2, buf0, 32, grid=grid(32), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((64, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf0, (4, 8), (1, 4), 0), out=buf1)
del buf0
buf2 = empty_strided_cuda((64, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, mul_2, x], Original ATen: [aten.add, aten.mul]
triton_poi_fused_add_mul_1.run(primals_4, buf1, primals_1, buf2, 4096, grid=grid(4096), stream=stream0)
del buf1
del primals_1
return (buf2, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((8, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((8, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((64, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class MyLinear(nn.Module):
"""Linear layer with equalized learning rate and custom learning rate multiplier."""
def __init__(self, input_size, output_size, gain=2 ** 0.5, use_wscale=
False, lrmul=1, bias=True):
super().__init__()
he_std = gain * input_size ** -0.5
if use_wscale:
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
else:
init_std = he_std / lrmul
self.w_mul = lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
if bias:
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
else:
self.bias = None
def forward(self, x):
bias = self.bias
if bias is not None:
bias = bias * self.b_mul
return F.linear(x, self.weight * self.w_mul, bias)
class StyleMod(nn.Module):
def __init__(self, latent_size, channels, use_wscale):
super(StyleMod, self).__init__()
self.lin = MyLinear(latent_size, channels * 2, gain=1.0, use_wscale
=use_wscale)
def forward(self, x, latent):
style = self.lin(latent)
shape = [-1, 2, x.size(1)] + (x.dim() - 2) * [1]
style = style.view(shape)
x = x * (style[:, 0] + 1.0) + style[:, 1]
return x
def get_inputs():
return [torch.rand([64, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'latent_size': 4, 'channels': 4, 'use_wscale': 1.0}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 16 % 4
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, None)
tmp1 = tl.load(in_ptr1 + (x1 + 8 * x2), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + x1, None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (4 + x1 + 8 * x2), None, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr2 + (4 + x1), None, eviction_policy='evict_last')
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp5 = tmp1 + tmp4
tmp6 = tmp5 + tmp3
tmp7 = tmp0 * tmp6
tmp10 = tmp9 * tmp3
tmp11 = tmp8 + tmp10
tmp12 = tmp7 + tmp11
tl.store(out_ptr0 + x3, tmp12, None)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (8,), (1,))
assert_size_stride(primals_2, (8, 4), (4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (64, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((8, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(32)](primals_2, buf0, 32, XBLOCK=32,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((64, 8), (8, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(buf0, (4, 8), (1, 4), 0), out=buf1)
del buf0
buf2 = empty_strided_cuda((64, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_1[grid(4096)](primals_4, buf1, primals_1,
buf2, 4096, XBLOCK=256, num_warps=4, num_stages=1)
del buf1
del primals_1
return buf2, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0)
class MyLinear(nn.Module):
"""Linear layer with equalized learning rate and custom learning rate multiplier."""
def __init__(self, input_size, output_size, gain=2 ** 0.5, use_wscale=
False, lrmul=1, bias=True):
super().__init__()
he_std = gain * input_size ** -0.5
if use_wscale:
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
else:
init_std = he_std / lrmul
self.w_mul = lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
if bias:
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
else:
self.bias = None
def forward(self, x):
bias = self.bias
if bias is not None:
bias = bias * self.b_mul
return F.linear(x, self.weight * self.w_mul, bias)
class StyleModNew(nn.Module):
def __init__(self, latent_size, channels, use_wscale):
super(StyleModNew, self).__init__()
self.lin = MyLinear(latent_size, channels * 2, gain=1.0, use_wscale
=use_wscale)
def forward(self, input_0, input_1):
primals_2 = self.lin.weight
primals_1 = self.lin.bias
primals_4 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
|
justinpinkney/ganspace
|
StyleMod
| false | 10,462 |
[
"Apache-2.0"
] | 0 |
7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
https://github.com/justinpinkney/ganspace/tree/7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
MultiHeadedAttention
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/x2/cx2hdvwyo7m5jvhhvtugzxqvmy6z4nsfhkkjhvgzbbm3cb6dsum2.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %mul_scalar : [num_users=1] = call_function[target=torch.ops.aten.mul.Scalar](args = (%permute_default, 1.0), kwargs = {})
# %clone_default : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_0 = async_compile.triton('triton_poi_fused_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + (4*y3)), tmp4, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/5j/c5jll3kxtd32cl7pwubrb5oky2mtzckfgip2xbwad7crvvp4zk4r.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_default_2, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_default_2, %amax_default), kwargs = {})
# %exp_default : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_tensor,), kwargs = {})
triton_poi_fused_1 = async_compile.triton('triton_poi_fused_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/kt/cktnex5febczl2ac6zugjmcksgsd5kjdufazv65vtepuwob3cb7a.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %sum_dim_int_list : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_default, [-1], True), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_default, %sum_dim_int_list), kwargs = {})
# %eq_scalar : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%view_default_2, -inf), kwargs = {})
# %logical_not_default : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%eq_scalar,), kwargs = {})
# %any_dim : [num_users=1] = call_function[target=torch.ops.aten.any.dim](args = (%logical_not_default, -1, True), kwargs = {})
# %logical_not_default_1 : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%any_dim,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_self : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%logical_not_default_1, %full_default, %div_tensor), kwargs = {})
triton_poi_fused_2 = async_compile.triton('triton_poi_fused_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr1 + (x2), xmask)
tmp26 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = float("-inf")
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = (tmp4 != 0)
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = (tmp9 != 0)
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = (tmp15 != 0)
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = (tmp21 != 0)
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + (x2), tmp35, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/vv/cvvnhithjvmvhfjufxwwzclfobkrgbyyteg66hp24r675f7elw4c.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %clone_default_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_3 = async_compile.triton('triton_poi_fused_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_3(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + (4*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/6t/c6t5a5ere3lqjiu7zh3uu4oxmpdoujdaqqmeunxqapgzo4m74uav.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone_4
# Graph fragment:
# %clone_4 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_4 = async_compile.triton('triton_poi_fused_clone_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_6, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_9, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf2)
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
stream0 = get_raw_stream(0)
triton_poi_fused_0.run(buf0, primals_3, buf3, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_3
buf4 = reinterpret_tensor(buf0, (4, 4, 1, 4), (16, 4, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_0.run(buf1, primals_5, buf4, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(buf5, buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf5
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_3.run(buf2, primals_8, buf8, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_8
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.bmm(reinterpret_tensor(buf7, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [linear_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf11)
del primals_11
return (reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (16, 4), (4, 1), 0), buf7, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0), reinterpret_tensor(buf10, (16, 4), (4, 1), 0), primals_10, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
from typing import Optional
import torch.nn.functional as F
from torch import nn
def attention(query, key, value, mask=None, dropout=None):
"""Compute 'Scaled Dot Product Attention'
"""
d_k = query.size(-1)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(d_k)
if mask is not None:
scores = scores.masked_fill(mask == 0, -1000000000.0)
weights = F.softmax(scores, dim=-1)
if dropout is not None:
weights = dropout(weights)
out = torch.matmul(weights, value)
return out, weights
class MultiHeadedAttention(nn.Module):
def __init__(self, h, d_model, dropout=0.1):
"""Take in model size and number of heads."""
super(MultiHeadedAttention, self).__init__()
assert d_model % h == 0
self.d_head = d_model // h
self.h = h
self.proj_q = nn.Linear(d_model, d_model)
self.proj_k = nn.Linear(d_model, d_model)
self.proj_v = nn.Linear(d_model, d_model)
self.ret_proj = nn.Linear(d_model, d_model)
self.dropout = nn.Dropout(p=dropout)
def forward(self, query: 'torch.Tensor', key: 'torch.Tensor', value:
'torch.Tensor', mask: 'Optional[torch.Tensor]'=None):
"""
query: (batch_size, seq_len, dmodel)
key: (batch_size, seq_len, dmodel)
value: (batch_size, seq_len, dmodel)
mask: (batch_size, seq_len)
"""
if mask is not None:
mask = mask.unsqueeze(1)
batch_size = query.size(0)
seq_len = query.size(1)
query = self.proj_q(query).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
key = self.proj_k(key).view(batch_size, seq_len, self.h, self.d_head
).transpose(1, 2)
value = self.proj_v(value).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
x, _ = attention(query, key, value, mask=mask, dropout=self.dropout)
x = x.transpose(1, 2).contiguous().view(batch_size, seq_len, self.h *
self.d_head)
return self.ret_proj(x)
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'h': 4, 'd_model': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import torch.nn.functional as F
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + 4 * y3), tmp4, xmask & ymask)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp25 = tl.load(in_ptr1 + x2, xmask)
tmp26 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp29 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp31 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = float('-inf')
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = tmp4 != 0
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = tmp9 != 0
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = tmp15 != 0
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp21 != 0
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + x2, tmp35, xmask)
@triton.jit
def triton_poi_fused_3(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + 4 * y3), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_6, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_9, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf2)
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_0[grid(16, 4)](buf0, primals_3, buf3, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_3
buf4 = reinterpret_tensor(buf0, (4, 4, 1, 4), (16, 4, 4, 1), 0)
del buf0
triton_poi_fused_0[grid(16, 4)](buf1, primals_5, buf4, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_1[grid(256)](buf5, buf6, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_2[grid(256)](buf5, buf6, buf7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf5
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_3[grid(16, 4)](buf2, primals_8, buf8, 16, 4,
XBLOCK=4, YBLOCK=8, num_warps=1, num_stages=1)
del primals_8
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0)
del buf2
extern_kernels.bmm(reinterpret_tensor(buf7, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf11)
del primals_11
return reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0
), reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_6, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_9, (16, 4), (4, 1), 0
), buf7, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0
), reinterpret_tensor(buf10, (16, 4), (4, 1), 0), primals_10
def attention(query, key, value, mask=None, dropout=None):
"""Compute 'Scaled Dot Product Attention'
"""
d_k = query.size(-1)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(d_k)
if mask is not None:
scores = scores.masked_fill(mask == 0, -1000000000.0)
weights = F.softmax(scores, dim=-1)
if dropout is not None:
weights = dropout(weights)
out = torch.matmul(weights, value)
return out, weights
class MultiHeadedAttentionNew(nn.Module):
def __init__(self, h, d_model, dropout=0.1):
"""Take in model size and number of heads."""
super(MultiHeadedAttentionNew, self).__init__()
assert d_model % h == 0
self.d_head = d_model // h
self.h = h
self.proj_q = nn.Linear(d_model, d_model)
self.proj_k = nn.Linear(d_model, d_model)
self.proj_v = nn.Linear(d_model, d_model)
self.ret_proj = nn.Linear(d_model, d_model)
self.dropout = nn.Dropout(p=dropout)
def forward(self, input_0, input_1, input_2):
primals_2 = self.proj_q.weight
primals_3 = self.proj_q.bias
primals_4 = self.proj_k.weight
primals_5 = self.proj_k.bias
primals_7 = self.proj_v.weight
primals_8 = self.proj_v.bias
primals_10 = self.ret_proj.weight
primals_11 = self.ret_proj.bias
primals_1 = input_0
primals_6 = input_1
primals_9 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
|
malhotraa/transformer-experiments
|
MultiHeadedAttention
| false | 10,463 |
[
"MIT"
] | 0 |
82931b89b14d26dbd6e4ffef8d6f2fd8b7279c0f
|
https://github.com/malhotraa/transformer-experiments/tree/82931b89b14d26dbd6e4ffef8d6f2fd8b7279c0f
|
GE2ELoss
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/i3/ci3z4sjonlh5q5onksb7pcs7gyh4blejtoz4atayshpru5wafc22.py
# Topologically Sorted Source Nodes: [new_centroids, new_centroids_3, new_centroids_4, new_centroids_7, new_centroids_8, new_centroids_11, new_centroids_12], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# new_centroids => cat
# new_centroids_11 => cat_19
# new_centroids_12 => cat_21
# new_centroids_3 => cat_5
# new_centroids_4 => cat_7
# new_centroids_7 => cat_12
# new_centroids_8 => cat_14
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_1, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_5 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_4, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_7 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_5, %select_11, %select_12],), kwargs = {})
# %cat_12 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_8, %select_11, %select_12],), kwargs = {})
# %cat_14 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_9, %select_12],), kwargs = {})
# %cat_19 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_12, %select_12],), kwargs = {})
# %cat_21 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_13],), kwargs = {})
triton_poi_fused_stack_0 = async_compile.triton('triton_poi_fused_stack_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_stack_0(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (8 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (12 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = 3.0
tmp11 = tmp9 / tmp10
tmp12 = tl.full(tmp11.shape, 0.0, tmp11.dtype)
tmp13 = tl.where(tmp4, tmp11, tmp12)
tmp14 = tmp0 >= tmp3
tmp15 = tl.full([1], 8, tl.int64)
tmp16 = tmp0 < tmp15
tmp17 = tmp14 & tmp16
tmp18 = tl.load(in_ptr0 + (16 + ((-4) + x0)), tmp17 & xmask, eviction_policy='evict_last', other=0.0)
tmp19 = tl.load(in_ptr0 + (20 + ((-4) + x0)), tmp17 & xmask, eviction_policy='evict_last', other=0.0)
tmp20 = tmp18 + tmp19
tmp21 = tl.load(in_ptr0 + (24 + ((-4) + x0)), tmp17 & xmask, eviction_policy='evict_last', other=0.0)
tmp22 = tmp20 + tmp21
tmp23 = tl.load(in_ptr0 + (28 + ((-4) + x0)), tmp17 & xmask, eviction_policy='evict_last', other=0.0)
tmp24 = tmp22 + tmp23
tmp25 = 4.0
tmp26 = tmp24 / tmp25
tmp27 = tl.full(tmp26.shape, 0.0, tmp26.dtype)
tmp28 = tl.where(tmp17, tmp26, tmp27)
tmp29 = tmp0 >= tmp15
tmp30 = tl.full([1], 12, tl.int64)
tmp31 = tmp0 < tmp30
tmp32 = tmp29 & tmp31
tmp33 = tl.load(in_ptr0 + (32 + ((-8) + x0)), tmp32 & xmask, eviction_policy='evict_last', other=0.0)
tmp34 = tl.load(in_ptr0 + (36 + ((-8) + x0)), tmp32 & xmask, eviction_policy='evict_last', other=0.0)
tmp35 = tmp33 + tmp34
tmp36 = tl.load(in_ptr0 + (40 + ((-8) + x0)), tmp32 & xmask, eviction_policy='evict_last', other=0.0)
tmp37 = tmp35 + tmp36
tmp38 = tl.load(in_ptr0 + (44 + ((-8) + x0)), tmp32 & xmask, eviction_policy='evict_last', other=0.0)
tmp39 = tmp37 + tmp38
tmp40 = tmp39 / tmp25
tmp41 = tl.full(tmp40.shape, 0.0, tmp40.dtype)
tmp42 = tl.where(tmp32, tmp40, tmp41)
tmp43 = tmp0 >= tmp30
tmp44 = tl.full([1], 16, tl.int64)
tmp45 = tmp0 < tmp44
tmp46 = tl.load(in_ptr0 + (48 + ((-12) + x0)), tmp43 & xmask, eviction_policy='evict_last', other=0.0)
tmp47 = tl.load(in_ptr0 + (52 + ((-12) + x0)), tmp43 & xmask, eviction_policy='evict_last', other=0.0)
tmp48 = tmp46 + tmp47
tmp49 = tl.load(in_ptr0 + (56 + ((-12) + x0)), tmp43 & xmask, eviction_policy='evict_last', other=0.0)
tmp50 = tmp48 + tmp49
tmp51 = tl.load(in_ptr0 + (60 + ((-12) + x0)), tmp43 & xmask, eviction_policy='evict_last', other=0.0)
tmp52 = tmp50 + tmp51
tmp53 = tmp52 / tmp25
tmp54 = tl.full(tmp53.shape, 0.0, tmp53.dtype)
tmp55 = tl.where(tmp43, tmp53, tmp54)
tmp56 = tl.where(tmp32, tmp42, tmp55)
tmp57 = tl.where(tmp17, tmp28, tmp56)
tmp58 = tl.where(tmp4, tmp13, tmp57)
tmp59 = tl.load(in_ptr0 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp60 = tmp59 + tmp5
tmp61 = tmp60 + tmp6
tmp62 = tmp61 / tmp10
tmp63 = tl.full(tmp62.shape, 0.0, tmp62.dtype)
tmp64 = tl.where(tmp4, tmp62, tmp63)
tmp65 = tl.where(tmp4, tmp64, tmp57)
tmp66 = tmp61 + tmp8
tmp67 = tmp66 / tmp25
tmp68 = tl.full(tmp67.shape, 0.0, tmp67.dtype)
tmp69 = tl.where(tmp4, tmp67, tmp68)
tmp70 = tmp19 + tmp21
tmp71 = tmp70 + tmp23
tmp72 = tmp71 / tmp10
tmp73 = tl.full(tmp72.shape, 0.0, tmp72.dtype)
tmp74 = tl.where(tmp17, tmp72, tmp73)
tmp75 = tl.where(tmp17, tmp74, tmp56)
tmp76 = tl.where(tmp4, tmp69, tmp75)
tmp77 = tmp22 / tmp10
tmp78 = tl.full(tmp77.shape, 0.0, tmp77.dtype)
tmp79 = tl.where(tmp17, tmp77, tmp78)
tmp80 = tl.where(tmp17, tmp79, tmp56)
tmp81 = tl.where(tmp4, tmp69, tmp80)
tmp82 = tmp34 + tmp36
tmp83 = tmp82 + tmp38
tmp84 = tmp83 / tmp10
tmp85 = tl.full(tmp84.shape, 0.0, tmp84.dtype)
tmp86 = tl.where(tmp32, tmp84, tmp85)
tmp87 = tl.where(tmp32, tmp86, tmp55)
tmp88 = tl.where(tmp17, tmp28, tmp87)
tmp89 = tl.where(tmp4, tmp69, tmp88)
tmp90 = tmp37 / tmp10
tmp91 = tl.full(tmp90.shape, 0.0, tmp90.dtype)
tmp92 = tl.where(tmp32, tmp90, tmp91)
tmp93 = tl.where(tmp32, tmp92, tmp55)
tmp94 = tl.where(tmp17, tmp28, tmp93)
tmp95 = tl.where(tmp4, tmp69, tmp94)
tmp96 = tmp47 + tmp49
tmp97 = tmp96 + tmp51
tmp98 = tmp97 / tmp10
tmp99 = tl.full(tmp98.shape, 0.0, tmp98.dtype)
tmp100 = tl.where(tmp43, tmp98, tmp99)
tmp101 = tl.where(tmp32, tmp42, tmp100)
tmp102 = tl.where(tmp17, tmp28, tmp101)
tmp103 = tl.where(tmp4, tmp69, tmp102)
tl.store(out_ptr0 + (x0), tmp58, xmask)
tl.store(out_ptr1 + (x0), tmp65, xmask)
tl.store(out_ptr2 + (x0), tmp76, xmask)
tl.store(out_ptr3 + (x0), tmp81, xmask)
tl.store(out_ptr4 + (x0), tmp89, xmask)
tl.store(out_ptr5 + (x0), tmp95, xmask)
tl.store(out_ptr6 + (x0), tmp103, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ep/cepvlwnn5h76ozdsbquyuiv4b32445eiuhfjkyovopyyh7r6lqdf.py
# Topologically Sorted Source Nodes: [norm, excl_2, excl_3], Original ATen: [aten.linalg_vector_norm, aten.cat, aten.mean]
# Source node to ATen node mapping:
# excl_2 => cat_1
# excl_3 => mean_2
# norm => pow_1, sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_4, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, None), kwargs = {})
# %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_3, %slice_4],), kwargs = {})
# %mean_2 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_1, [0]), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_1 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_1(in_ptr0, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp5 = tl.full([1, 1], 0, tl.int64)
tmp6 = tmp5 >= tmp5
tmp7 = tl.full([1, 1], 1, tl.int64)
tmp8 = tmp5 < tmp7
tmp9 = tl.load(in_ptr0 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp8, other=0.0)
tmp10 = tmp5 >= tmp7
tmp11 = tl.full([1, 1], 3, tl.int64)
tmp12 = tmp5 < tmp11
tmp13 = tl.load(in_ptr0 + (tl.broadcast_to(8 + r0 + (4*(-1)), [XBLOCK, RBLOCK])), tmp10, other=0.0)
tmp14 = tl.where(tmp8, tmp9, tmp13)
tmp15 = tmp7 >= tmp5
tmp16 = tmp7 < tmp7
tmp17 = tl.load(in_ptr0 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp16, other=0.0)
tmp18 = tmp7 >= tmp7
tmp19 = tmp7 < tmp11
tmp20 = tl.load(in_ptr0 + (tl.broadcast_to(8 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp18, other=0.0)
tmp21 = tl.where(tmp16, tmp17, tmp20)
tmp22 = tmp14 + tmp21
tmp23 = tl.full([1, 1], 2, tl.int64)
tmp24 = tmp23 >= tmp5
tmp25 = tmp23 < tmp7
tmp26 = tl.load(in_ptr0 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp25, other=0.0)
tmp27 = tmp23 >= tmp7
tmp28 = tmp23 < tmp11
tmp29 = tl.load(in_ptr0 + (tl.broadcast_to(8 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp27, other=0.0)
tmp30 = tl.where(tmp25, tmp26, tmp29)
tmp31 = tmp22 + tmp30
tmp32 = 3.0
tmp33 = tmp31 / tmp32
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp33, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/mf/cmfeaa5x7m36qp3iyiau26fjfz3dxqznxvc4k6n4bc2o7xru5x2l.py
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, norm_8, excl_10, excl_11], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
# Source node to ATen node mapping:
# excl_10 => cat_8
# excl_11 => mean_6
# new_centroids_1 => cat_2
# new_centroids_2 => cat_4
# norm_8 => pow_17, sum_9
# Graph fragment:
# %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_2, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_4 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_3, %select_10, %select_11, %select_12],), kwargs = {})
# %pow_17 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_28, 2), kwargs = {})
# %sum_9 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_17, None), kwargs = {})
# %cat_8 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_11, %slice_12],), kwargs = {})
# %mean_6 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_8, [0]), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_2 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_2(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (16 + r0), None)
tmp5 = tl.load(in_ptr0 + (20 + r0), None)
tmp7 = tl.load(in_ptr0 + (24 + r0), None)
tmp9 = tl.load(in_ptr0 + (28 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp14 = tmp13 >= tmp13
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK])), tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tmp19 = tl.full([1, 1], 3, tl.int64)
tmp20 = tmp13 < tmp19
tmp21 = tl.load(in_ptr0 + (tl.broadcast_to(24 + r0 + (4*(-1)), [XBLOCK, RBLOCK])), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp23 = tmp15 >= tmp13
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK])), tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp27 = tmp15 < tmp19
tmp28 = tl.load(in_ptr0 + (tl.broadcast_to(24 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp32 = tmp31 >= tmp13
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK])), tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp36 = tmp31 < tmp19
tmp37 = tl.load(in_ptr0 + (tl.broadcast_to(24 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr2 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp41, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/4j/c4jrilvt5mqtiyzl6ukdzjjkxjtxwj4vcmqdi3frlq5doee6agn3.py
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, new_centroids_5, new_centroids_6, norm_16, excl_18, excl_19, norm_18], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
# Source node to ATen node mapping:
# excl_18 => cat_15
# excl_19 => mean_10
# new_centroids_1 => cat_2
# new_centroids_2 => cat_4
# new_centroids_5 => cat_9
# new_centroids_6 => cat_11
# norm_16 => pow_33, sum_17
# norm_18 => pow_37, sum_19
# Graph fragment:
# %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_2, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_4 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_3, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_9 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_6, %select_11, %select_12],), kwargs = {})
# %cat_11 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_7, %select_11, %select_12],), kwargs = {})
# %pow_33 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_52, 2), kwargs = {})
# %sum_17 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_33, None), kwargs = {})
# %cat_15 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_19, %slice_20],), kwargs = {})
# %mean_10 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_15, [0]), kwargs = {})
# %pow_37 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_53, 2), kwargs = {})
# %sum_19 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_37, None), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_3 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32', 9: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {8: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 7), equal_to_1=(8,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_3(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (32 + r0), None)
tmp5 = tl.load(in_ptr0 + (36 + r0), None)
tmp7 = tl.load(in_ptr0 + (40 + r0), None)
tmp9 = tl.load(in_ptr0 + (44 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp14 = tmp13 >= tmp13
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK])), tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tmp19 = tl.full([1, 1], 3, tl.int64)
tmp20 = tmp13 < tmp19
tmp21 = tl.load(in_ptr0 + (tl.broadcast_to(40 + r0 + (4*(-1)), [XBLOCK, RBLOCK])), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp23 = tmp15 >= tmp13
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK])), tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp27 = tmp15 < tmp19
tmp28 = tl.load(in_ptr0 + (tl.broadcast_to(40 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp32 = tmp31 >= tmp13
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK])), tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp36 = tmp31 < tmp19
tmp37 = tl.load(in_ptr0 + (tl.broadcast_to(40 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tmp42 = tmp5 * tmp5
tmp43 = tl.broadcast_to(tmp42, [XBLOCK, RBLOCK])
tmp45 = tl.sum(tmp43, 1)[:, None]
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr2 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr4 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr5 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp41, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
tl.store(out_ptr6 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp45, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jy/cjyty4bxscbvyfultdbj4vrypnzsswd6efeo2rhnwgxg6mg72y7r.py
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, new_centroids_5, new_centroids_6, new_centroids_9, new_centroids_10, norm_24, excl_26, excl_27], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
# Source node to ATen node mapping:
# excl_26 => cat_22
# excl_27 => mean_14
# new_centroids_1 => cat_2
# new_centroids_10 => cat_18
# new_centroids_2 => cat_4
# new_centroids_5 => cat_9
# new_centroids_6 => cat_11
# new_centroids_9 => cat_16
# norm_24 => pow_49, sum_25
# Graph fragment:
# %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_2, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_4 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mean_3, %select_10, %select_11, %select_12],), kwargs = {})
# %cat_9 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_6, %select_11, %select_12],), kwargs = {})
# %cat_11 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_7, %select_11, %select_12],), kwargs = {})
# %cat_16 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_10, %select_12],), kwargs = {})
# %cat_18 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_11, %select_12],), kwargs = {})
# %pow_49 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_76, 2), kwargs = {})
# %sum_25 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_49, None), kwargs = {})
# %cat_22 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_27, %slice_28],), kwargs = {})
# %mean_14 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_22, [0]), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_4 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: 'i32', 10: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {9: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(9,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_4(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (48 + r0), None)
tmp5 = tl.load(in_ptr0 + (52 + r0), None)
tmp7 = tl.load(in_ptr0 + (56 + r0), None)
tmp9 = tl.load(in_ptr0 + (60 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp14 = tmp13 >= tmp13
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK])), tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tmp19 = tl.full([1, 1], 3, tl.int64)
tmp20 = tmp13 < tmp19
tmp21 = tl.load(in_ptr0 + (tl.broadcast_to(56 + r0 + (4*(-1)), [XBLOCK, RBLOCK])), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp23 = tmp15 >= tmp13
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK])), tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp27 = tmp15 < tmp19
tmp28 = tl.load(in_ptr0 + (tl.broadcast_to(56 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp32 = tmp31 >= tmp13
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK])), tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp36 = tmp31 < tmp19
tmp37 = tl.load(in_ptr0 + (tl.broadcast_to(56 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tl.store(out_ptr1 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr2 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr4 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr5 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr6 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp12, None)
tl.store(out_ptr7 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp41, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/wg/cwgqgf5md4z72c43htycwc3ktpi7aaquyuyylibbxopxvlx3xjpw.py
# Topologically Sorted Source Nodes: [norm_2], Original ATen: [aten.linalg_vector_norm]
# Source node to ATen node mapping:
# norm_2 => pow_5, sum_3
# Graph fragment:
# %pow_5 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_5, 2), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_5, None), kwargs = {})
triton_per_fused_linalg_vector_norm_5 = async_compile.triton('triton_per_fused_linalg_vector_norm_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_linalg_vector_norm_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_linalg_vector_norm_5(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (4 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/fn/cfnvzcqi44e6h2vvavig5ohe2okvf7vbtlnytu63pi7tyguwayv6.py
# Topologically Sorted Source Nodes: [excl_4, excl_5, norm_4, norm_6, new_centroids_5, new_centroids_6, new_centroids_9, new_centroids_10, new_centroids_13, new_centroids_14], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm, aten.stack]
# Source node to ATen node mapping:
# excl_4 => cat_3
# excl_5 => mean_3
# new_centroids_10 => cat_18
# new_centroids_13 => cat_23
# new_centroids_14 => cat_25
# new_centroids_5 => cat_9
# new_centroids_6 => cat_11
# new_centroids_9 => cat_16
# norm_4 => pow_9, sum_5
# norm_6 => pow_13, sum_7
# Graph fragment:
# %cat_3 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_5, %slice_6],), kwargs = {})
# %mean_3 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_3, [0]), kwargs = {})
# %pow_9 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_6, 2), kwargs = {})
# %sum_5 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_9, None), kwargs = {})
# %pow_13 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_7, 2), kwargs = {})
# %sum_7 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_13, None), kwargs = {})
# %cat_9 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_6, %select_11, %select_12],), kwargs = {})
# %cat_11 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %mean_7, %select_11, %select_12],), kwargs = {})
# %cat_16 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_10, %select_12],), kwargs = {})
# %cat_18 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_11, %select_12],), kwargs = {})
# %cat_23 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_14],), kwargs = {})
# %cat_25 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_15],), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_6 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: 'i32', 11: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {10: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), equal_to_1=(10,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_6(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7, out_ptr8, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (8 + r0), None)
tmp5 = tl.load(in_ptr0 + (12 + r0), None)
tmp39 = tl.load(in_ptr0 + (r0), None)
tmp40 = tl.load(in_ptr0 + (4 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp10 = tl.full([1, 1], 0, tl.int64)
tmp11 = tmp10 >= tmp10
tmp12 = tl.full([1, 1], 2, tl.int64)
tmp13 = tmp10 < tmp12
tmp14 = tl.load(in_ptr0 + (tl.broadcast_to(r0 + (4*0), [XBLOCK, RBLOCK])), tmp13, other=0.0)
tmp15 = tmp10 >= tmp12
tmp16 = tl.full([1, 1], 3, tl.int64)
tmp17 = tmp10 < tmp16
tmp18 = tl.load(in_ptr0 + (tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK])), tmp15, other=0.0)
tmp19 = tl.where(tmp13, tmp14, tmp18)
tmp20 = tl.full([1, 1], 1, tl.int64)
tmp21 = tmp20 >= tmp10
tmp22 = tmp20 < tmp12
tmp23 = tl.load(in_ptr0 + (tl.broadcast_to(r0 + (4*1), [XBLOCK, RBLOCK])), tmp22, other=0.0)
tmp24 = tmp20 >= tmp12
tmp25 = tmp20 < tmp16
tmp26 = tl.load(in_ptr0 + (tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK])), tmp24, other=0.0)
tmp27 = tl.where(tmp22, tmp23, tmp26)
tmp28 = tmp19 + tmp27
tmp29 = tmp12 >= tmp10
tmp30 = tmp12 < tmp12
tmp31 = tl.load(in_ptr0 + (tl.broadcast_to(r0 + (4*2), [XBLOCK, RBLOCK])), tmp30, other=0.0)
tmp32 = tmp12 >= tmp12
tmp33 = tmp12 < tmp16
tmp34 = tl.load(in_ptr0 + (tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK])), tmp32, other=0.0)
tmp35 = tl.where(tmp30, tmp31, tmp34)
tmp36 = tmp28 + tmp35
tmp37 = 3.0
tmp38 = tmp36 / tmp37
tmp41 = tmp39 + tmp40
tmp42 = tmp41 + tmp0
tmp43 = tmp42 + tmp5
tmp44 = 4.0
tmp45 = tmp43 / tmp44
tl.store(out_ptr2 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp38, None)
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr4 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr5 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr6 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr7 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr8 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/uv/cuv3j34526wda3guhmc53tumbwtq6kgrjwarktyyx43qheojwlvy.py
# Topologically Sorted Source Nodes: [cs_row], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# cs_row => cat_6
# Graph fragment:
# %cat_6 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%clamp_min, %clamp_min_1, %clamp_min_2, %clamp_min_3],), kwargs = {})
triton_poi_fused_cat_7 = async_compile.triton('triton_poi_fused_cat_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_7', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 24, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_7(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10, in_ptr11, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x0 = xindex % 4
x2 = xindex
tmp6 = tl.load(in_ptr1 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp32 = tl.load(in_ptr4 + (0))
tmp33 = tl.broadcast_to(tmp32, [XBLOCK])
tmp57 = tl.load(in_ptr7 + (0))
tmp58 = tl.broadcast_to(tmp57, [XBLOCK])
tmp81 = tl.load(in_ptr10 + (0))
tmp82 = tl.broadcast_to(tmp81, [XBLOCK])
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp8 = libdevice.sqrt(tmp7)
tmp9 = tl.load(in_ptr2 + (4*x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp10 = tmp9 * tmp9
tmp11 = tl.load(in_ptr2 + (1 + (4*x0)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tl.load(in_ptr2 + (2 + (4*x0)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tl.load(in_ptr2 + (3 + (4*x0)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = libdevice.sqrt(tmp19)
tmp21 = tmp8 * tmp20
tmp22 = tmp5 / tmp21
tmp23 = 1e-06
tmp24 = triton_helpers.maximum(tmp22, tmp23)
tmp25 = tl.full(tmp24.shape, 0.0, tmp24.dtype)
tmp26 = tl.where(tmp4, tmp24, tmp25)
tmp27 = tmp0 >= tmp3
tmp28 = tl.full([1], 2, tl.int64)
tmp29 = tmp0 < tmp28
tmp30 = tmp27 & tmp29
tmp31 = tl.load(in_ptr3 + (x0), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp34 = libdevice.sqrt(tmp33)
tmp35 = tl.load(in_ptr5 + (4*x0), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp36 = tmp35 * tmp35
tmp37 = tl.load(in_ptr5 + (1 + (4*x0)), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp38 = tmp37 * tmp37
tmp39 = tmp36 + tmp38
tmp40 = tl.load(in_ptr5 + (2 + (4*x0)), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp41 = tmp40 * tmp40
tmp42 = tmp39 + tmp41
tmp43 = tl.load(in_ptr5 + (3 + (4*x0)), tmp30 & xmask, eviction_policy='evict_last', other=0.0)
tmp44 = tmp43 * tmp43
tmp45 = tmp42 + tmp44
tmp46 = libdevice.sqrt(tmp45)
tmp47 = tmp34 * tmp46
tmp48 = tmp31 / tmp47
tmp49 = triton_helpers.maximum(tmp48, tmp23)
tmp50 = tl.full(tmp49.shape, 0.0, tmp49.dtype)
tmp51 = tl.where(tmp30, tmp49, tmp50)
tmp52 = tmp0 >= tmp28
tmp53 = tl.full([1], 3, tl.int64)
tmp54 = tmp0 < tmp53
tmp55 = tmp52 & tmp54
tmp56 = tl.load(in_ptr6 + (x0), tmp55 & xmask, eviction_policy='evict_last', other=0.0)
tmp59 = libdevice.sqrt(tmp58)
tmp60 = tl.load(in_ptr8 + (4*x0), tmp55 & xmask, eviction_policy='evict_last', other=0.0)
tmp61 = tmp60 * tmp60
tmp62 = tl.load(in_ptr8 + (1 + (4*x0)), tmp55 & xmask, eviction_policy='evict_last', other=0.0)
tmp63 = tmp62 * tmp62
tmp64 = tmp61 + tmp63
tmp65 = tl.load(in_ptr8 + (2 + (4*x0)), tmp55 & xmask, eviction_policy='evict_last', other=0.0)
tmp66 = tmp65 * tmp65
tmp67 = tmp64 + tmp66
tmp68 = tl.load(in_ptr8 + (3 + (4*x0)), tmp55 & xmask, eviction_policy='evict_last', other=0.0)
tmp69 = tmp68 * tmp68
tmp70 = tmp67 + tmp69
tmp71 = libdevice.sqrt(tmp70)
tmp72 = tmp59 * tmp71
tmp73 = tmp56 / tmp72
tmp74 = triton_helpers.maximum(tmp73, tmp23)
tmp75 = tl.full(tmp74.shape, 0.0, tmp74.dtype)
tmp76 = tl.where(tmp55, tmp74, tmp75)
tmp77 = tmp0 >= tmp53
tmp78 = tl.full([1], 4, tl.int64)
tmp79 = tmp0 < tmp78
tmp80 = tl.load(in_ptr9 + (x0), tmp77 & xmask, eviction_policy='evict_last', other=0.0)
tmp83 = libdevice.sqrt(tmp82)
tmp84 = tl.load(in_ptr11 + (4*x0), tmp77 & xmask, eviction_policy='evict_last', other=0.0)
tmp85 = tmp84 * tmp84
tmp86 = tl.load(in_ptr11 + (1 + (4*x0)), tmp77 & xmask, eviction_policy='evict_last', other=0.0)
tmp87 = tmp86 * tmp86
tmp88 = tmp85 + tmp87
tmp89 = tl.load(in_ptr11 + (2 + (4*x0)), tmp77 & xmask, eviction_policy='evict_last', other=0.0)
tmp90 = tmp89 * tmp89
tmp91 = tmp88 + tmp90
tmp92 = tl.load(in_ptr11 + (3 + (4*x0)), tmp77 & xmask, eviction_policy='evict_last', other=0.0)
tmp93 = tmp92 * tmp92
tmp94 = tmp91 + tmp93
tmp95 = libdevice.sqrt(tmp94)
tmp96 = tmp83 * tmp95
tmp97 = tmp80 / tmp96
tmp98 = triton_helpers.maximum(tmp97, tmp23)
tmp99 = tl.full(tmp98.shape, 0.0, tmp98.dtype)
tmp100 = tl.where(tmp77, tmp98, tmp99)
tmp101 = tl.where(tmp55, tmp76, tmp100)
tmp102 = tl.where(tmp30, tmp51, tmp101)
tmp103 = tl.where(tmp4, tmp26, tmp102)
tl.store(out_ptr0 + (x2), tmp103, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/qw/cqwtxu5kn2bvn6whkodf5hex3x352f66d5yqfcryzy5xmydy2txx.py
# Topologically Sorted Source Nodes: [norm_10, excl_12, excl_13, norm_12, norm_14, new_centroids_9, new_centroids_10, new_centroids_13, new_centroids_14], Original ATen: [aten.linalg_vector_norm, aten.cat, aten.mean, aten.stack]
# Source node to ATen node mapping:
# excl_12 => cat_10
# excl_13 => mean_7
# new_centroids_10 => cat_18
# new_centroids_13 => cat_23
# new_centroids_14 => cat_25
# new_centroids_9 => cat_16
# norm_10 => pow_21, sum_11
# norm_12 => pow_25, sum_13
# norm_14 => pow_29, sum_15
# Graph fragment:
# %pow_21 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_29, 2), kwargs = {})
# %sum_11 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_21, None), kwargs = {})
# %cat_10 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_13, %slice_14],), kwargs = {})
# %mean_7 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_10, [0]), kwargs = {})
# %pow_25 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_30, 2), kwargs = {})
# %sum_13 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_25, None), kwargs = {})
# %pow_29 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_31, 2), kwargs = {})
# %sum_15 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_29, None), kwargs = {})
# %cat_16 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_10, %select_12],), kwargs = {})
# %cat_18 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %mean_11, %select_12],), kwargs = {})
# %cat_23 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_14],), kwargs = {})
# %cat_25 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_15],), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_8 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: 'i32', 10: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {9: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=(9,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_8', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 3, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_8(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (20 + r0), None)
tmp5 = tl.load(in_ptr0 + (24 + r0), None)
tmp10 = tl.load(in_ptr0 + (28 + r0), None)
tmp44 = tl.load(in_ptr0 + (16 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp11 = tmp10 * tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp15 = tl.full([1, 1], 0, tl.int64)
tmp16 = tmp15 >= tmp15
tmp17 = tl.full([1, 1], 2, tl.int64)
tmp18 = tmp15 < tmp17
tmp19 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp18, other=0.0)
tmp20 = tmp15 >= tmp17
tmp21 = tl.full([1, 1], 3, tl.int64)
tmp22 = tmp15 < tmp21
tmp23 = tl.load(in_ptr0 + (tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK])), tmp20, other=0.0)
tmp24 = tl.where(tmp18, tmp19, tmp23)
tmp25 = tl.full([1, 1], 1, tl.int64)
tmp26 = tmp25 >= tmp15
tmp27 = tmp25 < tmp17
tmp28 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp27, other=0.0)
tmp29 = tmp25 >= tmp17
tmp30 = tmp25 < tmp21
tmp31 = tl.load(in_ptr0 + (tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK])), tmp29, other=0.0)
tmp32 = tl.where(tmp27, tmp28, tmp31)
tmp33 = tmp24 + tmp32
tmp34 = tmp17 >= tmp15
tmp35 = tmp17 < tmp17
tmp36 = tl.load(in_ptr0 + (tl.broadcast_to(16 + r0 + (4*2), [XBLOCK, RBLOCK])), tmp35, other=0.0)
tmp37 = tmp17 >= tmp17
tmp38 = tmp17 < tmp21
tmp39 = tl.load(in_ptr0 + (tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK])), tmp37, other=0.0)
tmp40 = tl.where(tmp35, tmp36, tmp39)
tmp41 = tmp33 + tmp40
tmp42 = 3.0
tmp43 = tmp41 / tmp42
tmp45 = tmp44 + tmp0
tmp46 = tmp45 + tmp5
tmp47 = tmp46 + tmp10
tmp48 = 4.0
tmp49 = tmp47 / tmp48
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp43, None)
tl.store(out_ptr4 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp49, None)
tl.store(out_ptr5 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp49, None)
tl.store(out_ptr6 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp49, None)
tl.store(out_ptr7 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp49, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, None)
tl.store(out_ptr2 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp14, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/te/cteub7ll6kedydjue4iwrcci3eookcmcdgpnser3qmghytygolqj.py
# Topologically Sorted Source Nodes: [excl_20, excl_21, norm_20, norm_22, new_centroids_13, new_centroids_14], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm, aten.stack]
# Source node to ATen node mapping:
# excl_20 => cat_17
# excl_21 => mean_11
# new_centroids_13 => cat_23
# new_centroids_14 => cat_25
# norm_20 => pow_41, sum_21
# norm_22 => pow_45, sum_23
# Graph fragment:
# %cat_17 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_21, %slice_22],), kwargs = {})
# %mean_11 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_17, [0]), kwargs = {})
# %pow_41 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_54, 2), kwargs = {})
# %sum_21 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_41, None), kwargs = {})
# %pow_45 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_55, 2), kwargs = {})
# %sum_23 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_45, None), kwargs = {})
# %cat_23 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_14],), kwargs = {})
# %cat_25 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_15],), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_stack_9 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_stack_9', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {6: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=(6,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_stack_9', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_9(in_ptr0, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (40 + r0), None)
tmp5 = tl.load(in_ptr0 + (44 + r0), None)
tmp39 = tl.load(in_ptr0 + (32 + r0), None)
tmp40 = tl.load(in_ptr0 + (36 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp10 = tl.full([1, 1], 0, tl.int64)
tmp11 = tmp10 >= tmp10
tmp12 = tl.full([1, 1], 2, tl.int64)
tmp13 = tmp10 < tmp12
tmp14 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp13, other=0.0)
tmp15 = tmp10 >= tmp12
tmp16 = tl.full([1, 1], 3, tl.int64)
tmp17 = tmp10 < tmp16
tmp18 = tl.load(in_ptr0 + (tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK])), tmp15, other=0.0)
tmp19 = tl.where(tmp13, tmp14, tmp18)
tmp20 = tl.full([1, 1], 1, tl.int64)
tmp21 = tmp20 >= tmp10
tmp22 = tmp20 < tmp12
tmp23 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp22, other=0.0)
tmp24 = tmp20 >= tmp12
tmp25 = tmp20 < tmp16
tmp26 = tl.load(in_ptr0 + (tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK])), tmp24, other=0.0)
tmp27 = tl.where(tmp22, tmp23, tmp26)
tmp28 = tmp19 + tmp27
tmp29 = tmp12 >= tmp10
tmp30 = tmp12 < tmp12
tmp31 = tl.load(in_ptr0 + (tl.broadcast_to(32 + r0 + (4*2), [XBLOCK, RBLOCK])), tmp30, other=0.0)
tmp32 = tmp12 >= tmp12
tmp33 = tmp12 < tmp16
tmp34 = tl.load(in_ptr0 + (tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK])), tmp32, other=0.0)
tmp35 = tl.where(tmp30, tmp31, tmp34)
tmp36 = tmp28 + tmp35
tmp37 = 3.0
tmp38 = tmp36 / tmp37
tmp41 = tmp39 + tmp40
tmp42 = tmp41 + tmp0
tmp43 = tmp42 + tmp5
tmp44 = 4.0
tmp45 = tmp43 / tmp44
tl.store(out_ptr2 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp38, None)
tl.store(out_ptr3 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr4 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp45, None)
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zz/czzgiodr25igbxfpbzkqi75p57p62kkx2z7qtcr7iljk7knobkfe.py
# Topologically Sorted Source Nodes: [norm_26], Original ATen: [aten.linalg_vector_norm]
# Source node to ATen node mapping:
# norm_26 => pow_53, sum_27
# Graph fragment:
# %pow_53 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_77, 2), kwargs = {})
# %sum_27 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_53, None), kwargs = {})
triton_per_fused_linalg_vector_norm_10 = async_compile.triton('triton_per_fused_linalg_vector_norm_10', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_linalg_vector_norm_10', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_linalg_vector_norm_10(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (52 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/36/c36assjyhb77brp4nnctqswixa5jroretdu3sslbnwzkvoxh4lw5.py
# Topologically Sorted Source Nodes: [excl_28, excl_29, norm_30], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm]
# Source node to ATen node mapping:
# excl_28 => cat_24
# excl_29 => mean_15
# norm_30 => pow_61, sum_31
# Graph fragment:
# %cat_24 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_29, %slice_30],), kwargs = {})
# %mean_15 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%cat_24, [0]), kwargs = {})
# %pow_61 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_79, 2), kwargs = {})
# %sum_31 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_61, None), kwargs = {})
triton_per_fused_cat_linalg_vector_norm_mean_11 = async_compile.triton('triton_per_fused_cat_linalg_vector_norm_mean_11', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 2), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cat_linalg_vector_norm_mean_11', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_11(in_ptr0, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp29 = tl.load(in_ptr0 + (60 + r0), None)
tmp0 = tl.full([1, 1], 0, tl.int64)
tmp1 = tmp0 >= tmp0
tmp2 = tl.full([1, 1], 2, tl.int64)
tmp3 = tmp0 < tmp2
tmp4 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0 + (4*0), [XBLOCK, RBLOCK])), tmp3, other=0.0)
tmp5 = tmp0 >= tmp2
tmp6 = tl.full([1, 1], 3, tl.int64)
tmp7 = tmp0 < tmp6
tmp8 = tl.load(in_ptr0 + (tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK])), tmp5, other=0.0)
tmp9 = tl.where(tmp3, tmp4, tmp8)
tmp10 = tl.full([1, 1], 1, tl.int64)
tmp11 = tmp10 >= tmp0
tmp12 = tmp10 < tmp2
tmp13 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0 + (4*1), [XBLOCK, RBLOCK])), tmp12, other=0.0)
tmp14 = tmp10 >= tmp2
tmp15 = tmp10 < tmp6
tmp16 = tl.load(in_ptr0 + (tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK])), tmp14, other=0.0)
tmp17 = tl.where(tmp12, tmp13, tmp16)
tmp18 = tmp9 + tmp17
tmp19 = tmp2 >= tmp0
tmp20 = tmp2 < tmp2
tmp21 = tl.load(in_ptr0 + (tl.broadcast_to(48 + r0 + (4*2), [XBLOCK, RBLOCK])), tmp20, other=0.0)
tmp22 = tmp2 >= tmp2
tmp23 = tmp2 < tmp6
tmp24 = tl.load(in_ptr0 + (tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK])), tmp22, other=0.0)
tmp25 = tl.where(tmp20, tmp21, tmp24)
tmp26 = tmp18 + tmp25
tmp27 = 3.0
tmp28 = tmp26 / tmp27
tmp30 = tmp29 * tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr0 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp28, None)
tl.store(out_ptr1 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp33, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/rb/crbhwfxwz4knodchqf4oud2ojiulhvbli6snxsgr56f42mcvtvqr.py
# Topologically Sorted Source Nodes: [norm_28], Original ATen: [aten.linalg_vector_norm]
# Source node to ATen node mapping:
# norm_28 => pow_57, sum_29
# Graph fragment:
# %pow_57 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_78, 2), kwargs = {})
# %sum_29 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_57, None), kwargs = {})
triton_per_fused_linalg_vector_norm_12 = async_compile.triton('triton_per_fused_linalg_vector_norm_12', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_linalg_vector_norm_12', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_linalg_vector_norm_12(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (56 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/cs/ccsc3gkgu5deyhztcxdjwui2gavgf62f6nhcxahznhzc4rpkrqgu.py
# Topologically Sorted Source Nodes: [new_centroids_15], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# new_centroids_15 => cat_26
# Graph fragment:
# %cat_26 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%select_34, %select_10, %select_11, %mean_16],), kwargs = {})
triton_poi_fused_stack_13 = async_compile.triton('triton_poi_fused_stack_13', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_13', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 15, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_stack_13(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (4 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (8 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = tl.load(in_ptr0 + (12 + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp11 = tmp9 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp4, tmp13, tmp14)
tmp16 = tmp0 >= tmp3
tmp17 = tl.full([1], 8, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr0 + (16 + ((-4) + x0)), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp21 = tl.load(in_ptr0 + (20 + ((-4) + x0)), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp22 = tmp20 + tmp21
tmp23 = tl.load(in_ptr0 + (24 + ((-4) + x0)), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp24 = tmp22 + tmp23
tmp25 = tl.load(in_ptr0 + (28 + ((-4) + x0)), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp26 = tmp24 + tmp25
tmp27 = tmp26 / tmp12
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp19, tmp27, tmp28)
tmp30 = tmp0 >= tmp17
tmp31 = tl.full([1], 12, tl.int64)
tmp32 = tmp0 < tmp31
tmp33 = tmp30 & tmp32
tmp34 = tl.load(in_ptr0 + (32 + ((-8) + x0)), tmp33 & xmask, eviction_policy='evict_last', other=0.0)
tmp35 = tl.load(in_ptr0 + (36 + ((-8) + x0)), tmp33 & xmask, eviction_policy='evict_last', other=0.0)
tmp36 = tmp34 + tmp35
tmp37 = tl.load(in_ptr0 + (40 + ((-8) + x0)), tmp33 & xmask, eviction_policy='evict_last', other=0.0)
tmp38 = tmp36 + tmp37
tmp39 = tl.load(in_ptr0 + (44 + ((-8) + x0)), tmp33 & xmask, eviction_policy='evict_last', other=0.0)
tmp40 = tmp38 + tmp39
tmp41 = tmp40 / tmp12
tmp42 = tl.full(tmp41.shape, 0.0, tmp41.dtype)
tmp43 = tl.where(tmp33, tmp41, tmp42)
tmp44 = tmp0 >= tmp31
tmp45 = tl.full([1], 16, tl.int64)
tmp46 = tmp0 < tmp45
tmp47 = tl.load(in_ptr0 + (48 + ((-12) + x0)), tmp44 & xmask, eviction_policy='evict_last', other=0.0)
tmp48 = tl.load(in_ptr0 + (52 + ((-12) + x0)), tmp44 & xmask, eviction_policy='evict_last', other=0.0)
tmp49 = tmp47 + tmp48
tmp50 = tl.load(in_ptr0 + (56 + ((-12) + x0)), tmp44 & xmask, eviction_policy='evict_last', other=0.0)
tmp51 = tmp49 + tmp50
tmp52 = 3.0
tmp53 = tmp51 / tmp52
tmp54 = tl.full(tmp53.shape, 0.0, tmp53.dtype)
tmp55 = tl.where(tmp44, tmp53, tmp54)
tmp56 = tl.where(tmp33, tmp43, tmp55)
tmp57 = tl.where(tmp19, tmp29, tmp56)
tmp58 = tl.where(tmp4, tmp15, tmp57)
tl.store(out_ptr0 + (x0), tmp58, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ny/cnyzf7svi64n4lep576ucnmgkjgspkxsy67schyxekoirfh3tw6i.py
# Topologically Sorted Source Nodes: [log_softmax, log_softmax_1, log_softmax_2, log_softmax_3, L_row, log_softmax_4, log_softmax_5, log_softmax_6, log_softmax_7, L_row_1, log_softmax_8, log_softmax_9, log_softmax_10, log_softmax_11, L_row_2, log_softmax_12, log_softmax_13, log_softmax_14, log_softmax_15], Original ATen: [aten._log_softmax, aten.stack, aten._log_softmax_backward_data]
# Source node to ATen node mapping:
# L_row => cat_29
# L_row_1 => cat_30
# L_row_2 => cat_31
# log_softmax => amax, exp, sub, sum_33
# log_softmax_1 => amax_1, exp_1, sub_2, sum_34
# log_softmax_10 => amax_10, exp_10, sub_20, sum_43
# log_softmax_11 => amax_11, exp_11, sub_22, sum_44
# log_softmax_12 => amax_12, exp_12, log_12, sub_24, sub_25, sum_45
# log_softmax_13 => amax_13, exp_13, log_13, sub_26, sub_27, sum_46
# log_softmax_14 => amax_14, exp_14, log_14, sub_28, sub_29, sum_47
# log_softmax_15 => amax_15, exp_15, log_15, sub_30, sub_31, sum_48
# log_softmax_2 => amax_2, exp_2, sub_4, sum_35
# log_softmax_3 => amax_3, exp_3, sub_6, sum_36
# log_softmax_4 => amax_4, exp_4, sub_8, sum_37
# log_softmax_5 => amax_5, exp_5, sub_10, sum_38
# log_softmax_6 => amax_6, exp_6, sub_12, sum_39
# log_softmax_7 => amax_7, exp_7, sub_14, sum_40
# log_softmax_8 => amax_8, exp_8, sub_16, sum_41
# log_softmax_9 => amax_9, exp_9, sub_18, sum_42
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_101, [0], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_101, %amax), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_33 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [0], True), kwargs = {})
# %amax_1 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_104, [0], True), kwargs = {})
# %sub_2 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_104, %amax_1), kwargs = {})
# %exp_1 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_2,), kwargs = {})
# %sum_34 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_1, [0], True), kwargs = {})
# %amax_2 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_107, [0], True), kwargs = {})
# %sub_4 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_107, %amax_2), kwargs = {})
# %exp_2 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_4,), kwargs = {})
# %sum_35 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_2, [0], True), kwargs = {})
# %amax_3 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_110, [0], True), kwargs = {})
# %sub_6 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_110, %amax_3), kwargs = {})
# %exp_3 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_6,), kwargs = {})
# %sum_36 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_3, [0], True), kwargs = {})
# %cat_29 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%unsqueeze_16, %unsqueeze_17, %unsqueeze_18, %unsqueeze_19],), kwargs = {})
# %amax_4 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_113, [0], True), kwargs = {})
# %sub_8 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_113, %amax_4), kwargs = {})
# %exp_4 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_8,), kwargs = {})
# %sum_37 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_4, [0], True), kwargs = {})
# %amax_5 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_116, [0], True), kwargs = {})
# %sub_10 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_116, %amax_5), kwargs = {})
# %exp_5 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_10,), kwargs = {})
# %sum_38 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_5, [0], True), kwargs = {})
# %amax_6 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_119, [0], True), kwargs = {})
# %sub_12 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_119, %amax_6), kwargs = {})
# %exp_6 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_12,), kwargs = {})
# %sum_39 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_6, [0], True), kwargs = {})
# %amax_7 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_122, [0], True), kwargs = {})
# %sub_14 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_122, %amax_7), kwargs = {})
# %exp_7 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_14,), kwargs = {})
# %sum_40 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_7, [0], True), kwargs = {})
# %cat_30 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%unsqueeze_20, %unsqueeze_21, %unsqueeze_22, %unsqueeze_23],), kwargs = {})
# %amax_8 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_125, [0], True), kwargs = {})
# %sub_16 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_125, %amax_8), kwargs = {})
# %exp_8 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_16,), kwargs = {})
# %sum_41 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_8, [0], True), kwargs = {})
# %amax_9 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_128, [0], True), kwargs = {})
# %sub_18 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_128, %amax_9), kwargs = {})
# %exp_9 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_18,), kwargs = {})
# %sum_42 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_9, [0], True), kwargs = {})
# %amax_10 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_131, [0], True), kwargs = {})
# %sub_20 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_131, %amax_10), kwargs = {})
# %exp_10 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_20,), kwargs = {})
# %sum_43 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_10, [0], True), kwargs = {})
# %amax_11 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_134, [0], True), kwargs = {})
# %sub_22 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_134, %amax_11), kwargs = {})
# %exp_11 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_22,), kwargs = {})
# %sum_44 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_11, [0], True), kwargs = {})
# %cat_31 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%unsqueeze_24, %unsqueeze_25, %unsqueeze_26, %unsqueeze_27],), kwargs = {})
# %amax_12 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_137, [0], True), kwargs = {})
# %sub_24 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_137, %amax_12), kwargs = {})
# %exp_12 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_24,), kwargs = {})
# %sum_45 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_12, [0], True), kwargs = {})
# %log_12 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_45,), kwargs = {})
# %sub_25 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_24, %log_12), kwargs = {})
# %amax_13 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_140, [0], True), kwargs = {})
# %sub_26 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_140, %amax_13), kwargs = {})
# %exp_13 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_26,), kwargs = {})
# %sum_46 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_13, [0], True), kwargs = {})
# %log_13 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_46,), kwargs = {})
# %sub_27 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_26, %log_13), kwargs = {})
# %amax_14 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_143, [0], True), kwargs = {})
# %sub_28 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_143, %amax_14), kwargs = {})
# %exp_14 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_28,), kwargs = {})
# %sum_47 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_14, [0], True), kwargs = {})
# %log_14 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_47,), kwargs = {})
# %sub_29 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_28, %log_14), kwargs = {})
# %amax_15 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%select_146, [0], True), kwargs = {})
# %sub_30 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%select_146, %amax_15), kwargs = {})
# %exp_15 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_30,), kwargs = {})
# %sum_48 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_15, [0], True), kwargs = {})
# %log_15 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_48,), kwargs = {})
# %sub_31 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_30, %log_15), kwargs = {})
# %exp_16 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_31,), kwargs = {})
# %exp_17 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_29,), kwargs = {})
# %exp_18 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_27,), kwargs = {})
# %exp_19 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_25,), kwargs = {})
triton_per_fused__log_softmax__log_softmax_backward_data_stack_14 = async_compile.triton('triton_per_fused__log_softmax__log_softmax_backward_data_stack_14', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: 'i32', 15: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {14: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 6, 7, 8, 9, 10, 11, 12, 13), equal_to_1=(14,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__log_softmax__log_softmax_backward_data_stack_14', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 30, 'num_reduction': 32, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__log_softmax__log_softmax_backward_data_stack_14(in_ptr0, in_ptr1, in_ptr2, out_ptr32, out_ptr33, out_ptr34, out_ptr35, out_ptr36, out_ptr37, out_ptr38, out_ptr39, out_ptr40, out_ptr41, out_ptr42, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp2 = tl.load(in_ptr1 + (r0), None)
tmp4 = tl.load(in_ptr2 + (0))
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp15 = tl.load(in_ptr1 + (4 + r0), None)
tmp26 = tl.load(in_ptr1 + (8 + r0), None)
tmp37 = tl.load(in_ptr1 + (12 + r0), None)
tmp48 = tl.load(in_ptr1 + (16 + r0), None)
tmp59 = tl.load(in_ptr1 + (20 + r0), None)
tmp70 = tl.load(in_ptr1 + (24 + r0), None)
tmp81 = tl.load(in_ptr1 + (28 + r0), None)
tmp92 = tl.load(in_ptr1 + (32 + r0), None)
tmp103 = tl.load(in_ptr1 + (36 + r0), None)
tmp114 = tl.load(in_ptr1 + (40 + r0), None)
tmp125 = tl.load(in_ptr1 + (44 + r0), None)
tmp136 = tl.load(in_ptr1 + (48 + r0), None)
tmp147 = tl.load(in_ptr1 + (52 + r0), None)
tmp158 = tl.load(in_ptr1 + (56 + r0), None)
tmp169 = tl.load(in_ptr1 + (60 + r0), None)
tmp185 = tl.load(in_ptr1 + (0))
tmp186 = tl.broadcast_to(tmp185, [XBLOCK, RBLOCK])
tmp199 = tl.load(in_ptr1 + (4))
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp213 = tl.load(in_ptr1 + (8))
tmp214 = tl.broadcast_to(tmp213, [XBLOCK, RBLOCK])
tmp226 = tl.load(in_ptr1 + (12))
tmp227 = tl.broadcast_to(tmp226, [XBLOCK, RBLOCK])
tmp239 = tl.load(in_ptr1 + (17))
tmp240 = tl.broadcast_to(tmp239, [XBLOCK, RBLOCK])
tmp249 = tl.load(in_ptr1 + (21))
tmp250 = tl.broadcast_to(tmp249, [XBLOCK, RBLOCK])
tmp259 = tl.load(in_ptr1 + (25))
tmp260 = tl.broadcast_to(tmp259, [XBLOCK, RBLOCK])
tmp269 = tl.load(in_ptr1 + (29))
tmp270 = tl.broadcast_to(tmp269, [XBLOCK, RBLOCK])
tmp282 = tl.load(in_ptr1 + (34))
tmp283 = tl.broadcast_to(tmp282, [XBLOCK, RBLOCK])
tmp292 = tl.load(in_ptr1 + (38))
tmp293 = tl.broadcast_to(tmp292, [XBLOCK, RBLOCK])
tmp302 = tl.load(in_ptr1 + (42))
tmp303 = tl.broadcast_to(tmp302, [XBLOCK, RBLOCK])
tmp312 = tl.load(in_ptr1 + (46))
tmp313 = tl.broadcast_to(tmp312, [XBLOCK, RBLOCK])
tmp3 = tmp1 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = triton_helpers.max2(tmp7, 1)[:, None]
tmp10 = tmp6 - tmp9
tmp11 = tl_math.exp(tmp10)
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp16 = tmp1 * tmp15
tmp17 = tmp16 + tmp5
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = triton_helpers.max2(tmp18, 1)[:, None]
tmp21 = tmp17 - tmp20
tmp22 = tl_math.exp(tmp21)
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.sum(tmp23, 1)[:, None]
tmp27 = tmp1 * tmp26
tmp28 = tmp27 + tmp5
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = triton_helpers.max2(tmp29, 1)[:, None]
tmp32 = tmp28 - tmp31
tmp33 = tl_math.exp(tmp32)
tmp34 = tl.broadcast_to(tmp33, [XBLOCK, RBLOCK])
tmp36 = tl.sum(tmp34, 1)[:, None]
tmp38 = tmp1 * tmp37
tmp39 = tmp38 + tmp5
tmp40 = tl.broadcast_to(tmp39, [XBLOCK, RBLOCK])
tmp42 = triton_helpers.max2(tmp40, 1)[:, None]
tmp43 = tmp39 - tmp42
tmp44 = tl_math.exp(tmp43)
tmp45 = tl.broadcast_to(tmp44, [XBLOCK, RBLOCK])
tmp47 = tl.sum(tmp45, 1)[:, None]
tmp49 = tmp1 * tmp48
tmp50 = tmp49 + tmp5
tmp51 = tl.broadcast_to(tmp50, [XBLOCK, RBLOCK])
tmp53 = triton_helpers.max2(tmp51, 1)[:, None]
tmp54 = tmp50 - tmp53
tmp55 = tl_math.exp(tmp54)
tmp56 = tl.broadcast_to(tmp55, [XBLOCK, RBLOCK])
tmp58 = tl.sum(tmp56, 1)[:, None]
tmp60 = tmp1 * tmp59
tmp61 = tmp60 + tmp5
tmp62 = tl.broadcast_to(tmp61, [XBLOCK, RBLOCK])
tmp64 = triton_helpers.max2(tmp62, 1)[:, None]
tmp65 = tmp61 - tmp64
tmp66 = tl_math.exp(tmp65)
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = tl.sum(tmp67, 1)[:, None]
tmp71 = tmp1 * tmp70
tmp72 = tmp71 + tmp5
tmp73 = tl.broadcast_to(tmp72, [XBLOCK, RBLOCK])
tmp75 = triton_helpers.max2(tmp73, 1)[:, None]
tmp76 = tmp72 - tmp75
tmp77 = tl_math.exp(tmp76)
tmp78 = tl.broadcast_to(tmp77, [XBLOCK, RBLOCK])
tmp80 = tl.sum(tmp78, 1)[:, None]
tmp82 = tmp1 * tmp81
tmp83 = tmp82 + tmp5
tmp84 = tl.broadcast_to(tmp83, [XBLOCK, RBLOCK])
tmp86 = triton_helpers.max2(tmp84, 1)[:, None]
tmp87 = tmp83 - tmp86
tmp88 = tl_math.exp(tmp87)
tmp89 = tl.broadcast_to(tmp88, [XBLOCK, RBLOCK])
tmp91 = tl.sum(tmp89, 1)[:, None]
tmp93 = tmp1 * tmp92
tmp94 = tmp93 + tmp5
tmp95 = tl.broadcast_to(tmp94, [XBLOCK, RBLOCK])
tmp97 = triton_helpers.max2(tmp95, 1)[:, None]
tmp98 = tmp94 - tmp97
tmp99 = tl_math.exp(tmp98)
tmp100 = tl.broadcast_to(tmp99, [XBLOCK, RBLOCK])
tmp102 = tl.sum(tmp100, 1)[:, None]
tmp104 = tmp1 * tmp103
tmp105 = tmp104 + tmp5
tmp106 = tl.broadcast_to(tmp105, [XBLOCK, RBLOCK])
tmp108 = triton_helpers.max2(tmp106, 1)[:, None]
tmp109 = tmp105 - tmp108
tmp110 = tl_math.exp(tmp109)
tmp111 = tl.broadcast_to(tmp110, [XBLOCK, RBLOCK])
tmp113 = tl.sum(tmp111, 1)[:, None]
tmp115 = tmp1 * tmp114
tmp116 = tmp115 + tmp5
tmp117 = tl.broadcast_to(tmp116, [XBLOCK, RBLOCK])
tmp119 = triton_helpers.max2(tmp117, 1)[:, None]
tmp120 = tmp116 - tmp119
tmp121 = tl_math.exp(tmp120)
tmp122 = tl.broadcast_to(tmp121, [XBLOCK, RBLOCK])
tmp124 = tl.sum(tmp122, 1)[:, None]
tmp126 = tmp1 * tmp125
tmp127 = tmp126 + tmp5
tmp128 = tl.broadcast_to(tmp127, [XBLOCK, RBLOCK])
tmp130 = triton_helpers.max2(tmp128, 1)[:, None]
tmp131 = tmp127 - tmp130
tmp132 = tl_math.exp(tmp131)
tmp133 = tl.broadcast_to(tmp132, [XBLOCK, RBLOCK])
tmp135 = tl.sum(tmp133, 1)[:, None]
tmp137 = tmp1 * tmp136
tmp138 = tmp137 + tmp5
tmp139 = tl.broadcast_to(tmp138, [XBLOCK, RBLOCK])
tmp141 = triton_helpers.max2(tmp139, 1)[:, None]
tmp142 = tmp138 - tmp141
tmp143 = tl_math.exp(tmp142)
tmp144 = tl.broadcast_to(tmp143, [XBLOCK, RBLOCK])
tmp146 = tl.sum(tmp144, 1)[:, None]
tmp148 = tmp1 * tmp147
tmp149 = tmp148 + tmp5
tmp150 = tl.broadcast_to(tmp149, [XBLOCK, RBLOCK])
tmp152 = triton_helpers.max2(tmp150, 1)[:, None]
tmp153 = tmp149 - tmp152
tmp154 = tl_math.exp(tmp153)
tmp155 = tl.broadcast_to(tmp154, [XBLOCK, RBLOCK])
tmp157 = tl.sum(tmp155, 1)[:, None]
tmp159 = tmp1 * tmp158
tmp160 = tmp159 + tmp5
tmp161 = tl.broadcast_to(tmp160, [XBLOCK, RBLOCK])
tmp163 = triton_helpers.max2(tmp161, 1)[:, None]
tmp164 = tmp160 - tmp163
tmp165 = tl_math.exp(tmp164)
tmp166 = tl.broadcast_to(tmp165, [XBLOCK, RBLOCK])
tmp168 = tl.sum(tmp166, 1)[:, None]
tmp170 = tmp1 * tmp169
tmp171 = tmp170 + tmp5
tmp172 = tl.broadcast_to(tmp171, [XBLOCK, RBLOCK])
tmp174 = triton_helpers.max2(tmp172, 1)[:, None]
tmp175 = tmp171 - tmp174
tmp176 = tl_math.exp(tmp175)
tmp177 = tl.broadcast_to(tmp176, [XBLOCK, RBLOCK])
tmp179 = tl.sum(tmp177, 1)[:, None]
tmp180 = r0
tmp181 = tl.full([1, 1], 0, tl.int64)
tmp182 = tmp180 >= tmp181
tmp183 = tl.full([1, 1], 1, tl.int64)
tmp184 = tmp180 < tmp183
tmp187 = tmp1 * tmp186
tmp188 = tmp187 + tmp5
tmp189 = tmp188 - tmp9
tmp190 = tl_math.log(tmp14)
tmp191 = tmp189 - tmp190
tmp192 = -tmp191
tmp193 = tl.full(tmp192.shape, 0.0, tmp192.dtype)
tmp194 = tl.where(tmp184, tmp192, tmp193)
tmp195 = tmp180 >= tmp183
tmp196 = tl.full([1, 1], 2, tl.int64)
tmp197 = tmp180 < tmp196
tmp198 = tmp195 & tmp197
tmp201 = tmp1 * tmp200
tmp202 = tmp201 + tmp5
tmp203 = tmp202 - tmp20
tmp204 = tl_math.log(tmp25)
tmp205 = tmp203 - tmp204
tmp206 = -tmp205
tmp207 = tl.full(tmp206.shape, 0.0, tmp206.dtype)
tmp208 = tl.where(tmp198, tmp206, tmp207)
tmp209 = tmp180 >= tmp196
tmp210 = tl.full([1, 1], 3, tl.int64)
tmp211 = tmp180 < tmp210
tmp212 = tmp209 & tmp211
tmp215 = tmp1 * tmp214
tmp216 = tmp215 + tmp5
tmp217 = tmp216 - tmp31
tmp218 = tl_math.log(tmp36)
tmp219 = tmp217 - tmp218
tmp220 = -tmp219
tmp221 = tl.full(tmp220.shape, 0.0, tmp220.dtype)
tmp222 = tl.where(tmp212, tmp220, tmp221)
tmp223 = tmp180 >= tmp210
tmp224 = tl.full([1, 1], 4, tl.int64)
tmp225 = tmp180 < tmp224
tmp228 = tmp1 * tmp227
tmp229 = tmp228 + tmp5
tmp230 = tmp229 - tmp42
tmp231 = tl_math.log(tmp47)
tmp232 = tmp230 - tmp231
tmp233 = -tmp232
tmp234 = tl.full(tmp233.shape, 0.0, tmp233.dtype)
tmp235 = tl.where(tmp223, tmp233, tmp234)
tmp236 = tl.where(tmp212, tmp222, tmp235)
tmp237 = tl.where(tmp198, tmp208, tmp236)
tmp238 = tl.where(tmp184, tmp194, tmp237)
tmp241 = tmp1 * tmp240
tmp242 = tmp241 + tmp5
tmp243 = tmp242 - tmp53
tmp244 = tl_math.log(tmp58)
tmp245 = tmp243 - tmp244
tmp246 = -tmp245
tmp247 = tl.full(tmp246.shape, 0.0, tmp246.dtype)
tmp248 = tl.where(tmp184, tmp246, tmp247)
tmp251 = tmp1 * tmp250
tmp252 = tmp251 + tmp5
tmp253 = tmp252 - tmp64
tmp254 = tl_math.log(tmp69)
tmp255 = tmp253 - tmp254
tmp256 = -tmp255
tmp257 = tl.full(tmp256.shape, 0.0, tmp256.dtype)
tmp258 = tl.where(tmp198, tmp256, tmp257)
tmp261 = tmp1 * tmp260
tmp262 = tmp261 + tmp5
tmp263 = tmp262 - tmp75
tmp264 = tl_math.log(tmp80)
tmp265 = tmp263 - tmp264
tmp266 = -tmp265
tmp267 = tl.full(tmp266.shape, 0.0, tmp266.dtype)
tmp268 = tl.where(tmp212, tmp266, tmp267)
tmp271 = tmp1 * tmp270
tmp272 = tmp271 + tmp5
tmp273 = tmp272 - tmp86
tmp274 = tl_math.log(tmp91)
tmp275 = tmp273 - tmp274
tmp276 = -tmp275
tmp277 = tl.full(tmp276.shape, 0.0, tmp276.dtype)
tmp278 = tl.where(tmp223, tmp276, tmp277)
tmp279 = tl.where(tmp212, tmp268, tmp278)
tmp280 = tl.where(tmp198, tmp258, tmp279)
tmp281 = tl.where(tmp184, tmp248, tmp280)
tmp284 = tmp1 * tmp283
tmp285 = tmp284 + tmp5
tmp286 = tmp285 - tmp97
tmp287 = tl_math.log(tmp102)
tmp288 = tmp286 - tmp287
tmp289 = -tmp288
tmp290 = tl.full(tmp289.shape, 0.0, tmp289.dtype)
tmp291 = tl.where(tmp184, tmp289, tmp290)
tmp294 = tmp1 * tmp293
tmp295 = tmp294 + tmp5
tmp296 = tmp295 - tmp108
tmp297 = tl_math.log(tmp113)
tmp298 = tmp296 - tmp297
tmp299 = -tmp298
tmp300 = tl.full(tmp299.shape, 0.0, tmp299.dtype)
tmp301 = tl.where(tmp198, tmp299, tmp300)
tmp304 = tmp1 * tmp303
tmp305 = tmp304 + tmp5
tmp306 = tmp305 - tmp119
tmp307 = tl_math.log(tmp124)
tmp308 = tmp306 - tmp307
tmp309 = -tmp308
tmp310 = tl.full(tmp309.shape, 0.0, tmp309.dtype)
tmp311 = tl.where(tmp212, tmp309, tmp310)
tmp314 = tmp1 * tmp313
tmp315 = tmp314 + tmp5
tmp316 = tmp315 - tmp130
tmp317 = tl_math.log(tmp135)
tmp318 = tmp316 - tmp317
tmp319 = -tmp318
tmp320 = tl.full(tmp319.shape, 0.0, tmp319.dtype)
tmp321 = tl.where(tmp223, tmp319, tmp320)
tmp322 = tl.where(tmp212, tmp311, tmp321)
tmp323 = tl.where(tmp198, tmp301, tmp322)
tmp324 = tl.where(tmp184, tmp291, tmp323)
tmp325 = tl_math.log(tmp146)
tmp326 = tmp142 - tmp325
tmp327 = tl_math.exp(tmp326)
tmp328 = tl_math.log(tmp157)
tmp329 = tmp153 - tmp328
tmp330 = tl_math.exp(tmp329)
tmp331 = tl_math.log(tmp168)
tmp332 = tmp164 - tmp331
tmp333 = tl_math.exp(tmp332)
tmp334 = tl_math.log(tmp179)
tmp335 = tmp175 - tmp334
tmp336 = tl_math.exp(tmp335)
tl.store(out_ptr32 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp238, None)
tl.store(out_ptr33 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp281, None)
tl.store(out_ptr34 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp324, None)
tl.store(out_ptr35 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp326, None)
tl.store(out_ptr36 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp327, None)
tl.store(out_ptr37 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp329, None)
tl.store(out_ptr38 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp330, None)
tl.store(out_ptr39 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp332, None)
tl.store(out_ptr40 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp333, None)
tl.store(out_ptr41 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp335, None)
tl.store(out_ptr42 + (tl.broadcast_to(r0, [XBLOCK, RBLOCK])), tmp336, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/p3/cp3hem6ijvssri53og7f2pda7nauqlpvhy5sv4mndarqqqeb6qmi.py
# Topologically Sorted Source Nodes: [L_row_3], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# L_row_3 => cat_32
# Graph fragment:
# %cat_32 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%unsqueeze_28, %unsqueeze_29, %unsqueeze_30, %unsqueeze_31],), kwargs = {})
triton_poi_fused_stack_15 = async_compile.triton('triton_poi_fused_stack_15', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_15', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_stack_15(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp5 = tl.load(in_ptr0 + (3))
tmp6 = tl.broadcast_to(tmp5, [XBLOCK])
tmp14 = tl.load(in_ptr1 + (3))
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tmp23 = tl.load(in_ptr2 + (3))
tmp24 = tl.broadcast_to(tmp23, [XBLOCK])
tmp31 = tl.load(in_ptr3 + (3))
tmp32 = tl.broadcast_to(tmp31, [XBLOCK])
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp7 = -tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 2, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tmp10 & tmp12
tmp16 = -tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp13, tmp16, tmp17)
tmp19 = tmp0 >= tmp11
tmp20 = tl.full([1], 3, tl.int64)
tmp21 = tmp0 < tmp20
tmp22 = tmp19 & tmp21
tmp25 = -tmp24
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp22, tmp25, tmp26)
tmp28 = tmp0 >= tmp20
tmp29 = tl.full([1], 4, tl.int64)
tmp30 = tmp0 < tmp29
tmp33 = -tmp32
tmp34 = tl.full(tmp33.shape, 0.0, tmp33.dtype)
tmp35 = tl.where(tmp28, tmp33, tmp34)
tmp36 = tl.where(tmp22, tmp27, tmp35)
tmp37 = tl.where(tmp13, tmp18, tmp36)
tmp38 = tl.where(tmp4, tmp9, tmp37)
tl.store(out_ptr0 + (x0), tmp38, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/it/citbft6scgowtafcadbm67lvmjscghqookf2irszk6yfd3lrovgj.py
# Topologically Sorted Source Nodes: [mean_17], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# mean_17 => mean_17
# Graph fragment:
# %mean_17 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%view_17,), kwargs = {})
triton_per_fused_mean_16 = async_compile.triton('triton_per_fused_mean_16', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_16', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_16(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tmp4 = 16.0
tmp5 = tmp3 / tmp4
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp5, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (), ())
assert_size_stride(primals_3, (), ())
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf17 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf21 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf38 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf42 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf59 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf63 = empty_strided_cuda((16, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [new_centroids, new_centroids_3, new_centroids_4, new_centroids_7, new_centroids_8, new_centroids_11, new_centroids_12], Original ATen: [aten.stack]
stream0 = get_raw_stream(0)
triton_poi_fused_stack_0.run(primals_1, buf0, buf17, buf21, buf38, buf42, buf59, buf63, 16, grid=grid(16), stream=stream0)
buf1 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mm], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((), (), torch.float32)
buf7 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf3 = reinterpret_tensor(buf7, (4, ), (1, ), 0) # alias
# Topologically Sorted Source Nodes: [norm, excl_2, excl_3], Original ATen: [aten.linalg_vector_norm, aten.cat, aten.mean]
triton_per_fused_cat_linalg_vector_norm_mean_1.run(primals_1, buf2, buf3, 1, 4, grid=grid(1), stream=stream0)
buf23 = empty_strided_cuda((), (), torch.float32)
buf4 = reinterpret_tensor(buf7, (4, ), (1, ), 4) # alias
buf14 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf11 = reinterpret_tensor(buf14, (4, ), (1, ), 4) # alias
buf28 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf24 = reinterpret_tensor(buf28, (4, ), (1, ), 4) # alias
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, norm_8, excl_10, excl_11], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
triton_per_fused_cat_linalg_vector_norm_mean_stack_2.run(primals_1, buf23, buf4, buf11, buf24, 1, 4, grid=grid(1), stream=stream0)
buf44 = empty_strided_cuda((), (), torch.float32)
buf5 = reinterpret_tensor(buf7, (4, ), (1, ), 8) # alias
buf12 = reinterpret_tensor(buf14, (4, ), (1, ), 8) # alias
buf26 = reinterpret_tensor(buf28, (4, ), (1, ), 8) # alias
buf35 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf33 = reinterpret_tensor(buf35, (4, ), (1, ), 8) # alias
buf49 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf45 = reinterpret_tensor(buf49, (4, ), (1, ), 8) # alias
buf51 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, new_centroids_5, new_centroids_6, norm_16, excl_18, excl_19, norm_18], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
triton_per_fused_cat_linalg_vector_norm_mean_stack_3.run(primals_1, buf44, buf5, buf12, buf26, buf33, buf45, buf51, 1, 4, grid=grid(1), stream=stream0)
buf65 = empty_strided_cuda((), (), torch.float32)
buf6 = reinterpret_tensor(buf7, (4, ), (1, ), 12) # alias
buf13 = reinterpret_tensor(buf14, (4, ), (1, ), 12) # alias
buf27 = reinterpret_tensor(buf28, (4, ), (1, ), 12) # alias
buf34 = reinterpret_tensor(buf35, (4, ), (1, ), 12) # alias
buf48 = reinterpret_tensor(buf49, (4, ), (1, ), 12) # alias
buf56 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf55 = reinterpret_tensor(buf56, (4, ), (1, ), 12) # alias
buf70 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf66 = reinterpret_tensor(buf70, (4, ), (1, ), 12) # alias
# Topologically Sorted Source Nodes: [new_centroids_1, new_centroids_2, new_centroids_5, new_centroids_6, new_centroids_9, new_centroids_10, norm_24, excl_26, excl_27], Original ATen: [aten.stack, aten.linalg_vector_norm, aten.cat, aten.mean]
triton_per_fused_cat_linalg_vector_norm_mean_stack_4.run(primals_1, buf65, buf6, buf13, buf27, buf34, buf48, buf55, buf66, 1, 4, grid=grid(1), stream=stream0)
del buf3
del buf4
del buf5
del buf6
buf8 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mm_1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 4), reinterpret_tensor(buf7, (4, 4), (1, 4), 0), out=buf8)
buf9 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [norm_2], Original ATen: [aten.linalg_vector_norm]
triton_per_fused_linalg_vector_norm_5.run(primals_1, buf9, 1, 4, grid=grid(1), stream=stream0)
buf16 = empty_strided_cuda((), (), torch.float32)
buf19 = empty_strided_cuda((), (), torch.float32)
buf10 = reinterpret_tensor(buf14, (4, ), (1, ), 0) # alias
buf25 = reinterpret_tensor(buf28, (4, ), (1, ), 0) # alias
buf32 = reinterpret_tensor(buf35, (4, ), (1, ), 0) # alias
buf46 = reinterpret_tensor(buf49, (4, ), (1, ), 0) # alias
buf53 = reinterpret_tensor(buf56, (4, ), (1, ), 0) # alias
buf67 = reinterpret_tensor(buf70, (4, ), (1, ), 0) # alias
buf77 = empty_strided_cuda((16, ), (1, ), torch.float32)
buf74 = reinterpret_tensor(buf77, (4, ), (1, ), 0) # alias
# Topologically Sorted Source Nodes: [excl_4, excl_5, norm_4, norm_6, new_centroids_5, new_centroids_6, new_centroids_9, new_centroids_10, new_centroids_13, new_centroids_14], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm, aten.stack]
triton_per_fused_cat_linalg_vector_norm_mean_stack_6.run(primals_1, buf16, buf19, buf10, buf25, buf32, buf46, buf53, buf67, buf74, 1, 4, grid=grid(1), stream=stream0)
del buf10
del buf11
del buf12
del buf13
buf15 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mm_2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 8), reinterpret_tensor(buf14, (4, 4), (1, 4), 0), out=buf15)
buf18 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mm_3], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 12), reinterpret_tensor(buf17, (4, 4), (1, 4), 0), out=buf18)
buf84 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
buf20 = reinterpret_tensor(buf84, (4, 4), (4, 1), 0) # alias
# Topologically Sorted Source Nodes: [cs_row], Original ATen: [aten.cat]
triton_poi_fused_cat_7.run(buf1, buf2, buf0, buf8, buf9, buf7, buf15, buf16, buf14, buf18, buf19, buf17, buf20, 16, grid=grid(16), stream=stream0)
del buf0
del buf14
del buf16
del buf17
del buf7
buf22 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [mm_4], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 16), reinterpret_tensor(buf21, (4, 4), (1, 4), 0), out=buf22)
del buf24
del buf25
del buf26
del buf27
buf29 = buf18; del buf18 # reuse
# Topologically Sorted Source Nodes: [mm_5], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 20), reinterpret_tensor(buf28, (4, 4), (1, 4), 0), out=buf29)
buf30 = buf9; del buf9 # reuse
buf37 = buf2; del buf2 # reuse
buf40 = buf19; del buf19 # reuse
buf31 = reinterpret_tensor(buf35, (4, ), (1, ), 4) # alias
buf47 = reinterpret_tensor(buf49, (4, ), (1, ), 4) # alias
buf54 = reinterpret_tensor(buf56, (4, ), (1, ), 4) # alias
buf68 = reinterpret_tensor(buf70, (4, ), (1, ), 4) # alias
buf75 = reinterpret_tensor(buf77, (4, ), (1, ), 4) # alias
# Topologically Sorted Source Nodes: [norm_10, excl_12, excl_13, norm_12, norm_14, new_centroids_9, new_centroids_10, new_centroids_13, new_centroids_14], Original ATen: [aten.linalg_vector_norm, aten.cat, aten.mean, aten.stack]
triton_per_fused_cat_linalg_vector_norm_mean_stack_8.run(primals_1, buf30, buf37, buf40, buf31, buf47, buf54, buf68, buf75, 1, 4, grid=grid(1), stream=stream0)
del buf31
del buf32
del buf33
del buf34
buf36 = buf15; del buf15 # reuse
# Topologically Sorted Source Nodes: [mm_6], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 24), reinterpret_tensor(buf35, (4, 4), (1, 4), 0), out=buf36)
buf39 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [mm_7], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 28), reinterpret_tensor(buf38, (4, 4), (1, 4), 0), out=buf39)
buf41 = reinterpret_tensor(buf84, (4, 4), (4, 1), 16) # alias
# Topologically Sorted Source Nodes: [cs_row_1], Original ATen: [aten.cat]
triton_poi_fused_cat_7.run(buf22, buf23, buf21, buf29, buf30, buf28, buf36, buf37, buf35, buf39, buf40, buf38, buf41, 16, grid=grid(16), stream=stream0)
del buf21
del buf23
del buf28
del buf30
del buf35
del buf38
buf43 = buf39; del buf39 # reuse
# Topologically Sorted Source Nodes: [mm_8], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 32), reinterpret_tensor(buf42, (4, 4), (1, 4), 0), out=buf43)
del buf45
del buf46
del buf47
del buf48
buf50 = buf36; del buf36 # reuse
# Topologically Sorted Source Nodes: [mm_9], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 36), reinterpret_tensor(buf49, (4, 4), (1, 4), 0), out=buf50)
buf58 = buf40; del buf40 # reuse
buf61 = buf37; del buf37 # reuse
buf52 = reinterpret_tensor(buf56, (4, ), (1, ), 8) # alias
buf69 = reinterpret_tensor(buf70, (4, ), (1, ), 8) # alias
buf76 = reinterpret_tensor(buf77, (4, ), (1, ), 8) # alias
# Topologically Sorted Source Nodes: [excl_20, excl_21, norm_20, norm_22, new_centroids_13, new_centroids_14], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm, aten.stack]
triton_per_fused_cat_linalg_vector_norm_mean_stack_9.run(primals_1, buf58, buf61, buf52, buf69, buf76, 1, 4, grid=grid(1), stream=stream0)
del buf52
del buf53
del buf54
del buf55
buf57 = buf29; del buf29 # reuse
# Topologically Sorted Source Nodes: [mm_10], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 40), reinterpret_tensor(buf56, (4, 4), (1, 4), 0), out=buf57)
buf60 = buf22; del buf22 # reuse
# Topologically Sorted Source Nodes: [mm_11], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 44), reinterpret_tensor(buf59, (4, 4), (1, 4), 0), out=buf60)
buf62 = reinterpret_tensor(buf84, (4, 4), (4, 1), 32) # alias
# Topologically Sorted Source Nodes: [cs_row_2], Original ATen: [aten.cat]
triton_poi_fused_cat_7.run(buf43, buf44, buf42, buf50, buf51, buf49, buf57, buf58, buf56, buf60, buf61, buf59, buf62, 16, grid=grid(16), stream=stream0)
del buf42
del buf44
del buf49
del buf56
buf64 = buf60; del buf60 # reuse
# Topologically Sorted Source Nodes: [mm_12], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 48), reinterpret_tensor(buf63, (4, 4), (1, 4), 0), out=buf64)
del buf66
del buf67
del buf68
del buf69
buf71 = buf57; del buf57 # reuse
# Topologically Sorted Source Nodes: [mm_13], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 52), reinterpret_tensor(buf70, (4, 4), (1, 4), 0), out=buf71)
buf72 = buf61; del buf61 # reuse
# Topologically Sorted Source Nodes: [norm_26], Original ATen: [aten.linalg_vector_norm]
triton_per_fused_linalg_vector_norm_10.run(primals_1, buf72, 1, 4, grid=grid(1), stream=stream0)
buf73 = reinterpret_tensor(buf77, (4, ), (1, ), 12) # alias
buf82 = buf58; del buf58 # reuse
# Topologically Sorted Source Nodes: [excl_28, excl_29, norm_30], Original ATen: [aten.cat, aten.mean, aten.linalg_vector_norm]
triton_per_fused_cat_linalg_vector_norm_mean_11.run(primals_1, buf73, buf82, 1, 4, grid=grid(1), stream=stream0)
del buf73
del buf74
del buf75
del buf76
buf78 = buf50; del buf50 # reuse
# Topologically Sorted Source Nodes: [mm_14], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 56), reinterpret_tensor(buf77, (4, 4), (1, 4), 0), out=buf78)
buf79 = buf51; del buf51 # reuse
# Topologically Sorted Source Nodes: [norm_28], Original ATen: [aten.linalg_vector_norm]
triton_per_fused_linalg_vector_norm_12.run(primals_1, buf79, 1, 4, grid=grid(1), stream=stream0)
buf80 = buf59; del buf59 # reuse
# Topologically Sorted Source Nodes: [new_centroids_15], Original ATen: [aten.stack]
triton_poi_fused_stack_13.run(primals_1, buf80, 16, grid=grid(16), stream=stream0)
buf81 = buf43; del buf43 # reuse
# Topologically Sorted Source Nodes: [mm_15], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 60), reinterpret_tensor(buf80, (4, 4), (1, 4), 0), out=buf81)
del primals_1
buf83 = reinterpret_tensor(buf84, (4, 4), (4, 1), 48) # alias
# Topologically Sorted Source Nodes: [cs_row_3], Original ATen: [aten.cat]
triton_poi_fused_cat_7.run(buf64, buf65, buf63, buf71, buf72, buf70, buf78, buf79, buf77, buf81, buf82, buf80, buf83, 16, grid=grid(16), stream=stream0)
del buf63
del buf65
del buf70
del buf72
del buf77
del buf79
buf125 = buf80; del buf80 # reuse
buf93 = reinterpret_tensor(buf125, (4, ), (1, ), 0) # alias
buf102 = reinterpret_tensor(buf125, (4, ), (1, ), 4) # alias
buf111 = reinterpret_tensor(buf125, (4, ), (1, ), 8) # alias
buf114 = reinterpret_tensor(buf81, (4, ), (1, ), 0); del buf81 # reuse
buf130 = reinterpret_tensor(buf78, (4, ), (1, ), 0); del buf78 # reuse
buf117 = reinterpret_tensor(buf71, (4, ), (1, ), 0); del buf71 # reuse
buf129 = reinterpret_tensor(buf64, (4, ), (1, ), 0); del buf64 # reuse
buf120 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf128 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf123 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf127 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax, log_softmax_1, log_softmax_2, log_softmax_3, L_row, log_softmax_4, log_softmax_5, log_softmax_6, log_softmax_7, L_row_1, log_softmax_8, log_softmax_9, log_softmax_10, log_softmax_11, L_row_2, log_softmax_12, log_softmax_13, log_softmax_14, log_softmax_15], Original ATen: [aten._log_softmax, aten.stack, aten._log_softmax_backward_data]
triton_per_fused__log_softmax__log_softmax_backward_data_stack_14.run(primals_2, buf84, primals_3, buf93, buf102, buf111, buf114, buf130, buf117, buf129, buf120, buf128, buf123, buf127, 1, 4, grid=grid(1), stream=stream0)
buf124 = reinterpret_tensor(buf125, (4, ), (1, ), 12) # alias
# Topologically Sorted Source Nodes: [L_row_3], Original ATen: [aten.stack]
triton_poi_fused_stack_15.run(buf114, buf117, buf120, buf123, buf124, 4, grid=grid(4), stream=stream0)
del buf114
del buf117
del buf120
del buf123
buf126 = buf82; del buf82 # reuse
buf131 = buf126; del buf126 # reuse
# Topologically Sorted Source Nodes: [mean_17], Original ATen: [aten.mean]
triton_per_fused_mean_16.run(buf131, buf125, 1, 16, grid=grid(1), stream=stream0)
del buf102
del buf111
del buf124
del buf125
del buf93
return (buf131, primals_2, primals_3, buf84, buf127, buf128, buf129, buf130, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((), (), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((), (), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.functional import F
from torch.nn import functional as F
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
class GE2ELoss(nn.Module):
def __init__(self, init_w=10.0, init_b=-5.0, loss_method='softmax'):
"""
Implementation of the Generalized End-to-End loss defined in https://arxiv.org/abs/1710.10467 [1]
Accepts an input of size (N, M, D)
where N is the number of speakers in the batch,
M is the number of utterances per speaker,
and D is the dimensionality of the embedding vector (e.g. d-vector)
Args:
- init_w (float): defines the initial value of w in Equation (5) of [1]
- init_b (float): definies the initial value of b in Equation (5) of [1]
"""
super(GE2ELoss, self).__init__()
self.w = nn.Parameter(torch.tensor(init_w))
self.b = nn.Parameter(torch.tensor(init_b))
self.loss_method = loss_method
assert self.loss_method in ['softmax', 'contrast']
if self.loss_method == 'softmax':
self.embed_loss = self.embed_loss_softmax
if self.loss_method == 'contrast':
self.embed_loss = self.embed_loss_contrast
def calc_new_centroids(self, dvecs, centroids, spkr, utt):
"""
Calculates the new centroids excluding the reference utterance
"""
excl = torch.cat((dvecs[spkr, :utt], dvecs[spkr, utt + 1:]))
excl = torch.mean(excl, 0)
new_centroids = []
for i, centroid in enumerate(centroids):
if i == spkr:
new_centroids.append(excl)
else:
new_centroids.append(centroid)
return torch.stack(new_centroids)
def calc_cosine_sim(self, dvecs, centroids):
"""
Make the cosine similarity matrix with dims (N,M,N)
"""
cos_sim_matrix = []
for spkr_idx, speaker in enumerate(dvecs):
cs_row = []
for utt_idx, utterance in enumerate(speaker):
new_centroids = self.calc_new_centroids(dvecs, centroids,
spkr_idx, utt_idx)
cs_row.append(torch.clamp(torch.mm(utterance.unsqueeze(1).
transpose(0, 1), new_centroids.transpose(0, 1)) / (
torch.norm(utterance) * torch.norm(new_centroids, dim=1
)), 1e-06))
cs_row = torch.cat(cs_row, dim=0)
cos_sim_matrix.append(cs_row)
return torch.stack(cos_sim_matrix)
def embed_loss_softmax(self, dvecs, cos_sim_matrix):
"""
Calculates the loss on each embedding $L(e_{ji})$ by taking softmax
"""
N, M, _ = dvecs.shape
L = []
for j in range(N):
L_row = []
for i in range(M):
L_row.append(-F.log_softmax(cos_sim_matrix[j, i], 0)[j])
L_row = torch.stack(L_row)
L.append(L_row)
return torch.stack(L)
def embed_loss_contrast(self, dvecs, cos_sim_matrix):
"""
Calculates the loss on each embedding $L(e_{ji})$ by contrast loss with closest centroid
"""
N, M, _ = dvecs.shape
L = []
for j in range(N):
L_row = []
for i in range(M):
centroids_sigmoids = torch.sigmoid(cos_sim_matrix[j, i])
excl_centroids_sigmoids = torch.cat((centroids_sigmoids[:j],
centroids_sigmoids[j + 1:]))
L_row.append(1.0 - torch.sigmoid(cos_sim_matrix[j, i, j]) +
torch.max(excl_centroids_sigmoids))
L_row = torch.stack(L_row)
L.append(L_row)
return torch.stack(L)
def forward(self, dvecs):
"""
Calculates the GE2E loss for an input of dimensions (num_speakers, num_utts_per_speaker, dvec_feats)
"""
centroids = torch.mean(dvecs, 1)
cos_sim_matrix = self.calc_cosine_sim(dvecs, centroids)
torch.clamp(self.w, 1e-06)
cos_sim_matrix = self.w * cos_sim_matrix + self.b
L = self.embed_loss(dvecs, cos_sim_matrix)
return L.mean()
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import torch.nn as nn
import torch.nn.functional as F
from torch.functional import F
from torch.nn import functional as F
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.distributions
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_stack_0(in_ptr0, out_ptr0, out_ptr1, out_ptr2,
out_ptr3, out_ptr4, out_ptr5, out_ptr6, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (8 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (12 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = 3.0
tmp11 = tmp9 / tmp10
tmp12 = tl.full(tmp11.shape, 0.0, tmp11.dtype)
tmp13 = tl.where(tmp4, tmp11, tmp12)
tmp14 = tmp0 >= tmp3
tmp15 = tl.full([1], 8, tl.int64)
tmp16 = tmp0 < tmp15
tmp17 = tmp14 & tmp16
tmp18 = tl.load(in_ptr0 + (16 + (-4 + x0)), tmp17 & xmask,
eviction_policy='evict_last', other=0.0)
tmp19 = tl.load(in_ptr0 + (20 + (-4 + x0)), tmp17 & xmask,
eviction_policy='evict_last', other=0.0)
tmp20 = tmp18 + tmp19
tmp21 = tl.load(in_ptr0 + (24 + (-4 + x0)), tmp17 & xmask,
eviction_policy='evict_last', other=0.0)
tmp22 = tmp20 + tmp21
tmp23 = tl.load(in_ptr0 + (28 + (-4 + x0)), tmp17 & xmask,
eviction_policy='evict_last', other=0.0)
tmp24 = tmp22 + tmp23
tmp25 = 4.0
tmp26 = tmp24 / tmp25
tmp27 = tl.full(tmp26.shape, 0.0, tmp26.dtype)
tmp28 = tl.where(tmp17, tmp26, tmp27)
tmp29 = tmp0 >= tmp15
tmp30 = tl.full([1], 12, tl.int64)
tmp31 = tmp0 < tmp30
tmp32 = tmp29 & tmp31
tmp33 = tl.load(in_ptr0 + (32 + (-8 + x0)), tmp32 & xmask,
eviction_policy='evict_last', other=0.0)
tmp34 = tl.load(in_ptr0 + (36 + (-8 + x0)), tmp32 & xmask,
eviction_policy='evict_last', other=0.0)
tmp35 = tmp33 + tmp34
tmp36 = tl.load(in_ptr0 + (40 + (-8 + x0)), tmp32 & xmask,
eviction_policy='evict_last', other=0.0)
tmp37 = tmp35 + tmp36
tmp38 = tl.load(in_ptr0 + (44 + (-8 + x0)), tmp32 & xmask,
eviction_policy='evict_last', other=0.0)
tmp39 = tmp37 + tmp38
tmp40 = tmp39 / tmp25
tmp41 = tl.full(tmp40.shape, 0.0, tmp40.dtype)
tmp42 = tl.where(tmp32, tmp40, tmp41)
tmp43 = tmp0 >= tmp30
tl.full([1], 16, tl.int64)
tmp46 = tl.load(in_ptr0 + (48 + (-12 + x0)), tmp43 & xmask,
eviction_policy='evict_last', other=0.0)
tmp47 = tl.load(in_ptr0 + (52 + (-12 + x0)), tmp43 & xmask,
eviction_policy='evict_last', other=0.0)
tmp48 = tmp46 + tmp47
tmp49 = tl.load(in_ptr0 + (56 + (-12 + x0)), tmp43 & xmask,
eviction_policy='evict_last', other=0.0)
tmp50 = tmp48 + tmp49
tmp51 = tl.load(in_ptr0 + (60 + (-12 + x0)), tmp43 & xmask,
eviction_policy='evict_last', other=0.0)
tmp52 = tmp50 + tmp51
tmp53 = tmp52 / tmp25
tmp54 = tl.full(tmp53.shape, 0.0, tmp53.dtype)
tmp55 = tl.where(tmp43, tmp53, tmp54)
tmp56 = tl.where(tmp32, tmp42, tmp55)
tmp57 = tl.where(tmp17, tmp28, tmp56)
tmp58 = tl.where(tmp4, tmp13, tmp57)
tmp59 = tl.load(in_ptr0 + x0, tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp60 = tmp59 + tmp5
tmp61 = tmp60 + tmp6
tmp62 = tmp61 / tmp10
tmp63 = tl.full(tmp62.shape, 0.0, tmp62.dtype)
tmp64 = tl.where(tmp4, tmp62, tmp63)
tmp65 = tl.where(tmp4, tmp64, tmp57)
tmp66 = tmp61 + tmp8
tmp67 = tmp66 / tmp25
tmp68 = tl.full(tmp67.shape, 0.0, tmp67.dtype)
tmp69 = tl.where(tmp4, tmp67, tmp68)
tmp70 = tmp19 + tmp21
tmp71 = tmp70 + tmp23
tmp72 = tmp71 / tmp10
tmp73 = tl.full(tmp72.shape, 0.0, tmp72.dtype)
tmp74 = tl.where(tmp17, tmp72, tmp73)
tmp75 = tl.where(tmp17, tmp74, tmp56)
tmp76 = tl.where(tmp4, tmp69, tmp75)
tmp77 = tmp22 / tmp10
tmp78 = tl.full(tmp77.shape, 0.0, tmp77.dtype)
tmp79 = tl.where(tmp17, tmp77, tmp78)
tmp80 = tl.where(tmp17, tmp79, tmp56)
tmp81 = tl.where(tmp4, tmp69, tmp80)
tmp82 = tmp34 + tmp36
tmp83 = tmp82 + tmp38
tmp84 = tmp83 / tmp10
tmp85 = tl.full(tmp84.shape, 0.0, tmp84.dtype)
tmp86 = tl.where(tmp32, tmp84, tmp85)
tmp87 = tl.where(tmp32, tmp86, tmp55)
tmp88 = tl.where(tmp17, tmp28, tmp87)
tmp89 = tl.where(tmp4, tmp69, tmp88)
tmp90 = tmp37 / tmp10
tmp91 = tl.full(tmp90.shape, 0.0, tmp90.dtype)
tmp92 = tl.where(tmp32, tmp90, tmp91)
tmp93 = tl.where(tmp32, tmp92, tmp55)
tmp94 = tl.where(tmp17, tmp28, tmp93)
tmp95 = tl.where(tmp4, tmp69, tmp94)
tmp96 = tmp47 + tmp49
tmp97 = tmp96 + tmp51
tmp98 = tmp97 / tmp10
tmp99 = tl.full(tmp98.shape, 0.0, tmp98.dtype)
tmp100 = tl.where(tmp43, tmp98, tmp99)
tmp101 = tl.where(tmp32, tmp42, tmp100)
tmp102 = tl.where(tmp17, tmp28, tmp101)
tmp103 = tl.where(tmp4, tmp69, tmp102)
tl.store(out_ptr0 + x0, tmp58, xmask)
tl.store(out_ptr1 + x0, tmp65, xmask)
tl.store(out_ptr2 + x0, tmp76, xmask)
tl.store(out_ptr3 + x0, tmp81, xmask)
tl.store(out_ptr4 + x0, tmp89, xmask)
tl.store(out_ptr5 + x0, tmp95, xmask)
tl.store(out_ptr6 + x0, tmp103, xmask)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_1(in_ptr0, out_ptr0,
out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp5 = tl.full([1, 1], 0, tl.int64)
tmp7 = tl.full([1, 1], 1, tl.int64)
tmp8 = tmp5 < tmp7
tmp9 = tl.load(in_ptr0 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp8,
other=0.0)
tmp10 = tmp5 >= tmp7
tl.full([1, 1], 3, tl.int64)
tmp13 = tl.load(in_ptr0 + tl.broadcast_to(8 + r0 + 4 * -1, [XBLOCK,
RBLOCK]), tmp10, other=0.0)
tmp14 = tl.where(tmp8, tmp9, tmp13)
tmp16 = tmp7 < tmp7
tmp17 = tl.load(in_ptr0 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp16,
other=0.0)
tmp18 = tmp7 >= tmp7
tmp20 = tl.load(in_ptr0 + tl.broadcast_to(8 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp18, other=0.0)
tmp21 = tl.where(tmp16, tmp17, tmp20)
tmp22 = tmp14 + tmp21
tmp23 = tl.full([1, 1], 2, tl.int64)
tmp25 = tmp23 < tmp7
tmp26 = tl.load(in_ptr0 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp25,
other=0.0)
tmp27 = tmp23 >= tmp7
tmp29 = tl.load(in_ptr0 + tl.broadcast_to(8 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp27, other=0.0)
tmp30 = tl.where(tmp25, tmp26, tmp29)
tmp31 = tmp22 + tmp30
tmp32 = 3.0
tmp33 = tmp31 / tmp32
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp33, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_2(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (16 + r0), None)
tmp5 = tl.load(in_ptr0 + (20 + r0), None)
tmp7 = tl.load(in_ptr0 + (24 + r0), None)
tmp9 = tl.load(in_ptr0 + (28 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK]),
tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tl.full([1, 1], 3, tl.int64)
tmp21 = tl.load(in_ptr0 + tl.broadcast_to(24 + r0 + 4 * -1, [XBLOCK,
RBLOCK]), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK]),
tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp28 = tl.load(in_ptr0 + tl.broadcast_to(24 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0, [XBLOCK, RBLOCK]),
tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp37 = tl.load(in_ptr0 + tl.broadcast_to(24 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr2 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp41, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_3(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, xnumel,
rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (32 + r0), None)
tmp5 = tl.load(in_ptr0 + (36 + r0), None)
tmp7 = tl.load(in_ptr0 + (40 + r0), None)
tmp9 = tl.load(in_ptr0 + (44 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK]),
tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tl.full([1, 1], 3, tl.int64)
tmp21 = tl.load(in_ptr0 + tl.broadcast_to(40 + r0 + 4 * -1, [XBLOCK,
RBLOCK]), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK]),
tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp28 = tl.load(in_ptr0 + tl.broadcast_to(40 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0, [XBLOCK, RBLOCK]),
tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp37 = tl.load(in_ptr0 + tl.broadcast_to(40 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tmp42 = tmp5 * tmp5
tmp43 = tl.broadcast_to(tmp42, [XBLOCK, RBLOCK])
tmp45 = tl.sum(tmp43, 1)[:, None]
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr2 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr4 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr5 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp41, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
tl.store(out_ptr6 + tl.full([XBLOCK, 1], 0, tl.int32), tmp45, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_4(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7,
xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (48 + r0), None)
tmp5 = tl.load(in_ptr0 + (52 + r0), None)
tmp7 = tl.load(in_ptr0 + (56 + r0), None)
tmp9 = tl.load(in_ptr0 + (60 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp0 + tmp5
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = 4.0
tmp12 = tmp10 / tmp11
tmp13 = tl.full([1, 1], 0, tl.int64)
tmp15 = tl.full([1, 1], 1, tl.int64)
tmp16 = tmp13 < tmp15
tmp17 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK]),
tmp16, other=0.0)
tmp18 = tmp13 >= tmp15
tl.full([1, 1], 3, tl.int64)
tmp21 = tl.load(in_ptr0 + tl.broadcast_to(56 + r0 + 4 * -1, [XBLOCK,
RBLOCK]), tmp18, other=0.0)
tmp22 = tl.where(tmp16, tmp17, tmp21)
tmp24 = tmp15 < tmp15
tmp25 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK]),
tmp24, other=0.0)
tmp26 = tmp15 >= tmp15
tmp28 = tl.load(in_ptr0 + tl.broadcast_to(56 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp26, other=0.0)
tmp29 = tl.where(tmp24, tmp25, tmp28)
tmp30 = tmp22 + tmp29
tmp31 = tl.full([1, 1], 2, tl.int64)
tmp33 = tmp31 < tmp15
tmp34 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0, [XBLOCK, RBLOCK]),
tmp33, other=0.0)
tmp35 = tmp31 >= tmp15
tmp37 = tl.load(in_ptr0 + tl.broadcast_to(56 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp35, other=0.0)
tmp38 = tl.where(tmp33, tmp34, tmp37)
tmp39 = tmp30 + tmp38
tmp40 = 3.0
tmp41 = tmp39 / tmp40
tl.store(out_ptr1 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr2 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr4 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr5 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr6 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp12, None)
tl.store(out_ptr7 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp41, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_linalg_vector_norm_5(in_ptr0, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (4 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_6(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7,
out_ptr8, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (8 + r0), None)
tmp5 = tl.load(in_ptr0 + (12 + r0), None)
tmp39 = tl.load(in_ptr0 + r0, None)
tmp40 = tl.load(in_ptr0 + (4 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp10 = tl.full([1, 1], 0, tl.int64)
tmp12 = tl.full([1, 1], 2, tl.int64)
tmp13 = tmp10 < tmp12
tmp14 = tl.load(in_ptr0 + tl.broadcast_to(r0 + 4 * 0, [XBLOCK, RBLOCK]),
tmp13, other=0.0)
tmp15 = tmp10 >= tmp12
tl.full([1, 1], 3, tl.int64)
tmp18 = tl.load(in_ptr0 + tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK]),
tmp15, other=0.0)
tmp19 = tl.where(tmp13, tmp14, tmp18)
tmp20 = tl.full([1, 1], 1, tl.int64)
tmp22 = tmp20 < tmp12
tmp23 = tl.load(in_ptr0 + tl.broadcast_to(r0 + 4 * 1, [XBLOCK, RBLOCK]),
tmp22, other=0.0)
tmp24 = tmp20 >= tmp12
tmp26 = tl.load(in_ptr0 + tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK]),
tmp24, other=0.0)
tmp27 = tl.where(tmp22, tmp23, tmp26)
tmp28 = tmp19 + tmp27
tmp30 = tmp12 < tmp12
tmp31 = tl.load(in_ptr0 + tl.broadcast_to(r0 + 4 * 2, [XBLOCK, RBLOCK]),
tmp30, other=0.0)
tmp32 = tmp12 >= tmp12
tmp34 = tl.load(in_ptr0 + tl.broadcast_to(12 + r0, [XBLOCK, RBLOCK]),
tmp32, other=0.0)
tmp35 = tl.where(tmp30, tmp31, tmp34)
tmp36 = tmp28 + tmp35
tmp37 = 3.0
tmp38 = tmp36 / tmp37
tmp41 = tmp39 + tmp40
tmp42 = tmp41 + tmp0
tmp43 = tmp42 + tmp5
tmp44 = 4.0
tmp45 = tmp43 / tmp44
tl.store(out_ptr2 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp38, None)
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr4 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr5 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr6 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr7 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr8 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, None)
@triton.jit
def triton_poi_fused_cat_7(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10, in_ptr11,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x0 = xindex % 4
x2 = xindex
tmp6 = tl.load(in_ptr1 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp32 = tl.load(in_ptr4 + 0)
tmp33 = tl.broadcast_to(tmp32, [XBLOCK])
tmp57 = tl.load(in_ptr7 + 0)
tmp58 = tl.broadcast_to(tmp57, [XBLOCK])
tmp81 = tl.load(in_ptr10 + 0)
tmp82 = tl.broadcast_to(tmp81, [XBLOCK])
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp8 = libdevice.sqrt(tmp7)
tmp9 = tl.load(in_ptr2 + 4 * x0, tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp10 = tmp9 * tmp9
tmp11 = tl.load(in_ptr2 + (1 + 4 * x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tl.load(in_ptr2 + (2 + 4 * x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tl.load(in_ptr2 + (3 + 4 * x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = libdevice.sqrt(tmp19)
tmp21 = tmp8 * tmp20
tmp22 = tmp5 / tmp21
tmp23 = 1e-06
tmp24 = triton_helpers.maximum(tmp22, tmp23)
tmp25 = tl.full(tmp24.shape, 0.0, tmp24.dtype)
tmp26 = tl.where(tmp4, tmp24, tmp25)
tmp27 = tmp0 >= tmp3
tmp28 = tl.full([1], 2, tl.int64)
tmp29 = tmp0 < tmp28
tmp30 = tmp27 & tmp29
tmp31 = tl.load(in_ptr3 + x0, tmp30 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp34 = libdevice.sqrt(tmp33)
tmp35 = tl.load(in_ptr5 + 4 * x0, tmp30 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp36 = tmp35 * tmp35
tmp37 = tl.load(in_ptr5 + (1 + 4 * x0), tmp30 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp38 = tmp37 * tmp37
tmp39 = tmp36 + tmp38
tmp40 = tl.load(in_ptr5 + (2 + 4 * x0), tmp30 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp41 = tmp40 * tmp40
tmp42 = tmp39 + tmp41
tmp43 = tl.load(in_ptr5 + (3 + 4 * x0), tmp30 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp44 = tmp43 * tmp43
tmp45 = tmp42 + tmp44
tmp46 = libdevice.sqrt(tmp45)
tmp47 = tmp34 * tmp46
tmp48 = tmp31 / tmp47
tmp49 = triton_helpers.maximum(tmp48, tmp23)
tmp50 = tl.full(tmp49.shape, 0.0, tmp49.dtype)
tmp51 = tl.where(tmp30, tmp49, tmp50)
tmp52 = tmp0 >= tmp28
tmp53 = tl.full([1], 3, tl.int64)
tmp54 = tmp0 < tmp53
tmp55 = tmp52 & tmp54
tmp56 = tl.load(in_ptr6 + x0, tmp55 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp59 = libdevice.sqrt(tmp58)
tmp60 = tl.load(in_ptr8 + 4 * x0, tmp55 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp61 = tmp60 * tmp60
tmp62 = tl.load(in_ptr8 + (1 + 4 * x0), tmp55 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp63 = tmp62 * tmp62
tmp64 = tmp61 + tmp63
tmp65 = tl.load(in_ptr8 + (2 + 4 * x0), tmp55 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp66 = tmp65 * tmp65
tmp67 = tmp64 + tmp66
tmp68 = tl.load(in_ptr8 + (3 + 4 * x0), tmp55 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp69 = tmp68 * tmp68
tmp70 = tmp67 + tmp69
tmp71 = libdevice.sqrt(tmp70)
tmp72 = tmp59 * tmp71
tmp73 = tmp56 / tmp72
tmp74 = triton_helpers.maximum(tmp73, tmp23)
tmp75 = tl.full(tmp74.shape, 0.0, tmp74.dtype)
tmp76 = tl.where(tmp55, tmp74, tmp75)
tmp77 = tmp0 >= tmp53
tl.full([1], 4, tl.int64)
tmp80 = tl.load(in_ptr9 + x0, tmp77 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp83 = libdevice.sqrt(tmp82)
tmp84 = tl.load(in_ptr11 + 4 * x0, tmp77 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp85 = tmp84 * tmp84
tmp86 = tl.load(in_ptr11 + (1 + 4 * x0), tmp77 & xmask, eviction_policy
='evict_last', other=0.0)
tmp87 = tmp86 * tmp86
tmp88 = tmp85 + tmp87
tmp89 = tl.load(in_ptr11 + (2 + 4 * x0), tmp77 & xmask, eviction_policy
='evict_last', other=0.0)
tmp90 = tmp89 * tmp89
tmp91 = tmp88 + tmp90
tmp92 = tl.load(in_ptr11 + (3 + 4 * x0), tmp77 & xmask, eviction_policy
='evict_last', other=0.0)
tmp93 = tmp92 * tmp92
tmp94 = tmp91 + tmp93
tmp95 = libdevice.sqrt(tmp94)
tmp96 = tmp83 * tmp95
tmp97 = tmp80 / tmp96
tmp98 = triton_helpers.maximum(tmp97, tmp23)
tmp99 = tl.full(tmp98.shape, 0.0, tmp98.dtype)
tmp100 = tl.where(tmp77, tmp98, tmp99)
tmp101 = tl.where(tmp55, tmp76, tmp100)
tmp102 = tl.where(tmp30, tmp51, tmp101)
tmp103 = tl.where(tmp4, tmp26, tmp102)
tl.store(out_ptr0 + x2, tmp103, xmask)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_8(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, out_ptr7,
xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (20 + r0), None)
tmp5 = tl.load(in_ptr0 + (24 + r0), None)
tmp10 = tl.load(in_ptr0 + (28 + r0), None)
tmp44 = tl.load(in_ptr0 + (16 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp11 = tmp10 * tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp15 = tl.full([1, 1], 0, tl.int64)
tmp17 = tl.full([1, 1], 2, tl.int64)
tmp18 = tmp15 < tmp17
tmp19 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp18, other=0.0)
tmp20 = tmp15 >= tmp17
tl.full([1, 1], 3, tl.int64)
tmp23 = tl.load(in_ptr0 + tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK]),
tmp20, other=0.0)
tmp24 = tl.where(tmp18, tmp19, tmp23)
tmp25 = tl.full([1, 1], 1, tl.int64)
tmp27 = tmp25 < tmp17
tmp28 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp27, other=0.0)
tmp29 = tmp25 >= tmp17
tmp31 = tl.load(in_ptr0 + tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK]),
tmp29, other=0.0)
tmp32 = tl.where(tmp27, tmp28, tmp31)
tmp33 = tmp24 + tmp32
tmp35 = tmp17 < tmp17
tmp36 = tl.load(in_ptr0 + tl.broadcast_to(16 + r0 + 4 * 2, [XBLOCK,
RBLOCK]), tmp35, other=0.0)
tmp37 = tmp17 >= tmp17
tmp39 = tl.load(in_ptr0 + tl.broadcast_to(28 + r0, [XBLOCK, RBLOCK]),
tmp37, other=0.0)
tmp40 = tl.where(tmp35, tmp36, tmp39)
tmp41 = tmp33 + tmp40
tmp42 = 3.0
tmp43 = tmp41 / tmp42
tmp45 = tmp44 + tmp0
tmp46 = tmp45 + tmp5
tmp47 = tmp46 + tmp10
tmp48 = 4.0
tmp49 = tmp47 / tmp48
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp43, None)
tl.store(out_ptr4 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp49, None)
tl.store(out_ptr5 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp49, None)
tl.store(out_ptr6 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp49, None)
tl.store(out_ptr7 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp49, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, None)
tl.store(out_ptr2 + tl.full([XBLOCK, 1], 0, tl.int32), tmp14, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_stack_9(in_ptr0, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK: tl.
constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (40 + r0), None)
tmp5 = tl.load(in_ptr0 + (44 + r0), None)
tmp39 = tl.load(in_ptr0 + (32 + r0), None)
tmp40 = tl.load(in_ptr0 + (36 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tmp6 = tmp5 * tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp10 = tl.full([1, 1], 0, tl.int64)
tmp12 = tl.full([1, 1], 2, tl.int64)
tmp13 = tmp10 < tmp12
tmp14 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp13, other=0.0)
tmp15 = tmp10 >= tmp12
tl.full([1, 1], 3, tl.int64)
tmp18 = tl.load(in_ptr0 + tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK]),
tmp15, other=0.0)
tmp19 = tl.where(tmp13, tmp14, tmp18)
tmp20 = tl.full([1, 1], 1, tl.int64)
tmp22 = tmp20 < tmp12
tmp23 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp22, other=0.0)
tmp24 = tmp20 >= tmp12
tmp26 = tl.load(in_ptr0 + tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK]),
tmp24, other=0.0)
tmp27 = tl.where(tmp22, tmp23, tmp26)
tmp28 = tmp19 + tmp27
tmp30 = tmp12 < tmp12
tmp31 = tl.load(in_ptr0 + tl.broadcast_to(32 + r0 + 4 * 2, [XBLOCK,
RBLOCK]), tmp30, other=0.0)
tmp32 = tmp12 >= tmp12
tmp34 = tl.load(in_ptr0 + tl.broadcast_to(44 + r0, [XBLOCK, RBLOCK]),
tmp32, other=0.0)
tmp35 = tl.where(tmp30, tmp31, tmp34)
tmp36 = tmp28 + tmp35
tmp37 = 3.0
tmp38 = tmp36 / tmp37
tmp41 = tmp39 + tmp40
tmp42 = tmp41 + tmp0
tmp43 = tmp42 + tmp5
tmp44 = 4.0
tmp45 = tmp43 / tmp44
tl.store(out_ptr2 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp38, None)
tl.store(out_ptr3 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr4 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp45, None)
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, None)
@triton.jit
def triton_per_fused_linalg_vector_norm_10(in_ptr0, out_ptr0, xnumel,
rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (52 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_per_fused_cat_linalg_vector_norm_mean_11(in_ptr0, out_ptr0,
out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp29 = tl.load(in_ptr0 + (60 + r0), None)
tmp0 = tl.full([1, 1], 0, tl.int64)
tmp2 = tl.full([1, 1], 2, tl.int64)
tmp3 = tmp0 < tmp2
tmp4 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0 + 4 * 0, [XBLOCK,
RBLOCK]), tmp3, other=0.0)
tmp5 = tmp0 >= tmp2
tl.full([1, 1], 3, tl.int64)
tmp8 = tl.load(in_ptr0 + tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK]),
tmp5, other=0.0)
tmp9 = tl.where(tmp3, tmp4, tmp8)
tmp10 = tl.full([1, 1], 1, tl.int64)
tmp12 = tmp10 < tmp2
tmp13 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0 + 4 * 1, [XBLOCK,
RBLOCK]), tmp12, other=0.0)
tmp14 = tmp10 >= tmp2
tmp16 = tl.load(in_ptr0 + tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK]),
tmp14, other=0.0)
tmp17 = tl.where(tmp12, tmp13, tmp16)
tmp18 = tmp9 + tmp17
tmp20 = tmp2 < tmp2
tmp21 = tl.load(in_ptr0 + tl.broadcast_to(48 + r0 + 4 * 2, [XBLOCK,
RBLOCK]), tmp20, other=0.0)
tmp22 = tmp2 >= tmp2
tmp24 = tl.load(in_ptr0 + tl.broadcast_to(60 + r0, [XBLOCK, RBLOCK]),
tmp22, other=0.0)
tmp25 = tl.where(tmp20, tmp21, tmp24)
tmp26 = tmp18 + tmp25
tmp27 = 3.0
tmp28 = tmp26 / tmp27
tmp30 = tmp29 * tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr0 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp28, None)
tl.store(out_ptr1 + tl.full([XBLOCK, 1], 0, tl.int32), tmp33, None)
@triton.jit
def triton_per_fused_linalg_vector_norm_12(in_ptr0, out_ptr0, xnumel,
rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (56 + r0), None)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.sum(tmp2, 1)[:, None]
tl.store(out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp4, None)
@triton.jit
def triton_poi_fused_stack_13(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp6 = tl.load(in_ptr0 + (4 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (8 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = tl.load(in_ptr0 + (12 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp11 = tmp9 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp4, tmp13, tmp14)
tmp16 = tmp0 >= tmp3
tmp17 = tl.full([1], 8, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr0 + (16 + (-4 + x0)), tmp19 & xmask,
eviction_policy='evict_last', other=0.0)
tmp21 = tl.load(in_ptr0 + (20 + (-4 + x0)), tmp19 & xmask,
eviction_policy='evict_last', other=0.0)
tmp22 = tmp20 + tmp21
tmp23 = tl.load(in_ptr0 + (24 + (-4 + x0)), tmp19 & xmask,
eviction_policy='evict_last', other=0.0)
tmp24 = tmp22 + tmp23
tmp25 = tl.load(in_ptr0 + (28 + (-4 + x0)), tmp19 & xmask,
eviction_policy='evict_last', other=0.0)
tmp26 = tmp24 + tmp25
tmp27 = tmp26 / tmp12
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp19, tmp27, tmp28)
tmp30 = tmp0 >= tmp17
tmp31 = tl.full([1], 12, tl.int64)
tmp32 = tmp0 < tmp31
tmp33 = tmp30 & tmp32
tmp34 = tl.load(in_ptr0 + (32 + (-8 + x0)), tmp33 & xmask,
eviction_policy='evict_last', other=0.0)
tmp35 = tl.load(in_ptr0 + (36 + (-8 + x0)), tmp33 & xmask,
eviction_policy='evict_last', other=0.0)
tmp36 = tmp34 + tmp35
tmp37 = tl.load(in_ptr0 + (40 + (-8 + x0)), tmp33 & xmask,
eviction_policy='evict_last', other=0.0)
tmp38 = tmp36 + tmp37
tmp39 = tl.load(in_ptr0 + (44 + (-8 + x0)), tmp33 & xmask,
eviction_policy='evict_last', other=0.0)
tmp40 = tmp38 + tmp39
tmp41 = tmp40 / tmp12
tmp42 = tl.full(tmp41.shape, 0.0, tmp41.dtype)
tmp43 = tl.where(tmp33, tmp41, tmp42)
tmp44 = tmp0 >= tmp31
tl.full([1], 16, tl.int64)
tmp47 = tl.load(in_ptr0 + (48 + (-12 + x0)), tmp44 & xmask,
eviction_policy='evict_last', other=0.0)
tmp48 = tl.load(in_ptr0 + (52 + (-12 + x0)), tmp44 & xmask,
eviction_policy='evict_last', other=0.0)
tmp49 = tmp47 + tmp48
tmp50 = tl.load(in_ptr0 + (56 + (-12 + x0)), tmp44 & xmask,
eviction_policy='evict_last', other=0.0)
tmp51 = tmp49 + tmp50
tmp52 = 3.0
tmp53 = tmp51 / tmp52
tmp54 = tl.full(tmp53.shape, 0.0, tmp53.dtype)
tmp55 = tl.where(tmp44, tmp53, tmp54)
tmp56 = tl.where(tmp33, tmp43, tmp55)
tmp57 = tl.where(tmp19, tmp29, tmp56)
tmp58 = tl.where(tmp4, tmp15, tmp57)
tl.store(out_ptr0 + x0, tmp58, xmask)
@triton.jit
def triton_per_fused__log_softmax__log_softmax_backward_data_stack_14(in_ptr0,
in_ptr1, in_ptr2, out_ptr32, out_ptr33, out_ptr34, out_ptr35, out_ptr36,
out_ptr37, out_ptr38, out_ptr39, out_ptr40, out_ptr41, out_ptr42,
xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp2 = tl.load(in_ptr1 + r0, None)
tmp4 = tl.load(in_ptr2 + 0)
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp15 = tl.load(in_ptr1 + (4 + r0), None)
tmp26 = tl.load(in_ptr1 + (8 + r0), None)
tmp37 = tl.load(in_ptr1 + (12 + r0), None)
tmp48 = tl.load(in_ptr1 + (16 + r0), None)
tmp59 = tl.load(in_ptr1 + (20 + r0), None)
tmp70 = tl.load(in_ptr1 + (24 + r0), None)
tmp81 = tl.load(in_ptr1 + (28 + r0), None)
tmp92 = tl.load(in_ptr1 + (32 + r0), None)
tmp103 = tl.load(in_ptr1 + (36 + r0), None)
tmp114 = tl.load(in_ptr1 + (40 + r0), None)
tmp125 = tl.load(in_ptr1 + (44 + r0), None)
tmp136 = tl.load(in_ptr1 + (48 + r0), None)
tmp147 = tl.load(in_ptr1 + (52 + r0), None)
tmp158 = tl.load(in_ptr1 + (56 + r0), None)
tmp169 = tl.load(in_ptr1 + (60 + r0), None)
tmp185 = tl.load(in_ptr1 + 0)
tmp186 = tl.broadcast_to(tmp185, [XBLOCK, RBLOCK])
tmp199 = tl.load(in_ptr1 + 4)
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp213 = tl.load(in_ptr1 + 8)
tmp214 = tl.broadcast_to(tmp213, [XBLOCK, RBLOCK])
tmp226 = tl.load(in_ptr1 + 12)
tmp227 = tl.broadcast_to(tmp226, [XBLOCK, RBLOCK])
tmp239 = tl.load(in_ptr1 + 17)
tmp240 = tl.broadcast_to(tmp239, [XBLOCK, RBLOCK])
tmp249 = tl.load(in_ptr1 + 21)
tmp250 = tl.broadcast_to(tmp249, [XBLOCK, RBLOCK])
tmp259 = tl.load(in_ptr1 + 25)
tmp260 = tl.broadcast_to(tmp259, [XBLOCK, RBLOCK])
tmp269 = tl.load(in_ptr1 + 29)
tmp270 = tl.broadcast_to(tmp269, [XBLOCK, RBLOCK])
tmp282 = tl.load(in_ptr1 + 34)
tmp283 = tl.broadcast_to(tmp282, [XBLOCK, RBLOCK])
tmp292 = tl.load(in_ptr1 + 38)
tmp293 = tl.broadcast_to(tmp292, [XBLOCK, RBLOCK])
tmp302 = tl.load(in_ptr1 + 42)
tmp303 = tl.broadcast_to(tmp302, [XBLOCK, RBLOCK])
tmp312 = tl.load(in_ptr1 + 46)
tmp313 = tl.broadcast_to(tmp312, [XBLOCK, RBLOCK])
tmp3 = tmp1 * tmp2
tmp6 = tmp3 + tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = triton_helpers.max2(tmp7, 1)[:, None]
tmp10 = tmp6 - tmp9
tmp11 = tl_math.exp(tmp10)
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp16 = tmp1 * tmp15
tmp17 = tmp16 + tmp5
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = triton_helpers.max2(tmp18, 1)[:, None]
tmp21 = tmp17 - tmp20
tmp22 = tl_math.exp(tmp21)
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.sum(tmp23, 1)[:, None]
tmp27 = tmp1 * tmp26
tmp28 = tmp27 + tmp5
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = triton_helpers.max2(tmp29, 1)[:, None]
tmp32 = tmp28 - tmp31
tmp33 = tl_math.exp(tmp32)
tmp34 = tl.broadcast_to(tmp33, [XBLOCK, RBLOCK])
tmp36 = tl.sum(tmp34, 1)[:, None]
tmp38 = tmp1 * tmp37
tmp39 = tmp38 + tmp5
tmp40 = tl.broadcast_to(tmp39, [XBLOCK, RBLOCK])
tmp42 = triton_helpers.max2(tmp40, 1)[:, None]
tmp43 = tmp39 - tmp42
tmp44 = tl_math.exp(tmp43)
tmp45 = tl.broadcast_to(tmp44, [XBLOCK, RBLOCK])
tmp47 = tl.sum(tmp45, 1)[:, None]
tmp49 = tmp1 * tmp48
tmp50 = tmp49 + tmp5
tmp51 = tl.broadcast_to(tmp50, [XBLOCK, RBLOCK])
tmp53 = triton_helpers.max2(tmp51, 1)[:, None]
tmp54 = tmp50 - tmp53
tmp55 = tl_math.exp(tmp54)
tmp56 = tl.broadcast_to(tmp55, [XBLOCK, RBLOCK])
tmp58 = tl.sum(tmp56, 1)[:, None]
tmp60 = tmp1 * tmp59
tmp61 = tmp60 + tmp5
tmp62 = tl.broadcast_to(tmp61, [XBLOCK, RBLOCK])
tmp64 = triton_helpers.max2(tmp62, 1)[:, None]
tmp65 = tmp61 - tmp64
tmp66 = tl_math.exp(tmp65)
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = tl.sum(tmp67, 1)[:, None]
tmp71 = tmp1 * tmp70
tmp72 = tmp71 + tmp5
tmp73 = tl.broadcast_to(tmp72, [XBLOCK, RBLOCK])
tmp75 = triton_helpers.max2(tmp73, 1)[:, None]
tmp76 = tmp72 - tmp75
tmp77 = tl_math.exp(tmp76)
tmp78 = tl.broadcast_to(tmp77, [XBLOCK, RBLOCK])
tmp80 = tl.sum(tmp78, 1)[:, None]
tmp82 = tmp1 * tmp81
tmp83 = tmp82 + tmp5
tmp84 = tl.broadcast_to(tmp83, [XBLOCK, RBLOCK])
tmp86 = triton_helpers.max2(tmp84, 1)[:, None]
tmp87 = tmp83 - tmp86
tmp88 = tl_math.exp(tmp87)
tmp89 = tl.broadcast_to(tmp88, [XBLOCK, RBLOCK])
tmp91 = tl.sum(tmp89, 1)[:, None]
tmp93 = tmp1 * tmp92
tmp94 = tmp93 + tmp5
tmp95 = tl.broadcast_to(tmp94, [XBLOCK, RBLOCK])
tmp97 = triton_helpers.max2(tmp95, 1)[:, None]
tmp98 = tmp94 - tmp97
tmp99 = tl_math.exp(tmp98)
tmp100 = tl.broadcast_to(tmp99, [XBLOCK, RBLOCK])
tmp102 = tl.sum(tmp100, 1)[:, None]
tmp104 = tmp1 * tmp103
tmp105 = tmp104 + tmp5
tmp106 = tl.broadcast_to(tmp105, [XBLOCK, RBLOCK])
tmp108 = triton_helpers.max2(tmp106, 1)[:, None]
tmp109 = tmp105 - tmp108
tmp110 = tl_math.exp(tmp109)
tmp111 = tl.broadcast_to(tmp110, [XBLOCK, RBLOCK])
tmp113 = tl.sum(tmp111, 1)[:, None]
tmp115 = tmp1 * tmp114
tmp116 = tmp115 + tmp5
tmp117 = tl.broadcast_to(tmp116, [XBLOCK, RBLOCK])
tmp119 = triton_helpers.max2(tmp117, 1)[:, None]
tmp120 = tmp116 - tmp119
tmp121 = tl_math.exp(tmp120)
tmp122 = tl.broadcast_to(tmp121, [XBLOCK, RBLOCK])
tmp124 = tl.sum(tmp122, 1)[:, None]
tmp126 = tmp1 * tmp125
tmp127 = tmp126 + tmp5
tmp128 = tl.broadcast_to(tmp127, [XBLOCK, RBLOCK])
tmp130 = triton_helpers.max2(tmp128, 1)[:, None]
tmp131 = tmp127 - tmp130
tmp132 = tl_math.exp(tmp131)
tmp133 = tl.broadcast_to(tmp132, [XBLOCK, RBLOCK])
tmp135 = tl.sum(tmp133, 1)[:, None]
tmp137 = tmp1 * tmp136
tmp138 = tmp137 + tmp5
tmp139 = tl.broadcast_to(tmp138, [XBLOCK, RBLOCK])
tmp141 = triton_helpers.max2(tmp139, 1)[:, None]
tmp142 = tmp138 - tmp141
tmp143 = tl_math.exp(tmp142)
tmp144 = tl.broadcast_to(tmp143, [XBLOCK, RBLOCK])
tmp146 = tl.sum(tmp144, 1)[:, None]
tmp148 = tmp1 * tmp147
tmp149 = tmp148 + tmp5
tmp150 = tl.broadcast_to(tmp149, [XBLOCK, RBLOCK])
tmp152 = triton_helpers.max2(tmp150, 1)[:, None]
tmp153 = tmp149 - tmp152
tmp154 = tl_math.exp(tmp153)
tmp155 = tl.broadcast_to(tmp154, [XBLOCK, RBLOCK])
tmp157 = tl.sum(tmp155, 1)[:, None]
tmp159 = tmp1 * tmp158
tmp160 = tmp159 + tmp5
tmp161 = tl.broadcast_to(tmp160, [XBLOCK, RBLOCK])
tmp163 = triton_helpers.max2(tmp161, 1)[:, None]
tmp164 = tmp160 - tmp163
tmp165 = tl_math.exp(tmp164)
tmp166 = tl.broadcast_to(tmp165, [XBLOCK, RBLOCK])
tmp168 = tl.sum(tmp166, 1)[:, None]
tmp170 = tmp1 * tmp169
tmp171 = tmp170 + tmp5
tmp172 = tl.broadcast_to(tmp171, [XBLOCK, RBLOCK])
tmp174 = triton_helpers.max2(tmp172, 1)[:, None]
tmp175 = tmp171 - tmp174
tmp176 = tl_math.exp(tmp175)
tmp177 = tl.broadcast_to(tmp176, [XBLOCK, RBLOCK])
tmp179 = tl.sum(tmp177, 1)[:, None]
tmp180 = r0
tl.full([1, 1], 0, tl.int64)
tmp183 = tl.full([1, 1], 1, tl.int64)
tmp184 = tmp180 < tmp183
tmp187 = tmp1 * tmp186
tmp188 = tmp187 + tmp5
tmp189 = tmp188 - tmp9
tmp190 = tl_math.log(tmp14)
tmp191 = tmp189 - tmp190
tmp192 = -tmp191
tmp193 = tl.full(tmp192.shape, 0.0, tmp192.dtype)
tmp194 = tl.where(tmp184, tmp192, tmp193)
tmp195 = tmp180 >= tmp183
tmp196 = tl.full([1, 1], 2, tl.int64)
tmp197 = tmp180 < tmp196
tmp198 = tmp195 & tmp197
tmp201 = tmp1 * tmp200
tmp202 = tmp201 + tmp5
tmp203 = tmp202 - tmp20
tmp204 = tl_math.log(tmp25)
tmp205 = tmp203 - tmp204
tmp206 = -tmp205
tmp207 = tl.full(tmp206.shape, 0.0, tmp206.dtype)
tmp208 = tl.where(tmp198, tmp206, tmp207)
tmp209 = tmp180 >= tmp196
tmp210 = tl.full([1, 1], 3, tl.int64)
tmp211 = tmp180 < tmp210
tmp212 = tmp209 & tmp211
tmp215 = tmp1 * tmp214
tmp216 = tmp215 + tmp5
tmp217 = tmp216 - tmp31
tmp218 = tl_math.log(tmp36)
tmp219 = tmp217 - tmp218
tmp220 = -tmp219
tmp221 = tl.full(tmp220.shape, 0.0, tmp220.dtype)
tmp222 = tl.where(tmp212, tmp220, tmp221)
tmp223 = tmp180 >= tmp210
tl.full([1, 1], 4, tl.int64)
tmp228 = tmp1 * tmp227
tmp229 = tmp228 + tmp5
tmp230 = tmp229 - tmp42
tmp231 = tl_math.log(tmp47)
tmp232 = tmp230 - tmp231
tmp233 = -tmp232
tmp234 = tl.full(tmp233.shape, 0.0, tmp233.dtype)
tmp235 = tl.where(tmp223, tmp233, tmp234)
tmp236 = tl.where(tmp212, tmp222, tmp235)
tmp237 = tl.where(tmp198, tmp208, tmp236)
tmp238 = tl.where(tmp184, tmp194, tmp237)
tmp241 = tmp1 * tmp240
tmp242 = tmp241 + tmp5
tmp243 = tmp242 - tmp53
tmp244 = tl_math.log(tmp58)
tmp245 = tmp243 - tmp244
tmp246 = -tmp245
tmp247 = tl.full(tmp246.shape, 0.0, tmp246.dtype)
tmp248 = tl.where(tmp184, tmp246, tmp247)
tmp251 = tmp1 * tmp250
tmp252 = tmp251 + tmp5
tmp253 = tmp252 - tmp64
tmp254 = tl_math.log(tmp69)
tmp255 = tmp253 - tmp254
tmp256 = -tmp255
tmp257 = tl.full(tmp256.shape, 0.0, tmp256.dtype)
tmp258 = tl.where(tmp198, tmp256, tmp257)
tmp261 = tmp1 * tmp260
tmp262 = tmp261 + tmp5
tmp263 = tmp262 - tmp75
tmp264 = tl_math.log(tmp80)
tmp265 = tmp263 - tmp264
tmp266 = -tmp265
tmp267 = tl.full(tmp266.shape, 0.0, tmp266.dtype)
tmp268 = tl.where(tmp212, tmp266, tmp267)
tmp271 = tmp1 * tmp270
tmp272 = tmp271 + tmp5
tmp273 = tmp272 - tmp86
tmp274 = tl_math.log(tmp91)
tmp275 = tmp273 - tmp274
tmp276 = -tmp275
tmp277 = tl.full(tmp276.shape, 0.0, tmp276.dtype)
tmp278 = tl.where(tmp223, tmp276, tmp277)
tmp279 = tl.where(tmp212, tmp268, tmp278)
tmp280 = tl.where(tmp198, tmp258, tmp279)
tmp281 = tl.where(tmp184, tmp248, tmp280)
tmp284 = tmp1 * tmp283
tmp285 = tmp284 + tmp5
tmp286 = tmp285 - tmp97
tmp287 = tl_math.log(tmp102)
tmp288 = tmp286 - tmp287
tmp289 = -tmp288
tmp290 = tl.full(tmp289.shape, 0.0, tmp289.dtype)
tmp291 = tl.where(tmp184, tmp289, tmp290)
tmp294 = tmp1 * tmp293
tmp295 = tmp294 + tmp5
tmp296 = tmp295 - tmp108
tmp297 = tl_math.log(tmp113)
tmp298 = tmp296 - tmp297
tmp299 = -tmp298
tmp300 = tl.full(tmp299.shape, 0.0, tmp299.dtype)
tmp301 = tl.where(tmp198, tmp299, tmp300)
tmp304 = tmp1 * tmp303
tmp305 = tmp304 + tmp5
tmp306 = tmp305 - tmp119
tmp307 = tl_math.log(tmp124)
tmp308 = tmp306 - tmp307
tmp309 = -tmp308
tmp310 = tl.full(tmp309.shape, 0.0, tmp309.dtype)
tmp311 = tl.where(tmp212, tmp309, tmp310)
tmp314 = tmp1 * tmp313
tmp315 = tmp314 + tmp5
tmp316 = tmp315 - tmp130
tmp317 = tl_math.log(tmp135)
tmp318 = tmp316 - tmp317
tmp319 = -tmp318
tmp320 = tl.full(tmp319.shape, 0.0, tmp319.dtype)
tmp321 = tl.where(tmp223, tmp319, tmp320)
tmp322 = tl.where(tmp212, tmp311, tmp321)
tmp323 = tl.where(tmp198, tmp301, tmp322)
tmp324 = tl.where(tmp184, tmp291, tmp323)
tmp325 = tl_math.log(tmp146)
tmp326 = tmp142 - tmp325
tmp327 = tl_math.exp(tmp326)
tmp328 = tl_math.log(tmp157)
tmp329 = tmp153 - tmp328
tmp330 = tl_math.exp(tmp329)
tmp331 = tl_math.log(tmp168)
tmp332 = tmp164 - tmp331
tmp333 = tl_math.exp(tmp332)
tmp334 = tl_math.log(tmp179)
tmp335 = tmp175 - tmp334
tmp336 = tl_math.exp(tmp335)
tl.store(out_ptr32 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp238, None)
tl.store(out_ptr33 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp281, None)
tl.store(out_ptr34 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp324, None)
tl.store(out_ptr35 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp326, None)
tl.store(out_ptr36 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp327, None)
tl.store(out_ptr37 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp329, None)
tl.store(out_ptr38 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp330, None)
tl.store(out_ptr39 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp332, None)
tl.store(out_ptr40 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp333, None)
tl.store(out_ptr41 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp335, None)
tl.store(out_ptr42 + tl.broadcast_to(r0, [XBLOCK, RBLOCK]), tmp336, None)
@triton.jit
def triton_poi_fused_stack_15(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp5 = tl.load(in_ptr0 + 3)
tmp6 = tl.broadcast_to(tmp5, [XBLOCK])
tmp14 = tl.load(in_ptr1 + 3)
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tmp23 = tl.load(in_ptr2 + 3)
tmp24 = tl.broadcast_to(tmp23, [XBLOCK])
tmp31 = tl.load(in_ptr3 + 3)
tmp32 = tl.broadcast_to(tmp31, [XBLOCK])
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp7 = -tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 2, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tmp10 & tmp12
tmp16 = -tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp13, tmp16, tmp17)
tmp19 = tmp0 >= tmp11
tmp20 = tl.full([1], 3, tl.int64)
tmp21 = tmp0 < tmp20
tmp22 = tmp19 & tmp21
tmp25 = -tmp24
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp22, tmp25, tmp26)
tmp28 = tmp0 >= tmp20
tl.full([1], 4, tl.int64)
tmp33 = -tmp32
tmp34 = tl.full(tmp33.shape, 0.0, tmp33.dtype)
tmp35 = tl.where(tmp28, tmp33, tmp34)
tmp36 = tl.where(tmp22, tmp27, tmp35)
tmp37 = tl.where(tmp13, tmp18, tmp36)
tmp38 = tl.where(tmp4, tmp9, tmp37)
tl.store(out_ptr0 + x0, tmp38, xmask)
@triton.jit
def triton_per_fused_mean_16(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tmp4 = 16.0
tmp5 = tmp3 / tmp4
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp5, None)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (), ())
assert_size_stride(primals_3, (), ())
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16,), (1,), torch.float32)
buf17 = empty_strided_cuda((16,), (1,), torch.float32)
buf21 = empty_strided_cuda((16,), (1,), torch.float32)
buf38 = empty_strided_cuda((16,), (1,), torch.float32)
buf42 = empty_strided_cuda((16,), (1,), torch.float32)
buf59 = empty_strided_cuda((16,), (1,), torch.float32)
buf63 = empty_strided_cuda((16,), (1,), torch.float32)
get_raw_stream(0)
triton_poi_fused_stack_0[grid(16)](primals_1, buf0, buf17, buf21,
buf38, buf42, buf59, buf63, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf1 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 0),
reinterpret_tensor(buf0, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((), (), torch.float32)
buf7 = empty_strided_cuda((16,), (1,), torch.float32)
buf3 = reinterpret_tensor(buf7, (4,), (1,), 0)
triton_per_fused_cat_linalg_vector_norm_mean_1[grid(1)](primals_1,
buf2, buf3, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
buf23 = empty_strided_cuda((), (), torch.float32)
buf4 = reinterpret_tensor(buf7, (4,), (1,), 4)
buf14 = empty_strided_cuda((16,), (1,), torch.float32)
buf11 = reinterpret_tensor(buf14, (4,), (1,), 4)
buf28 = empty_strided_cuda((16,), (1,), torch.float32)
buf24 = reinterpret_tensor(buf28, (4,), (1,), 4)
triton_per_fused_cat_linalg_vector_norm_mean_stack_2[grid(1)](primals_1
, buf23, buf4, buf11, buf24, 1, 4, XBLOCK=1, num_warps=2,
num_stages=1)
buf44 = empty_strided_cuda((), (), torch.float32)
buf5 = reinterpret_tensor(buf7, (4,), (1,), 8)
buf12 = reinterpret_tensor(buf14, (4,), (1,), 8)
buf26 = reinterpret_tensor(buf28, (4,), (1,), 8)
buf35 = empty_strided_cuda((16,), (1,), torch.float32)
buf33 = reinterpret_tensor(buf35, (4,), (1,), 8)
buf49 = empty_strided_cuda((16,), (1,), torch.float32)
buf45 = reinterpret_tensor(buf49, (4,), (1,), 8)
buf51 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_cat_linalg_vector_norm_mean_stack_3[grid(1)](primals_1
, buf44, buf5, buf12, buf26, buf33, buf45, buf51, 1, 4, XBLOCK=
1, num_warps=2, num_stages=1)
buf65 = empty_strided_cuda((), (), torch.float32)
buf6 = reinterpret_tensor(buf7, (4,), (1,), 12)
buf13 = reinterpret_tensor(buf14, (4,), (1,), 12)
buf27 = reinterpret_tensor(buf28, (4,), (1,), 12)
buf34 = reinterpret_tensor(buf35, (4,), (1,), 12)
buf48 = reinterpret_tensor(buf49, (4,), (1,), 12)
buf56 = empty_strided_cuda((16,), (1,), torch.float32)
buf55 = reinterpret_tensor(buf56, (4,), (1,), 12)
buf70 = empty_strided_cuda((16,), (1,), torch.float32)
buf66 = reinterpret_tensor(buf70, (4,), (1,), 12)
triton_per_fused_cat_linalg_vector_norm_mean_stack_4[grid(1)](primals_1
, buf65, buf6, buf13, buf27, buf34, buf48, buf55, buf66, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
del buf3
del buf4
del buf5
del buf6
buf8 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 4),
reinterpret_tensor(buf7, (4, 4), (1, 4), 0), out=buf8)
buf9 = empty_strided_cuda((), (), torch.float32)
triton_per_fused_linalg_vector_norm_5[grid(1)](primals_1, buf9, 1,
4, XBLOCK=1, num_warps=2, num_stages=1)
buf16 = empty_strided_cuda((), (), torch.float32)
buf19 = empty_strided_cuda((), (), torch.float32)
buf10 = reinterpret_tensor(buf14, (4,), (1,), 0)
buf25 = reinterpret_tensor(buf28, (4,), (1,), 0)
buf32 = reinterpret_tensor(buf35, (4,), (1,), 0)
buf46 = reinterpret_tensor(buf49, (4,), (1,), 0)
buf53 = reinterpret_tensor(buf56, (4,), (1,), 0)
buf67 = reinterpret_tensor(buf70, (4,), (1,), 0)
buf77 = empty_strided_cuda((16,), (1,), torch.float32)
buf74 = reinterpret_tensor(buf77, (4,), (1,), 0)
triton_per_fused_cat_linalg_vector_norm_mean_stack_6[grid(1)](primals_1
, buf16, buf19, buf10, buf25, buf32, buf46, buf53, buf67, buf74,
1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del buf10
del buf11
del buf12
del buf13
buf15 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 8),
reinterpret_tensor(buf14, (4, 4), (1, 4), 0), out=buf15)
buf18 = empty_strided_cuda((1, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 12),
reinterpret_tensor(buf17, (4, 4), (1, 4), 0), out=buf18)
buf84 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
buf20 = reinterpret_tensor(buf84, (4, 4), (4, 1), 0)
triton_poi_fused_cat_7[grid(16)](buf1, buf2, buf0, buf8, buf9, buf7,
buf15, buf16, buf14, buf18, buf19, buf17, buf20, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del buf0
del buf14
del buf16
del buf17
del buf7
buf22 = buf8
del buf8
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 16),
reinterpret_tensor(buf21, (4, 4), (1, 4), 0), out=buf22)
del buf24
del buf25
del buf26
del buf27
buf29 = buf18
del buf18
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 20),
reinterpret_tensor(buf28, (4, 4), (1, 4), 0), out=buf29)
buf30 = buf9
del buf9
buf37 = buf2
del buf2
buf40 = buf19
del buf19
buf31 = reinterpret_tensor(buf35, (4,), (1,), 4)
buf47 = reinterpret_tensor(buf49, (4,), (1,), 4)
buf54 = reinterpret_tensor(buf56, (4,), (1,), 4)
buf68 = reinterpret_tensor(buf70, (4,), (1,), 4)
buf75 = reinterpret_tensor(buf77, (4,), (1,), 4)
triton_per_fused_cat_linalg_vector_norm_mean_stack_8[grid(1)](primals_1
, buf30, buf37, buf40, buf31, buf47, buf54, buf68, buf75, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
del buf31
del buf32
del buf33
del buf34
buf36 = buf15
del buf15
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 24),
reinterpret_tensor(buf35, (4, 4), (1, 4), 0), out=buf36)
buf39 = buf1
del buf1
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 28),
reinterpret_tensor(buf38, (4, 4), (1, 4), 0), out=buf39)
buf41 = reinterpret_tensor(buf84, (4, 4), (4, 1), 16)
triton_poi_fused_cat_7[grid(16)](buf22, buf23, buf21, buf29, buf30,
buf28, buf36, buf37, buf35, buf39, buf40, buf38, buf41, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del buf21
del buf23
del buf28
del buf30
del buf35
del buf38
buf43 = buf39
del buf39
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 32),
reinterpret_tensor(buf42, (4, 4), (1, 4), 0), out=buf43)
del buf45
del buf46
del buf47
del buf48
buf50 = buf36
del buf36
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 36),
reinterpret_tensor(buf49, (4, 4), (1, 4), 0), out=buf50)
buf58 = buf40
del buf40
buf61 = buf37
del buf37
buf52 = reinterpret_tensor(buf56, (4,), (1,), 8)
buf69 = reinterpret_tensor(buf70, (4,), (1,), 8)
buf76 = reinterpret_tensor(buf77, (4,), (1,), 8)
triton_per_fused_cat_linalg_vector_norm_mean_stack_9[grid(1)](primals_1
, buf58, buf61, buf52, buf69, buf76, 1, 4, XBLOCK=1, num_warps=
2, num_stages=1)
del buf52
del buf53
del buf54
del buf55
buf57 = buf29
del buf29
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 40),
reinterpret_tensor(buf56, (4, 4), (1, 4), 0), out=buf57)
buf60 = buf22
del buf22
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 44),
reinterpret_tensor(buf59, (4, 4), (1, 4), 0), out=buf60)
buf62 = reinterpret_tensor(buf84, (4, 4), (4, 1), 32)
triton_poi_fused_cat_7[grid(16)](buf43, buf44, buf42, buf50, buf51,
buf49, buf57, buf58, buf56, buf60, buf61, buf59, buf62, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del buf42
del buf44
del buf49
del buf56
buf64 = buf60
del buf60
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 48),
reinterpret_tensor(buf63, (4, 4), (1, 4), 0), out=buf64)
del buf66
del buf67
del buf68
del buf69
buf71 = buf57
del buf57
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 52),
reinterpret_tensor(buf70, (4, 4), (1, 4), 0), out=buf71)
buf72 = buf61
del buf61
triton_per_fused_linalg_vector_norm_10[grid(1)](primals_1, buf72, 1,
4, XBLOCK=1, num_warps=2, num_stages=1)
buf73 = reinterpret_tensor(buf77, (4,), (1,), 12)
buf82 = buf58
del buf58
triton_per_fused_cat_linalg_vector_norm_mean_11[grid(1)](primals_1,
buf73, buf82, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del buf73
del buf74
del buf75
del buf76
buf78 = buf50
del buf50
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 56),
reinterpret_tensor(buf77, (4, 4), (1, 4), 0), out=buf78)
buf79 = buf51
del buf51
triton_per_fused_linalg_vector_norm_12[grid(1)](primals_1, buf79, 1,
4, XBLOCK=1, num_warps=2, num_stages=1)
buf80 = buf59
del buf59
triton_poi_fused_stack_13[grid(16)](primals_1, buf80, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf81 = buf43
del buf43
extern_kernels.mm(reinterpret_tensor(primals_1, (1, 4), (1, 1), 60),
reinterpret_tensor(buf80, (4, 4), (1, 4), 0), out=buf81)
del primals_1
buf83 = reinterpret_tensor(buf84, (4, 4), (4, 1), 48)
triton_poi_fused_cat_7[grid(16)](buf64, buf65, buf63, buf71, buf72,
buf70, buf78, buf79, buf77, buf81, buf82, buf80, buf83, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del buf63
del buf65
del buf70
del buf72
del buf77
del buf79
buf125 = buf80
del buf80
buf93 = reinterpret_tensor(buf125, (4,), (1,), 0)
buf102 = reinterpret_tensor(buf125, (4,), (1,), 4)
buf111 = reinterpret_tensor(buf125, (4,), (1,), 8)
buf114 = reinterpret_tensor(buf81, (4,), (1,), 0)
del buf81
buf130 = reinterpret_tensor(buf78, (4,), (1,), 0)
del buf78
buf117 = reinterpret_tensor(buf71, (4,), (1,), 0)
del buf71
buf129 = reinterpret_tensor(buf64, (4,), (1,), 0)
del buf64
buf120 = empty_strided_cuda((4,), (1,), torch.float32)
buf128 = empty_strided_cuda((4,), (1,), torch.float32)
buf123 = empty_strided_cuda((4,), (1,), torch.float32)
buf127 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused__log_softmax__log_softmax_backward_data_stack_14[grid
(1)](primals_2, buf84, primals_3, buf93, buf102, buf111, buf114,
buf130, buf117, buf129, buf120, buf128, buf123, buf127, 1, 4,
XBLOCK=1, num_warps=2, num_stages=1)
buf124 = reinterpret_tensor(buf125, (4,), (1,), 12)
triton_poi_fused_stack_15[grid(4)](buf114, buf117, buf120, buf123,
buf124, 4, XBLOCK=4, num_warps=1, num_stages=1)
del buf114
del buf117
del buf120
del buf123
buf126 = buf82
del buf82
buf131 = buf126
del buf126
triton_per_fused_mean_16[grid(1)](buf131, buf125, 1, 16, XBLOCK=1,
num_warps=2, num_stages=1)
del buf102
del buf111
del buf124
del buf125
del buf93
return buf131, primals_2, primals_3, buf84, buf127, buf128, buf129, buf130
class GE2ELossNew(nn.Module):
def __init__(self, init_w=10.0, init_b=-5.0, loss_method='softmax'):
"""
Implementation of the Generalized End-to-End loss defined in https://arxiv.org/abs/1710.10467 [1]
Accepts an input of size (N, M, D)
where N is the number of speakers in the batch,
M is the number of utterances per speaker,
and D is the dimensionality of the embedding vector (e.g. d-vector)
Args:
- init_w (float): defines the initial value of w in Equation (5) of [1]
- init_b (float): definies the initial value of b in Equation (5) of [1]
"""
super(GE2ELossNew, self).__init__()
self.w = nn.Parameter(torch.tensor(init_w))
self.b = nn.Parameter(torch.tensor(init_b))
self.loss_method = loss_method
assert self.loss_method in ['softmax', 'contrast']
if self.loss_method == 'softmax':
self.embed_loss = self.embed_loss_softmax
if self.loss_method == 'contrast':
self.embed_loss = self.embed_loss_contrast
def calc_new_centroids(self, dvecs, centroids, spkr, utt):
"""
Calculates the new centroids excluding the reference utterance
"""
excl = torch.cat((dvecs[spkr, :utt], dvecs[spkr, utt + 1:]))
excl = torch.mean(excl, 0)
new_centroids = []
for i, centroid in enumerate(centroids):
if i == spkr:
new_centroids.append(excl)
else:
new_centroids.append(centroid)
return torch.stack(new_centroids)
def calc_cosine_sim(self, dvecs, centroids):
"""
Make the cosine similarity matrix with dims (N,M,N)
"""
cos_sim_matrix = []
for spkr_idx, speaker in enumerate(dvecs):
cs_row = []
for utt_idx, utterance in enumerate(speaker):
new_centroids = self.calc_new_centroids(dvecs, centroids,
spkr_idx, utt_idx)
cs_row.append(torch.clamp(torch.mm(utterance.unsqueeze(1).
transpose(0, 1), new_centroids.transpose(0, 1)) / (
torch.norm(utterance) * torch.norm(new_centroids, dim=1
)), 1e-06))
cs_row = torch.cat(cs_row, dim=0)
cos_sim_matrix.append(cs_row)
return torch.stack(cos_sim_matrix)
def embed_loss_softmax(self, dvecs, cos_sim_matrix):
"""
Calculates the loss on each embedding $L(e_{ji})$ by taking softmax
"""
N, M, _ = dvecs.shape
L = []
for j in range(N):
L_row = []
for i in range(M):
L_row.append(-F.log_softmax(cos_sim_matrix[j, i], 0)[j])
L_row = torch.stack(L_row)
L.append(L_row)
return torch.stack(L)
def embed_loss_contrast(self, dvecs, cos_sim_matrix):
"""
Calculates the loss on each embedding $L(e_{ji})$ by contrast loss with closest centroid
"""
N, M, _ = dvecs.shape
L = []
for j in range(N):
L_row = []
for i in range(M):
centroids_sigmoids = torch.sigmoid(cos_sim_matrix[j, i])
excl_centroids_sigmoids = torch.cat((centroids_sigmoids[:j],
centroids_sigmoids[j + 1:]))
L_row.append(1.0 - torch.sigmoid(cos_sim_matrix[j, i, j]) +
torch.max(excl_centroids_sigmoids))
L_row = torch.stack(L_row)
L.append(L_row)
return torch.stack(L)
def forward(self, input_0):
primals_2 = self.w
primals_3 = self.b
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
greenstar1151/pytorch-benchmark
|
GE2ELoss
| false | 10,464 |
[
"BSD-3-Clause"
] | 0 |
8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
https://github.com/greenstar1151/pytorch-benchmark/tree/8b7808d3be6b7ca1d57f1812e35fd2df5e470f8b
|
MyLinear
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/fl/cflw6zjzdk2wqtau7m6nsei5vavjfijzxhb37zaa3xp4yxpw5yb2.py
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, 1), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2o/c2oqkq7zaubqmw7vuixxlseb2ff5jzqqbyczicxlmsahuxwdpdyp.py
# Topologically Sorted Source Nodes: [bias], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# bias => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, 1), kwargs = {})
triton_poi_fused_mul_1 = async_compile.triton('triton_poi_fused_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, ), (1, ))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_2, buf0, 16, grid=grid(16), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [bias], Original ATen: [aten.mul]
triton_poi_fused_mul_1.run(primals_1, buf1, 4, grid=grid(4), stream=stream0)
del primals_1
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [bias, linear], Original ATen: [aten.mul, aten.addmm]
extern_kernels.addmm(buf1, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf0
del buf1
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class MyLinear(nn.Module):
"""Linear layer with equalized learning rate and custom learning rate multiplier."""
def __init__(self, input_size, output_size, gain=2 ** 0.5, use_wscale=
False, lrmul=1, bias=True):
super().__init__()
he_std = gain * input_size ** -0.5
if use_wscale:
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
else:
init_std = he_std / lrmul
self.w_mul = lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
if bias:
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
else:
self.bias = None
def forward(self, x):
bias = self.bias
if bias is not None:
bias = bias * self.b_mul
return F.linear(x, self.weight * self.w_mul, bias)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'output_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4,), (1,))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(16)](primals_2, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_1, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_1
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(buf1, reinterpret_tensor(primals_3, (64, 4), (
4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1,
beta=1, out=buf2)
del buf0
del buf1
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0)
class MyLinearNew(nn.Module):
"""Linear layer with equalized learning rate and custom learning rate multiplier."""
def __init__(self, input_size, output_size, gain=2 ** 0.5, use_wscale=
False, lrmul=1, bias=True):
super().__init__()
he_std = gain * input_size ** -0.5
if use_wscale:
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
else:
init_std = he_std / lrmul
self.w_mul = lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
if bias:
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
else:
self.bias = None
def forward(self, input_0):
primals_2 = self.weight
primals_1 = self.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
justinpinkney/ganspace
|
MyLinear
| false | 10,465 |
[
"Apache-2.0"
] | 0 |
7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
https://github.com/justinpinkney/ganspace/tree/7dc76d1d2ddad21d946a7ceb375efe5d5316fb3f
|
VAE
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/g3/cg3el2gn3jo2uczn6kvxebxonhlsgf4gykdxpouwhsyjf55b5gdg.py
# Topologically Sorted Source Nodes: [h1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# h1 => relu
# Graph fragment:
# %add_tensor_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_2, %primals_3), kwargs = {})
# %relu : [num_users=3] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_2,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 500
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ki/ckicw4ebs4esmjq6arnmoylvxkanfgm3d6liitiqr6rwy5d63nvv.py
# Topologically Sorted Source Nodes: [mul, std, mul_1, z], Original ATen: [aten.mul, aten.exp, aten.add]
# Source node to ATen node mapping:
# mul => mul
# mul_1 => mul_1
# std => exp
# z => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%addmm_2, 0.5), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%mul,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%randn, %exp), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%addmm_1, %mul_1), kwargs = {})
triton_poi_fused_add_exp_mul_1 = async_compile.triton('triton_poi_fused_add_exp_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_exp_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 20
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = tl.load(in_ptr2 + (x0), xmask)
tmp3 = 0.5
tmp4 = tmp2 * tmp3
tmp5 = tl_math.exp(tmp4)
tmp6 = tmp1 * tmp5
tmp7 = tmp0 + tmp6
tl.store(out_ptr0 + (x0), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/hb/chbjjrtszu6f3bhry7ireqcm3ie3twpz5s7g7owb3zuauqhiqcby.py
# Topologically Sorted Source Nodes: [sigmoid], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid => sigmoid
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_11), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_sigmoid_2 = async_compile.triton('triton_poi_fused_sigmoid_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (500, 784), (784, 1))
assert_size_stride(primals_3, (500, ), (1, ))
assert_size_stride(primals_4, (5, 500), (500, 1))
assert_size_stride(primals_5, (5, ), (1, ))
assert_size_stride(primals_6, (5, 500), (500, 1))
assert_size_stride(primals_7, (5, ), (1, ))
assert_size_stride(primals_8, (500, 5), (5, 1))
assert_size_stride(primals_9, (500, ), (1, ))
assert_size_stride(primals_10, (784, 500), (500, 1))
assert_size_stride(primals_11, (784, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784, 500), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [h1], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 2000, grid=grid(2000), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
# Topologically Sorted Source Nodes: [mu], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, buf1, reinterpret_tensor(primals_4, (500, 5), (1, 500), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
# Topologically Sorted Source Nodes: [logvar], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf1, reinterpret_tensor(primals_6, (500, 5), (1, 500), 0), alpha=1, beta=1, out=buf3)
del primals_7
# Topologically Sorted Source Nodes: [eps], Original ATen: [aten.randn_like]
buf4 = torch.ops.aten.randn.default([4, 5], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, std, mul_1, z], Original ATen: [aten.mul, aten.exp, aten.add]
triton_poi_fused_add_exp_mul_1.run(buf2, buf5, buf3, buf6, 20, grid=grid(20), stream=stream0)
buf7 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf6, reinterpret_tensor(primals_8, (5, 500), (1, 5), 0), out=buf7)
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [h3], Original ATen: [aten.relu]
triton_poi_fused_relu_0.run(buf8, primals_9, 2000, grid=grid(2000), stream=stream0)
del primals_9
buf9 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf8, reinterpret_tensor(primals_10, (500, 784), (1, 500), 0), out=buf9)
buf10 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [sigmoid], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_2.run(buf10, primals_11, 3136, grid=grid(3136), stream=stream0)
del primals_11
return (buf10, buf2, buf3, primals_1, buf1, buf3, buf5, buf6, buf8, buf10, primals_10, primals_8, primals_6, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((500, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((500, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((5, 500), (500, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((5, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((5, 500), (500, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((5, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((500, 5), (5, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((500, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((784, 500), (500, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((784, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.data
class VAE(nn.Module):
def __init__(self):
super(VAE, self).__init__()
self.fc1 = nn.Linear(784, 500)
self.fc21 = nn.Linear(500, 5)
self.fc22 = nn.Linear(500, 5)
self.fc3 = nn.Linear(5, 500)
self.fc4 = nn.Linear(500, 784)
def encode(self, x):
h1 = F.relu(self.fc1(x))
return self.fc21(h1), self.fc22(h1)
def reparameterize(self, mu, logvar):
std = torch.exp(0.5 * logvar)
eps = torch.randn_like(std)
return mu + eps * std
def decode(self, z):
h3 = F.relu(self.fc3(z))
return torch.sigmoid(self.fc4(h3))
def forward(self, x):
mu, logvar = self.encode(x.view(-1, 784))
z = self.reparameterize(mu, logvar)
return self.decode(z), mu, logvar
def get_inputs():
return [torch.rand([4, 784])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch import device
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 2000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 500
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 20
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask)
tmp2 = tl.load(in_ptr2 + x0, xmask)
tmp3 = 0.5
tmp4 = tmp2 * tmp3
tmp5 = tl_math.exp(tmp4)
tmp6 = tmp1 * tmp5
tmp7 = tmp0 + tmp6
tl.store(out_ptr0 + x0, tmp7, xmask)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (500, 784), (784, 1))
assert_size_stride(primals_3, (500,), (1,))
assert_size_stride(primals_4, (5, 500), (500, 1))
assert_size_stride(primals_5, (5,), (1,))
assert_size_stride(primals_6, (5, 500), (500, 1))
assert_size_stride(primals_7, (5,), (1,))
assert_size_stride(primals_8, (500, 5), (5, 1))
assert_size_stride(primals_9, (500,), (1,))
assert_size_stride(primals_10, (784, 500), (500, 1))
assert_size_stride(primals_11, (784,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784,
500), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(2000)](buf1, primals_3, 2000, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
extern_kernels.addmm(primals_5, buf1, reinterpret_tensor(primals_4,
(500, 5), (1, 500), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
extern_kernels.addmm(primals_7, buf1, reinterpret_tensor(primals_6,
(500, 5), (1, 500), 0), alpha=1, beta=1, out=buf3)
del primals_7
buf4 = torch.ops.aten.randn.default([4, 5], dtype=torch.float32,
device=device(type='cuda', index=0), pin_memory=False)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 5), (5, 1), torch.float32)
triton_poi_fused_add_exp_mul_1[grid(20)](buf2, buf5, buf3, buf6, 20,
XBLOCK=32, num_warps=1, num_stages=1)
buf7 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
extern_kernels.mm(buf6, reinterpret_tensor(primals_8, (5, 500), (1,
5), 0), out=buf7)
buf8 = buf7
del buf7
triton_poi_fused_relu_0[grid(2000)](buf8, primals_9, 2000, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_9
buf9 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
extern_kernels.mm(buf8, reinterpret_tensor(primals_10, (500, 784),
(1, 500), 0), out=buf9)
buf10 = buf9
del buf9
triton_poi_fused_sigmoid_2[grid(3136)](buf10, primals_11, 3136,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_11
return (buf10, buf2, buf3, primals_1, buf1, buf3, buf5, buf6, buf8,
buf10, primals_10, primals_8, primals_6, primals_4)
class VAENew(nn.Module):
def __init__(self):
super(VAENew, self).__init__()
self.fc1 = nn.Linear(784, 500)
self.fc21 = nn.Linear(500, 5)
self.fc22 = nn.Linear(500, 5)
self.fc3 = nn.Linear(5, 500)
self.fc4 = nn.Linear(500, 784)
def encode(self, x):
h1 = F.relu(self.fc1(x))
return self.fc21(h1), self.fc22(h1)
def reparameterize(self, mu, logvar):
std = torch.exp(0.5 * logvar)
eps = torch.randn_like(std)
return mu + eps * std
def decode(self, z):
h3 = F.relu(self.fc3(z))
return torch.sigmoid(self.fc4(h3))
def forward(self, input_0):
primals_2 = self.fc1.weight
primals_3 = self.fc1.bias
primals_4 = self.fc21.weight
primals_5 = self.fc21.bias
primals_6 = self.fc22.weight
primals_7 = self.fc22.bias
primals_8 = self.fc3.weight
primals_9 = self.fc3.bias
primals_10 = self.fc4.weight
primals_11 = self.fc4.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0], output[1], output[2]
|
mcabbott/Avalon.jl
|
VAE
| false | 10,466 |
[
"MIT"
] | 0 |
6885bcc8204952a2396e762ce51432d9969c4138
|
https://github.com/mcabbott/Avalon.jl/tree/6885bcc8204952a2396e762ce51432d9969c4138
|
ContinuousNet
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/y7/cy7463cf27qpra2f6ndigmu6ve4q6o3cbvgetgqehejtevv6yfa5.py
# Topologically Sorted Source Nodes: [a1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# a1 => clamp_max, clamp_min
# Graph fragment:
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_1, 0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 6), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_1, 0), kwargs = {})
# %ge_1 : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%view_1, 6), kwargs = {})
# %bitwise_or_1 : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le_1, %ge_1), kwargs = {})
triton_poi_fused_hardtanh_hardtanh_backward_0 = async_compile.triton('triton_poi_fused_hardtanh_hardtanh_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_hardtanh_hardtanh_backward_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/oz/coz2vvm7j43dgp2scg63a7xyihnzn7xcwkbiwytoo3i3naiahfwp.py
# Topologically Sorted Source Nodes: [tanh, mu], Original ATen: [aten.tanh, aten.mul]
# Source node to ATen node mapping:
# mu => mul
# tanh => tanh
# Graph fragment:
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%view_3,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh, 2), kwargs = {})
triton_poi_fused_mul_tanh_1 = async_compile.triton('triton_poi_fused_mul_tanh_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_tanh_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_tanh_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = libdevice.tanh(tmp0)
tmp2 = 2.0
tmp3 = tmp1 * tmp2
tl.store(out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/az/cazly2utbmgqy4jzdtmpz56naibvqlqicln4ij6ptdvbf4ytmvkc.py
# Topologically Sorted Source Nodes: [softplus, sigma], Original ATen: [aten.softplus, aten.add]
# Source node to ATen node mapping:
# sigma => add
# softplus => exp, gt, log1p, where
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%view_5,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%exp,), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%view_5, 20), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %view_5, %log1p), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%where, 0.001), kwargs = {})
triton_poi_fused_add_softplus_2 = async_compile.triton('triton_poi_fused_add_softplus_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_softplus_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_softplus_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 20.0
tmp2 = tmp0 > tmp1
tmp3 = tl_math.exp(tmp0)
tmp4 = libdevice.log1p(tmp3)
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = 0.001
tmp7 = tmp5 + tmp6
tl.store(out_ptr0 + (x0), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zv/czv7olwo7h6ucl4yfkx5wdow4pryybtkxjpmco7mpv23b5hl3wfb.py
# Topologically Sorted Source Nodes: [c1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# c1 => clamp_max_1, clamp_min_1
# Graph fragment:
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_7, 0), kwargs = {})
# %clamp_max_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min_1, 6), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_7, 0), kwargs = {})
# %ge : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%view_7, 6), kwargs = {})
# %bitwise_or : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le, %ge), kwargs = {})
triton_poi_fused_hardtanh_hardtanh_backward_3 = async_compile.triton('triton_poi_fused_hardtanh_hardtanh_backward_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_hardtanh_hardtanh_backward_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x3 = (xindex // 1600)
x5 = xindex % 1600
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x4), tmp6, xmask)
tl.store(out_ptr1 + (x5 + (1664*x3)), tmp9, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (200, 4), (4, 1))
assert_size_stride(primals_2, (200, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 200), (200, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 200), (200, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (100, 4), (4, 1))
assert_size_stride(primals_9, (100, ), (1, ))
assert_size_stride(primals_10, (1, 100), (100, 1))
assert_size_stride(primals_11, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 200), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1), torch.float32)
buf11 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1), torch.bool)
# Topologically Sorted Source Nodes: [a1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_hardtanh_hardtanh_backward_0.run(buf0, primals_2, buf1, buf11, 12800, grid=grid(12800), stream=stream0)
del buf0
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 200), (200, 1), 0), reinterpret_tensor(primals_4, (200, 4), (1, 200), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tanh, mu], Original ATen: [aten.tanh, aten.mul]
triton_poi_fused_mul_tanh_1.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf1, (64, 200), (200, 1), 0), reinterpret_tensor(primals_6, (200, 4), (1, 200), 0), alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softplus, sigma], Original ATen: [aten.softplus, aten.add]
triton_poi_fused_add_softplus_2.run(buf4, buf5, 256, grid=grid(256), stream=stream0)
buf6 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 100), (1, 4), 0), out=buf6)
del primals_8
buf7 = empty_strided_cuda((4, 4, 4, 100), (1600, 400, 100, 1), torch.float32)
buf10 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1), torch.bool)
# Topologically Sorted Source Nodes: [c1], Original ATen: [aten.hardtanh, aten.hardtanh_backward]
triton_poi_fused_hardtanh_hardtanh_backward_3.run(buf6, primals_9, buf7, buf10, 6400, grid=grid(6400), stream=stream0)
del buf6
del primals_9
buf9 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [values], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, reinterpret_tensor(buf7, (64, 100), (100, 1), 0), reinterpret_tensor(primals_10, (100, 1), (1, 100), 0), alpha=1, beta=1, out=buf9)
del primals_11
return (buf3, buf5, reinterpret_tensor(buf9, (4, 4, 4, 1), (16, 4, 1, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 200), (200, 1), 0), buf2, buf4, reinterpret_tensor(buf7, (64, 100), (100, 1), 0), primals_10, buf10, primals_6, primals_4, buf11, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((200, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((200, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 200), (200, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 200), (200, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((100, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((100, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((1, 100), (100, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class ContinuousNet(nn.Module):
def __init__(self, s_dim, a_dim):
super(ContinuousNet, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.a1 = nn.Linear(s_dim, 200)
self.mu = nn.Linear(200, a_dim)
self.sigma = nn.Linear(200, a_dim)
self.c1 = nn.Linear(s_dim, 100)
self.v = nn.Linear(100, 1)
set_init([self.a1, self.mu, self.sigma, self.c1, self.v])
self.distribution = torch.distributions.Normal
def forward(self, x):
a1 = F.relu6(self.a1(x))
mu = 2 * F.tanh(self.mu(a1))
sigma = F.softplus(self.sigma(a1)) + 0.001
c1 = F.relu6(self.c1(x))
values = self.v(c1)
return mu, sigma, values
def choose_action(self, s):
self.training = False
mu, sigma, _ = self.forward(s)
m = self.distribution(mu.view(1).data, sigma.view(1).data)
return m.sample().numpy()
def loss_func(self, s, a, v_t):
self.train()
mu, sigma, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
m = self.distribution(mu, sigma)
log_prob = m.log_prob(a)
entropy = 0.5 + 0.5 * math.log(2 * math.pi) + torch.log(m.scale)
exp_v = log_prob * td.detach() + 0.005 * entropy
a_loss = -exp_v
total_loss = (a_loss + c_loss).mean()
return total_loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'s_dim': 4, 'a_dim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused_mul_tanh_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = libdevice.tanh(tmp0)
tmp2 = 2.0
tmp3 = tmp1 * tmp2
tl.store(out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_poi_fused_add_softplus_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 20.0
tmp2 = tmp0 > tmp1
tmp3 = tl_math.exp(tmp0)
tmp4 = libdevice.log1p(tmp3)
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = 0.001
tmp7 = tmp5 + tmp6
tl.store(out_ptr0 + x0, tmp7, xmask)
@triton.jit
def triton_poi_fused_hardtanh_hardtanh_backward_3(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x3 = xindex // 1600
x5 = xindex % 1600
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x4, tmp6, xmask)
tl.store(out_ptr1 + (x5 + 1664 * x3), tmp9, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (200, 4), (4, 1))
assert_size_stride(primals_2, (200,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 200), (200, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 200), (200, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (100, 4), (4, 1))
assert_size_stride(primals_9, (100,), (1,))
assert_size_stride(primals_10, (1, 100), (100, 1))
assert_size_stride(primals_11, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 200), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1),
torch.float32)
buf11 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_hardtanh_hardtanh_backward_0[grid(12800)](buf0,
primals_2, buf1, buf11, 12800, XBLOCK=256, num_warps=4,
num_stages=1)
del buf0
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 200),
(200, 1), 0), reinterpret_tensor(primals_4, (200, 4), (1, 200),
0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_mul_tanh_1[grid(256)](buf2, buf3, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf1, (64, 200),
(200, 1), 0), reinterpret_tensor(primals_6, (200, 4), (1, 200),
0), alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_softplus_2[grid(256)](buf4, buf5, 256, XBLOCK=
128, num_warps=4, num_stages=1)
buf6 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 100), (1, 4), 0), out=buf6)
del primals_8
buf7 = empty_strided_cuda((4, 4, 4, 100), (1600, 400, 100, 1),
torch.float32)
buf10 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1),
torch.bool)
triton_poi_fused_hardtanh_hardtanh_backward_3[grid(6400)](buf6,
primals_9, buf7, buf10, 6400, XBLOCK=256, num_warps=4, num_stages=1
)
del buf6
del primals_9
buf9 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_11, reinterpret_tensor(buf7, (64, 100),
(100, 1), 0), reinterpret_tensor(primals_10, (100, 1), (1, 100),
0), alpha=1, beta=1, out=buf9)
del primals_11
return buf3, buf5, reinterpret_tensor(buf9, (4, 4, 4, 1), (16, 4, 1, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 200), (200, 1), 0
), buf2, buf4, reinterpret_tensor(buf7, (64, 100), (100, 1), 0
), primals_10, buf10, primals_6, primals_4, buf11
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class ContinuousNetNew(nn.Module):
def __init__(self, s_dim, a_dim):
super(ContinuousNetNew, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.a1 = nn.Linear(s_dim, 200)
self.mu = nn.Linear(200, a_dim)
self.sigma = nn.Linear(200, a_dim)
self.c1 = nn.Linear(s_dim, 100)
self.v = nn.Linear(100, 1)
set_init([self.a1, self.mu, self.sigma, self.c1, self.v])
self.distribution = torch.distributions.Normal
def choose_action(self, s):
self.training = False
mu, sigma, _ = self.forward(s)
m = self.distribution(mu.view(1).data, sigma.view(1).data)
return m.sample().numpy()
def loss_func(self, s, a, v_t):
self.train()
mu, sigma, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
m = self.distribution(mu, sigma)
log_prob = m.log_prob(a)
entropy = 0.5 + 0.5 * math.log(2 * math.pi) + torch.log(m.scale)
exp_v = log_prob * td.detach() + 0.005 * entropy
a_loss = -exp_v
total_loss = (a_loss + c_loss).mean()
return total_loss
def forward(self, input_0):
primals_1 = self.a1.weight
primals_2 = self.a1.bias
primals_4 = self.mu.weight
primals_5 = self.mu.bias
primals_6 = self.sigma.weight
primals_7 = self.sigma.bias
primals_8 = self.c1.weight
primals_9 = self.c1.bias
primals_10 = self.v.weight
primals_11 = self.v.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0], output[1], output[2]
|
lws803/pytorch-A3C
|
ContinuousNet
| false | 10,467 |
[
"MIT"
] | 0 |
944e7f42a8fa54b7d6efbe169d8a3467b20a0f7f
|
https://github.com/lws803/pytorch-A3C/tree/944e7f42a8fa54b7d6efbe169d8a3467b20a0f7f
|
FullAttention
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/ay/caylcn737p2wwjm32cacv462xdgdut6ho32ptwxfu34t3i2tr75z.py
# Topologically Sorted Source Nodes: [QK], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# QK => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x3 = (xindex // 64)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2) + (16*x1) + (64*x3)), xmask)
tl.store(out_ptr0 + (x4), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ri/cricgdtr5c24l63g746gjtdd45qor3pkzmi7qmyygyd24ejrijb7.py
# Topologically Sorted Source Nodes: [QK], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# QK => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 16
y1 = (yindex // 16)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x2) + (64*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ya/cyacjmbos4hqjbfqrids3ws7umxa2yjx6uocj6nk4q5qb3ifqsdu.py
# Topologically Sorted Source Nodes: [mul, A], Original ATen: [aten.mul, aten._softmax]
# Source node to ATen node mapping:
# A => amax, clone_2, exp, sub
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, 0.5), kwargs = {})
# %clone_2 : [num_users=2] = call_function[target=torch.ops.aten.clone.default](args = (%mul,), kwargs = {memory_format: torch.contiguous_format})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%clone_2, [2], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%clone_2, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_mul_2 = async_compile.triton('triton_poi_fused__softmax_mul_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_mul_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_mul_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tl_math.exp(tmp14)
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/vb/cvb3oguu73uopye5dab6pedxgpbqv7hv76dtm7n5cat2xumuixgo.py
# Topologically Sorted Source Nodes: [queried_values], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# queried_values => clone_3
# Graph fragment:
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [QK], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(arg1_1, buf0, 256, grid=grid(256), stream=stream0)
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [QK], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(arg0_1, buf1, 64, 4, grid=grid(64, 4), stream=stream0)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [QK], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0), out=buf2)
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 4, 1, 16), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [mul, A], Original ATen: [aten.mul, aten._softmax]
triton_poi_fused__softmax_mul_2.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [queried_values], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [queried_values], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(arg2_1, buf5, 256, grid=grid(256), stream=stream0)
del arg2_1
buf6 = reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [queried_values], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1), 0), out=buf6)
del buf4
buf7 = reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
return (buf7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
from torch.nn import Dropout
class FullAttention(Module):
def __init__(self, use_dropout=False, attention_dropout=0.1):
super().__init__()
self.use_dropout = use_dropout
self.dropout = Dropout(attention_dropout)
def forward(self, queries, keys, values, q_mask=None, kv_mask=None):
""" Multi-head scaled dot-product attention, a.k.a full attention.
Args:
queries: [N, L, H, D]
keys: [N, S, H, D]
values: [N, S, H, D]
q_mask: [N, L]
kv_mask: [N, S]
Returns:
queried_values: (N, L, H, D)
"""
QK = torch.einsum('nlhd,nshd->nlsh', queries, keys)
if kv_mask is not None:
QK.masked_fill_(~(q_mask[:, :, None, None] * kv_mask[:, None, :,
None]), float('-inf'))
softmax_temp = 1.0 / queries.size(3) ** 0.5
A = torch.softmax(softmax_temp * QK, dim=2)
if self.use_dropout:
A = self.dropout(A)
queried_values = torch.einsum('nlsh,nshd->nlhd', A, values)
return queried_values.contiguous()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.nn import Module
from torch.nn import Dropout
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x3 = xindex // 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), xmask)
tl.store(out_ptr0 + x4, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 16
y1 = yindex // 16
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x2 + 64 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_mul_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tl_math.exp(tmp14)
tl.store(out_ptr0 + x2, tmp15, xmask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(256)](arg1_1, buf0, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
triton_poi_fused_clone_1[grid(64, 4)](arg0_1, buf1, 64, 4, XBLOCK=4,
YBLOCK=32, num_warps=4, num_stages=1)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0), out=buf2)
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 4, 1, 16), 0)
del buf1
triton_poi_fused__softmax_mul_2[grid(256)](buf2, buf3, 256, XBLOCK=
256, num_warps=4, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0)
del buf2
triton_poi_fused_clone_3[grid(256)](buf3, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0)
del buf3
triton_poi_fused_clone_0[grid(256)](arg2_1, buf5, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg2_1
buf6 = reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0)
del buf0
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1), 0), out=buf6)
del buf4
buf7 = reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf5
triton_poi_fused_clone_0[grid(256)](buf6, buf7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf6
return buf7,
class FullAttentionNew(Module):
def __init__(self, use_dropout=False, attention_dropout=0.1):
super().__init__()
self.use_dropout = use_dropout
self.dropout = Dropout(attention_dropout)
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0]
|
lee-vius/LoFTR
|
FullAttention
| false | 10,468 |
[
"Apache-2.0"
] | 0 |
dd9add373a20696fb6f020f4fda38bca7a91cdd9
|
https://github.com/lee-vius/LoFTR/tree/dd9add373a20696fb6f020f4fda38bca7a91cdd9
|
LRN
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/gv/cgvon7iygyhjm2kmwtta5t2r3z2byfrr4qwpcmym3h4h6yzxvtvp.py
# Topologically Sorted Source Nodes: [div, div_1, mul, add, div_2, x], Original ATen: [aten.pow, aten.avg_pool2d, aten.mul, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# div => pow_1
# div_1 => avg_pool2d
# div_2 => pow_2
# mul => mul
# x => div
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 2), kwargs = {})
# %avg_pool2d : [num_users=1] = call_function[target=torch.ops.aten.avg_pool2d.default](args = (%pow_1, [1, 1], [1, 1]), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%avg_pool2d, 0.0001), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 2.0), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add, 0.75), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %pow_2), kwargs = {})
triton_poi_fused_add_avg_pool2d_div_mul_pow_0 = async_compile.triton('triton_poi_fused_add_avg_pool2d_div_mul_pow_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_avg_pool2d_div_mul_pow_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_avg_pool2d_div_mul_pow_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tmp0 * tmp0
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0001
tmp5 = tmp3 * tmp4
tmp6 = 2.0
tmp7 = tmp5 + tmp6
tmp8 = 0.75
tmp9 = libdevice.pow(tmp7, tmp8)
tmp10 = tmp0 / tmp9
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [div, div_1, mul, add, div_2, x], Original ATen: [aten.pow, aten.avg_pool2d, aten.mul, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_avg_pool2d_div_mul_pow_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class LRN(nn.Module):
def __init__(self, local_size=1, alpha=0.0001, beta=0.75,
ACROSS_CHANNELS=False):
super(LRN, self).__init__()
self.ACROSS_CHANNELS = ACROSS_CHANNELS
if self.ACROSS_CHANNELS:
self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
stride=1, padding=(int((local_size - 1.0) / 2), 0, 0))
else:
self.average = nn.AvgPool2d(kernel_size=local_size, stride=1,
padding=int((local_size - 1.0) / 2))
self.alpha = alpha
self.beta = beta
def forward(self, x):
if self.ACROSS_CHANNELS:
div = x.pow(2).unsqueeze(1)
div = self.average(div).squeeze(1)
div = div.mul(self.alpha).add(2.0).pow(self.beta)
else:
div = x.pow(2)
div = self.average(div)
div = div.mul(self.alpha).add(2.0).pow(self.beta)
x = x.div(div)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_avg_pool2d_div_mul_pow_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tmp0 * tmp0
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0001
tmp5 = tmp3 * tmp4
tmp6 = 2.0
tmp7 = tmp5 + tmp6
tmp8 = 0.75
tmp9 = libdevice.pow(tmp7, tmp8)
tmp10 = tmp0 / tmp9
tl.store(out_ptr0 + x0, tmp10, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_avg_pool2d_div_mul_pow_0[grid(256)](arg0_1,
buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class LRNNew(nn.Module):
def __init__(self, local_size=1, alpha=0.0001, beta=0.75,
ACROSS_CHANNELS=False):
super(LRNNew, self).__init__()
self.ACROSS_CHANNELS = ACROSS_CHANNELS
if self.ACROSS_CHANNELS:
self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
stride=1, padding=(int((local_size - 1.0) / 2), 0, 0))
else:
self.average = nn.AvgPool2d(kernel_size=local_size, stride=1,
padding=int((local_size - 1.0) / 2))
self.alpha = alpha
self.beta = beta
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
melster1010/VIAME
|
LRN
| false | 10,469 |
[
"BSD-3-Clause"
] | 0 |
0062265088aae65effbfcd130bfb874c343c785f
|
https://github.com/melster1010/VIAME/tree/0062265088aae65effbfcd130bfb874c343c785f
|
LinearAttention
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/og/cogo4pfcy56l7ulvsx6und2vmih3qojwpvkiazyzh7hsptmp4vos.py
# Topologically Sorted Source Nodes: [elu_1, K, sum_1], Original ATen: [aten.elu, aten.add, aten.sum]
# Source node to ATen node mapping:
# K => add_1
# elu_1 => expm1_1, gt_1, mul_3, mul_4, mul_5, where_1
# sum_1 => sum_1
# Graph fragment:
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%arg1_1, 0), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg1_1, 1.0), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg1_1, 1.0), kwargs = {})
# %expm1_1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul_4,), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1_1, 1.0), kwargs = {})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt_1, %mul_3, %mul_5), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%where_1, 1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_1, [1]), kwargs = {})
triton_poi_fused_add_elu_sum_0 = async_compile.triton('triton_poi_fused_add_elu_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_elu_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_elu_sum_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp25 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tmp10 = tmp9 > tmp1
tmp11 = tmp9 * tmp3
tmp12 = libdevice.expm1(tmp11)
tmp13 = tmp12 * tmp3
tmp14 = tl.where(tmp10, tmp11, tmp13)
tmp15 = tmp14 + tmp3
tmp16 = tmp8 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = libdevice.expm1(tmp19)
tmp21 = tmp20 * tmp3
tmp22 = tl.where(tmp18, tmp19, tmp21)
tmp23 = tmp22 + tmp3
tmp24 = tmp16 + tmp23
tmp26 = tmp25 > tmp1
tmp27 = tmp25 * tmp3
tmp28 = libdevice.expm1(tmp27)
tmp29 = tmp28 * tmp3
tmp30 = tl.where(tmp26, tmp27, tmp29)
tmp31 = tmp30 + tmp3
tmp32 = tmp24 + tmp31
tl.store(out_ptr0 + (x2), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ui/cuicp7lcen5tosk4olu3x3oaa7yz6cvucka5zwp4hr2qbvjrvvxh.py
# Topologically Sorted Source Nodes: [einsum_1, einsum_2], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# einsum_1 => clone_2
# einsum_2 => clone_3
# Graph fragment:
# %clone_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
# %clone_3 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_13,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x3 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tl.store(out_ptr0 + (x0 + (4*x2) + (16*x1) + (64*x3)), tmp8, xmask)
tl.store(out_ptr1 + (x0 + (4*x2) + (16*x1) + (64*x3)), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ds/cdscs66baluy2qjp2egcmspwpssho6a6gwzskgkoo4xnrapvrluw.py
# Topologically Sorted Source Nodes: [KV], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# KV => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_2 = async_compile.triton('triton_poi_fused_clone_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 4
y1 = (yindex // 4)
tmp0 = tl.load(in_ptr0 + (x2 + (16*y3)), xmask & ymask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tl.store(out_ptr0 + (y0 + (4*x2) + (64*y1)), tmp8, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/mf/cmffnm5icoqilxjyfgljttokrgpbt5d2oz7xydx6dxdlysmdthuj.py
# Topologically Sorted Source Nodes: [KV], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# KV => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_3 = async_compile.triton('triton_poi_fused_clone_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x3 = (xindex // 64)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2) + (16*x1) + (64*x3)), xmask)
tmp1 = 0.25
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x4), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/kd/ckd25s7vujyv5hxcsgrxyfy25zilz32mxdkf3yx2dqm3ovvchilg.py
# Topologically Sorted Source Nodes: [queried_values, contiguous], Original ATen: [aten.mul, aten.clone]
# Source node to ATen node mapping:
# contiguous => clone_4
# queried_values => mul_8
# Graph fragment:
# %mul_8 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_11, 4), kwargs = {})
# %clone_4 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%mul_8,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_mul_4 = async_compile.triton('triton_poi_fused_clone_mul_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_mul_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_mul_4(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = (xindex // 4)
x5 = xindex
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x3 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x4), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (x5), xmask)
tmp1 = 1e-06
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 1, tl.int32)
tmp4 = tmp3 / tmp2
tmp5 = 1.0
tmp6 = tmp4 * tmp5
tmp8 = tmp6 * tmp7
tmp9 = 4.0
tmp10 = tmp8 * tmp9
tl.store(out_ptr0 + (x0 + (4*x2) + (16*x1) + (64*x3)), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [elu_1, K, sum_1], Original ATen: [aten.elu, aten.add, aten.sum]
stream0 = get_raw_stream(0)
triton_poi_fused_add_elu_sum_0.run(arg1_1, buf0, 64, grid=grid(64), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [einsum_1, einsum_2], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(arg0_1, buf1, buf6, 256, grid=grid(256), stream=stream0)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 1), (4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [einsum_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf0, (16, 4, 1), (4, 1, 1), 0), out=buf2)
del buf0
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [KV], Original ATen: [aten.clone]
triton_poi_fused_clone_2.run(arg1_1, buf3, 16, 16, grid=grid(16, 16), stream=stream0)
del arg1_1
buf4 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [KV], Original ATen: [aten.clone]
triton_poi_fused_clone_3.run(arg2_1, buf4, 256, grid=grid(256), stream=stream0)
del arg2_1
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [KV], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0), out=buf5)
del buf3
buf7 = reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [einsum_2], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1), 0), buf5, out=buf7)
del buf5
buf8 = reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf6 # reuse
# Topologically Sorted Source Nodes: [queried_values, contiguous], Original ATen: [aten.mul, aten.clone]
triton_poi_fused_clone_mul_4.run(buf2, buf7, buf8, 256, grid=grid(256), stream=stream0)
del buf2
del buf7
return (buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from torch.nn import Module
import torch
def elu_feature_map(x):
return torch.nn.functional.elu(x) + 1
class LinearAttention(Module):
def __init__(self, eps=1e-06):
super().__init__()
self.feature_map = elu_feature_map
self.eps = eps
def forward(self, queries, keys, values, q_mask=None, kv_mask=None):
""" Multi-Head linear attention proposed in "Transformers are RNNs"
Args:
queries: [N, L, H, D]
keys: [N, S, H, D]
values: [N, S, H, D]
q_mask: [N, L]
kv_mask: [N, S]
Returns:
queried_values: (N, L, H, D)
"""
Q = self.feature_map(queries)
K = self.feature_map(keys)
if q_mask is not None:
Q = Q * q_mask[:, :, None, None]
if kv_mask is not None:
K = K * kv_mask[:, :, None, None]
values = values * kv_mask[:, :, None, None]
v_length = values.size(1)
values = values / v_length
KV = torch.einsum('nshd,nshv->nhdv', K, values)
Z = 1 / (torch.einsum('nlhd,nhd->nlh', Q, K.sum(dim=1)) + self.eps)
queried_values = torch.einsum('nlhd,nhdv,nlh->nlhv', Q, KV, Z
) * v_length
return queried_values.contiguous()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch.nn import Module
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_elu_sum_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp25 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tmp10 = tmp9 > tmp1
tmp11 = tmp9 * tmp3
tmp12 = libdevice.expm1(tmp11)
tmp13 = tmp12 * tmp3
tmp14 = tl.where(tmp10, tmp11, tmp13)
tmp15 = tmp14 + tmp3
tmp16 = tmp8 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = libdevice.expm1(tmp19)
tmp21 = tmp20 * tmp3
tmp22 = tl.where(tmp18, tmp19, tmp21)
tmp23 = tmp22 + tmp3
tmp24 = tmp16 + tmp23
tmp26 = tmp25 > tmp1
tmp27 = tmp25 * tmp3
tmp28 = libdevice.expm1(tmp27)
tmp29 = tmp28 * tmp3
tmp30 = tl.where(tmp26, tmp27, tmp29)
tmp31 = tmp30 + tmp3
tmp32 = tmp24 + tmp31
tl.store(out_ptr0 + x2, tmp32, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x3 = xindex // 64
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tl.store(out_ptr0 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), tmp8, xmask)
tl.store(out_ptr1 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), tmp8, xmask)
@triton.jit
def triton_poi_fused_clone_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 16
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 4
y1 = yindex // 4
tmp0 = tl.load(in_ptr0 + (x2 + 16 * y3), xmask & ymask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = tmp7 + tmp3
tl.store(out_ptr0 + (y0 + 4 * x2 + 64 * y1), tmp8, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x3 = xindex // 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), xmask)
tmp1 = 0.25
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x4, tmp2, xmask)
@triton.jit
def triton_poi_fused_clone_mul_4(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex // 4
x5 = xindex
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x3 = xindex // 64
tmp0 = tl.load(in_ptr0 + x4, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + x5, xmask)
tmp1 = 1e-06
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 1, tl.int32)
tmp4 = tmp3 / tmp2
tmp5 = 1.0
tmp6 = tmp4 * tmp5
tmp8 = tmp6 * tmp7
tmp9 = 4.0
tmp10 = tmp8 * tmp9
tl.store(out_ptr0 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), tmp10, xmask)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_elu_sum_0[grid(64)](arg1_1, buf0, 64, XBLOCK=
64, num_warps=1, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
triton_poi_fused_clone_1[grid(256)](arg0_1, buf1, buf6, 256, XBLOCK
=256, num_warps=4, num_stages=1)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 1), (4, 1, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf0, (16, 4, 1), (4, 1, 1), 0), out=buf2)
del buf0
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4, 1), (64, 16, 4, 1, 1), 0)
del buf1
triton_poi_fused_clone_2[grid(16, 16)](arg1_1, buf3, 16, 16, XBLOCK
=16, YBLOCK=16, num_warps=4, num_stages=1)
del arg1_1
buf4 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
triton_poi_fused_clone_3[grid(256)](arg2_1, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg2_1
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0), out=buf5)
del buf3
buf7 = reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0)
del buf4
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1),
0), buf5, out=buf7)
del buf5
buf8 = reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf6
triton_poi_fused_clone_mul_4[grid(256)](buf2, buf7, buf8, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del buf2
del buf7
return buf8,
def elu_feature_map(x):
return torch.nn.functional.elu(x) + 1
class LinearAttentionNew(Module):
def __init__(self, eps=1e-06):
super().__init__()
self.feature_map = elu_feature_map
self.eps = eps
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0]
|
lee-vius/LoFTR
|
LinearAttention
| false | 10,470 |
[
"Apache-2.0"
] | 0 |
dd9add373a20696fb6f020f4fda38bca7a91cdd9
|
https://github.com/lee-vius/LoFTR/tree/dd9add373a20696fb6f020f4fda38bca7a91cdd9
|
Descendant
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yk/cykarwk3qy3pmem22bdbqxnd35nfbs5r3u5357r5uoderrncltwa.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[524288],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 288000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 3600) % 20
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/pr/cprkqcgqlxorbpnularzqdmik33w6yjgt5sc4pc7na35izz6l3qp.py
# Topologically Sorted Source Nodes: [max_pool2d, conv_out], Original ATen: [aten.max_pool2d_with_indices, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# conv_out => relu
# max_pool2d => _low_memory_max_pool2d_with_offsets, getitem_1
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[131072],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1(in_ptr0, out_ptr0, out_ptr1, out_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 72000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = (xindex // 30)
x2 = (xindex // 18000)
x4 = xindex % 18000
x5 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp19 = 0.0
tmp20 = tmp18 <= tmp19
tl.store(out_ptr0 + (x4 + (18048*x2)), tmp15, xmask)
tl.store(out_ptr1 + (x5), tmp18, xmask)
tl.store(out_ptr2 + (x4 + (18048*x2)), tmp20, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (20, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (20, ), (1, ))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 20, 60, 60), (72000, 3600, 60, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_2, 288000, grid=grid(288000), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1), torch.int8)
buf3 = empty_strided_cuda((4, 20, 30, 30), (18000, 900, 30, 1), torch.float32)
buf4 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1), torch.bool)
# Topologically Sorted Source Nodes: [max_pool2d, conv_out], Original ATen: [aten.max_pool2d_with_indices, aten.relu, aten.threshold_backward]
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1.run(buf1, buf2, buf3, buf4, 72000, grid=grid(72000), stream=stream0)
return (buf3, primals_1, primals_3, buf1, buf2, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((20, 1, 5, 5), (25, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((20, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 1, 64, 64), (4096, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
import torch.nn.functional as F
class Descendant(nn.Module):
"""Descendant descendantEncoder model for ADDA."""
def __init__(self):
"""Init Descendant descendantEncoder."""
super(Descendant, self).__init__()
self.restored = False
self.conv1 = nn.Conv2d(1, 20, kernel_size=5)
self.pool1 = nn.MaxPool2d(kernel_size=2)
def forward(self, input):
"""Forward the Descendant."""
conv_out = F.relu(self.pool1(self.conv1(input)))
out = conv_out
return out
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 288000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 3600 % 20
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1(in_ptr0,
out_ptr0, out_ptr1, out_ptr2, xnumel, XBLOCK: tl.constexpr):
xnumel = 72000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = xindex // 30
x2 = xindex // 18000
x4 = xindex % 18000
x5 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 120 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp19 = 0.0
tmp20 = tmp18 <= tmp19
tl.store(out_ptr0 + (x4 + 18048 * x2), tmp15, xmask)
tl.store(out_ptr1 + x5, tmp18, xmask)
tl.store(out_ptr2 + (x4 + 18048 * x2), tmp20, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (20, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (20,), (1,))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 20, 60, 60), (72000, 3600, 60, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(288000)](buf1, primals_2,
288000, XBLOCK=512, num_warps=8, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1),
torch.int8)
buf3 = empty_strided_cuda((4, 20, 30, 30), (18000, 900, 30, 1),
torch.float32)
buf4 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1),
torch.bool)
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_1[grid
(72000)](buf1, buf2, buf3, buf4, 72000, XBLOCK=512, num_warps=8,
num_stages=1)
return buf3, primals_1, primals_3, buf1, buf2, buf4
class DescendantNew(nn.Module):
"""Descendant descendantEncoder model for ADDA."""
def __init__(self):
"""Init Descendant descendantEncoder."""
super(DescendantNew, self).__init__()
self.restored = False
self.conv1 = nn.Conv2d(1, 20, kernel_size=5)
self.pool1 = nn.MaxPool2d(kernel_size=2)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
lindagaw/Kadara
|
Descendant
| false | 10,471 |
[
"MIT"
] | 0 |
f1059b69a581344ca460c8df02ac3f73f3fbcba1
|
https://github.com/lindagaw/Kadara/tree/f1059b69a581344ca460c8df02ac3f73f3fbcba1
|
Block
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/xx/cxxsy7adcakcv5c6imnaqsvf3ebhgbph77vaoembzakbqdrjycbc.py
# Topologically Sorted Source Nodes: [mean, std, sub, mul, add, truediv, x], Original ATen: [aten.mean, aten.std, aten.sub, aten.mul, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# mean => mean
# mul => mul
# std => sqrt, var
# sub => sub
# truediv => div
# x => add_1
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [-1], True), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%primals_1, [-1]), kwargs = {correction: 1.0, keepdim: True})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %sub), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sqrt, 1e-06), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul, %add), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%div, %primals_3), kwargs = {})
triton_poi_fused_add_div_mean_mul_std_sub_0 = async_compile.triton('triton_poi_fused_add_div_mean_mul_std_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_mean_mul_std_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_mean_mul_std_sub_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp8 = tmp6 + tmp7
tmp9 = 4.0
tmp10 = tmp8 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp0 * tmp11
tmp13 = tmp2 - tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp3 - tmp10
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp18 = tmp5 - tmp10
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp21 = tmp7 - tmp10
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = 3.0
tmp25 = tmp23 / tmp24
tmp26 = libdevice.sqrt(tmp25)
tmp27 = 1e-06
tmp28 = tmp26 + tmp27
tmp29 = tmp12 / tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + (x2), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/zd/czdzwoe4pjzwt5uqsqnvjev2fqbcqpnioydhfcnlbkgk527hzabw.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %mul_scalar : [num_users=1] = call_function[target=torch.ops.aten.mul.Scalar](args = (%permute_default, 1.0), kwargs = {})
# %clone_default : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_1 = async_compile.triton('triton_poi_fused_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + (4*y3)), tmp4, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3o/c3oaxjlvvp4afd7ykuqeeshinhwcjxyjj3abqkwr3yqqinwm4e4f.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_default_2, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_default_2, %amax_default), kwargs = {})
# %exp_default : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_tensor,), kwargs = {})
triton_poi_fused_2 = async_compile.triton('triton_poi_fused_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2w/c2wy7nnmjg5t737ww7jovvpbthc2goznalidqgo6rnve5cme7k3g.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %sum_dim_int_list : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_default, [-1], True), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_default, %sum_dim_int_list), kwargs = {})
# %eq_scalar : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%view_default_2, -inf), kwargs = {})
# %logical_not_default : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%eq_scalar,), kwargs = {})
# %any_dim : [num_users=1] = call_function[target=torch.ops.aten.any.dim](args = (%logical_not_default, -1, True), kwargs = {})
# %logical_not_default_1 : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%any_dim,), kwargs = {})
# %full_default_2 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_self : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%logical_not_default_1, %full_default_2, %div_tensor), kwargs = {})
triton_poi_fused_3 = async_compile.triton('triton_poi_fused_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr1 + (x2), xmask)
tmp26 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = float("-inf")
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = (tmp4 != 0)
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = (tmp9 != 0)
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = (tmp15 != 0)
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = (tmp21 != 0)
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + (x2), tmp35, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/cc/ccc5cubvwrrtlmcdkkaw7lrpubmmsow2eocc44vbuqtnlv363xv3.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %clone_default_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_4 = async_compile.triton('triton_poi_fused_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_4(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + (4*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/we/cwe54p4p4jvwbdktkpj3wy2coheu6f3r3dgvi7ozm7xjfk4mgbwx.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone_4
# Graph fragment:
# %clone_4 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_5 = async_compile.triton('triton_poi_fused_clone_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_5(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/c3/cc3wjckmv52z5p6lagnrhsfwt53rzdfhvzlxkm5tgkwbs3kuzwax.py
# Topologically Sorted Source Nodes: [res_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# res_1 => add_2
# Graph fragment:
# %add_2 : [num_users=5] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %add_1), kwargs = {})
triton_poi_fused_add_6 = async_compile.triton('triton_poi_fused_add_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_6', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_6(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/nr/cnrllyxkh5tmxzmndufu4v6arhz3nvwpyestxuypvazb6covpgb6.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_19,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_7 = async_compile.triton('triton_poi_fused_relu_threshold_backward_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_7', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_7(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/he/chevf4d6tadiz3y2a2abr2lj2bvo3wyfykoivwj2s4xedp3vdjuf.py
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# out_1 => add_5
# Graph fragment:
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_2, %view_21), kwargs = {})
triton_poi_fused_add_8 = async_compile.triton('triton_poi_fused_add_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_8', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_8(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_out_ptr0 + (x2), xmask)
tmp2 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tmp0 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
assert_size_stride(primals_12, (4, ), (1, ))
assert_size_stride(primals_13, (4, ), (1, ))
assert_size_stride(primals_14, (16, 4), (4, 1))
assert_size_stride(primals_15, (16, ), (1, ))
assert_size_stride(primals_16, (4, 16), (16, 1))
assert_size_stride(primals_17, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, std, sub, mul, add, truediv, x], Original ATen: [aten.mean, aten.std, aten.sub, aten.mul, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_mean_mul_std_sub_0.run(primals_2, primals_1, primals_3, buf0, 64, grid=grid(64), stream=stream0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(buf1, primals_5, buf4, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_5
buf5 = reinterpret_tensor(buf1, (4, 4, 1, 4), (16, 4, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(buf2, primals_7, buf5, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_7
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_3.run(buf6, buf7, buf8, 256, grid=grid(256), stream=stream0)
del buf6
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_4.run(buf3, primals_9, buf9, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_9
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.bmm(reinterpret_tensor(buf8, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf9, (16, 4, 1), (4, 1, 0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_5.run(buf10, buf11, 16, 4, grid=grid(16, 4), stream=stream0)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0); del buf10 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0); del buf12 # reuse
# Topologically Sorted Source Nodes: [res_1], Original ATen: [aten.add]
triton_poi_fused_add_6.run(buf13, primals_11, buf0, 64, grid=grid(64), stream=stream0)
del primals_11
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean_2, std_2, sub_1, mul_1, add_3, truediv_2, add_4], Original ATen: [aten.mean, aten.std, aten.sub, aten.mul, aten.add, aten.div]
triton_poi_fused_add_div_mean_mul_std_sub_0.run(primals_12, buf13, primals_13, buf14, 64, grid=grid(64), stream=stream0)
del primals_13
buf15 = reinterpret_tensor(buf7, (16, 16), (16, 1), 0); del buf7 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf14, (16, 4), (4, 1), 0), reinterpret_tensor(primals_14, (4, 16), (1, 4), 0), out=buf15)
buf16 = reinterpret_tensor(buf15, (4, 4, 16), (64, 16, 1), 0); del buf15 # reuse
buf19 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_7.run(buf16, primals_15, buf19, 256, grid=grid(256), stream=stream0)
del primals_15
buf17 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf16, (16, 16), (16, 1), 0), reinterpret_tensor(primals_16, (16, 4), (1, 16), 0), out=buf17)
buf18 = reinterpret_tensor(buf17, (4, 4, 4), (16, 4, 1), 0); del buf17 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.add]
triton_poi_fused_add_8.run(buf18, buf13, primals_17, 64, grid=grid(64), stream=stream0)
del primals_17
return (buf18, primals_1, primals_12, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf8, reinterpret_tensor(buf9, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 4), 0), reinterpret_tensor(buf11, (16, 4), (4, 1), 0), buf13, reinterpret_tensor(buf14, (16, 4), (4, 1), 0), reinterpret_tensor(buf16, (16, 16), (16, 1), 0), primals_16, buf19, primals_14, primals_10, primals_8, primals_6, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((4, 16), (16, 1), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
from typing import Optional
import torch.nn.functional as F
from torch import nn
def attention(query, key, value, mask=None, dropout=None):
"""Compute 'Scaled Dot Product Attention'
"""
d_k = query.size(-1)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(d_k)
if mask is not None:
scores = scores.masked_fill(mask == 0, -1000000000.0)
weights = F.softmax(scores, dim=-1)
if dropout is not None:
weights = dropout(weights)
out = torch.matmul(weights, value)
return out, weights
class LayerNorm(nn.Module):
"""Construct a layernorm module (See citation for details)."""
def __init__(self, features: 'int', eps: 'float'=1e-06):
super(LayerNorm, self).__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x: 'torch.Tensor'):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class MultiHeadedAttention(nn.Module):
def __init__(self, h, d_model, dropout=0.1):
"""Take in model size and number of heads."""
super(MultiHeadedAttention, self).__init__()
assert d_model % h == 0
self.d_head = d_model // h
self.h = h
self.proj_q = nn.Linear(d_model, d_model)
self.proj_k = nn.Linear(d_model, d_model)
self.proj_v = nn.Linear(d_model, d_model)
self.ret_proj = nn.Linear(d_model, d_model)
self.dropout = nn.Dropout(p=dropout)
def forward(self, query: 'torch.Tensor', key: 'torch.Tensor', value:
'torch.Tensor', mask: 'Optional[torch.Tensor]'=None):
"""
query: (batch_size, seq_len, dmodel)
key: (batch_size, seq_len, dmodel)
value: (batch_size, seq_len, dmodel)
mask: (batch_size, seq_len)
"""
if mask is not None:
mask = mask.unsqueeze(1)
batch_size = query.size(0)
seq_len = query.size(1)
query = self.proj_q(query).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
key = self.proj_k(key).view(batch_size, seq_len, self.h, self.d_head
).transpose(1, 2)
value = self.proj_v(value).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
x, _ = attention(query, key, value, mask=mask, dropout=self.dropout)
x = x.transpose(1, 2).contiguous().view(batch_size, seq_len, self.h *
self.d_head)
return self.ret_proj(x)
class PositionwiseFeedForward(nn.Module):
"""Implements FFN equation."""
def __init__(self, d_model, d_ff, dropout=0.1):
super(PositionwiseFeedForward, self).__init__()
self.w_1 = nn.Linear(d_model, d_ff)
self.w_2 = nn.Linear(d_ff, d_model)
self.dropout = nn.Dropout(dropout)
def forward(self, x):
return self.w_2(self.dropout(F.relu(self.w_1(x))))
class Block(nn.Module):
"""A standard Decoder block for GPT."""
def __init__(self, d_model: 'int', n_heads: 'int', dropout: 'float'=0.1):
super(Block, self).__init__()
self.d_model = d_model
self.d_inner = 4 * self.d_model
self.n_heads = n_heads
self.dropout = dropout
self.layer_norm1 = LayerNorm(self.d_model)
self.layer_norm2 = LayerNorm(self.d_model)
self.multi_head_attn = MultiHeadedAttention(self.n_heads, self.
d_model, self.dropout)
self.feed_fwd = PositionwiseFeedForward(d_model, self.d_inner, self
.dropout)
def forward(self, x: 'torch.Tensor', mask: 'Optional[torch.Tensor]'=None):
x = self.layer_norm1(x)
attn_out = self.multi_head_attn(x, x, x, mask)
res_1 = attn_out + x
feed_fwd_out = self.feed_fwd(self.layer_norm2(res_1))
out = res_1 + feed_fwd_out
return out
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'d_model': 4, 'n_heads': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import math
from typing import Optional
import torch.nn.functional as F
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_div_mean_mul_std_sub_0(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp8 = tmp6 + tmp7
tmp9 = 4.0
tmp10 = tmp8 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp0 * tmp11
tmp13 = tmp2 - tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp3 - tmp10
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp18 = tmp5 - tmp10
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp21 = tmp7 - tmp10
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = 3.0
tmp25 = tmp23 / tmp24
tmp26 = libdevice.sqrt(tmp25)
tmp27 = 1e-06
tmp28 = tmp26 + tmp27
tmp29 = tmp12 / tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + x2, tmp31, xmask)
@triton.jit
def triton_poi_fused_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + 4 * y3), tmp4, xmask & ymask)
@triton.jit
def triton_poi_fused_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp25 = tl.load(in_ptr1 + x2, xmask)
tmp26 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp29 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp31 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = float('-inf')
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = tmp4 != 0
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = tmp9 != 0
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = tmp15 != 0
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp21 != 0
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + x2, tmp35, xmask)
@triton.jit
def triton_poi_fused_4(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + 4 * y3), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_5(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_6(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_7(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_8(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_out_ptr0 + x2, xmask)
tmp2 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tmp0 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
assert_size_stride(primals_12, (4,), (1,))
assert_size_stride(primals_13, (4,), (1,))
assert_size_stride(primals_14, (16, 4), (4, 1))
assert_size_stride(primals_15, (16,), (1,))
assert_size_stride(primals_16, (4, 16), (16, 1))
assert_size_stride(primals_17, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_mean_mul_std_sub_0[grid(64)](primals_2,
primals_1, primals_3, buf0, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_1[grid(16, 4)](buf1, primals_5, buf4, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_5
buf5 = reinterpret_tensor(buf1, (4, 4, 1, 4), (16, 4, 4, 1), 0)
del buf1
triton_poi_fused_1[grid(16, 4)](buf2, primals_7, buf5, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_7
buf6 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf5, (16, 1, 4), (4, 0, 1), 0), out=buf6)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_2[grid(256)](buf6, buf7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_3[grid(256)](buf6, buf7, buf8, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf6
buf9 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf2
triton_poi_fused_4[grid(16, 4)](buf3, primals_9, buf9, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_9
buf10 = reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 1), 0)
del buf3
extern_kernels.bmm(reinterpret_tensor(buf8, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf9, (16, 4, 1), (4, 1, 0), 0), out=buf10)
buf11 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_5[grid(16, 4)](buf10, buf11, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf12 = reinterpret_tensor(buf10, (16, 4), (4, 1), 0)
del buf10
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0)
del buf12
triton_poi_fused_add_6[grid(64)](buf13, primals_11, buf0, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_11
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_div_mean_mul_std_sub_0[grid(64)](primals_12,
buf13, primals_13, buf14, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_13
buf15 = reinterpret_tensor(buf7, (16, 16), (16, 1), 0)
del buf7
extern_kernels.mm(reinterpret_tensor(buf14, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_14, (4, 16), (1, 4), 0), out=buf15)
buf16 = reinterpret_tensor(buf15, (4, 4, 16), (64, 16, 1), 0)
del buf15
buf19 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_7[grid(256)](buf16,
primals_15, buf19, 256, XBLOCK=256, num_warps=4, num_stages=1)
del primals_15
buf17 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf16, (16, 16), (16, 1), 0),
reinterpret_tensor(primals_16, (16, 4), (1, 16), 0), out=buf17)
buf18 = reinterpret_tensor(buf17, (4, 4, 4), (16, 4, 1), 0)
del buf17
triton_poi_fused_add_8[grid(64)](buf18, buf13, primals_17, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_17
return (buf18, primals_1, primals_12, reinterpret_tensor(buf0, (16, 4),
(4, 1), 0), buf8, reinterpret_tensor(buf9, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 1, 4), (4, 1, 1), 0),
reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 4), 0),
reinterpret_tensor(buf11, (16, 4), (4, 1), 0), buf13,
reinterpret_tensor(buf14, (16, 4), (4, 1), 0), reinterpret_tensor(
buf16, (16, 16), (16, 1), 0), primals_16, buf19, primals_14,
primals_10, primals_8, primals_6, primals_4)
def attention(query, key, value, mask=None, dropout=None):
"""Compute 'Scaled Dot Product Attention'
"""
d_k = query.size(-1)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(d_k)
if mask is not None:
scores = scores.masked_fill(mask == 0, -1000000000.0)
weights = F.softmax(scores, dim=-1)
if dropout is not None:
weights = dropout(weights)
out = torch.matmul(weights, value)
return out, weights
class LayerNorm(nn.Module):
"""Construct a layernorm module (See citation for details)."""
def __init__(self, features: 'int', eps: 'float'=1e-06):
super(LayerNorm, self).__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x: 'torch.Tensor'):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class MultiHeadedAttention(nn.Module):
def __init__(self, h, d_model, dropout=0.1):
"""Take in model size and number of heads."""
super(MultiHeadedAttention, self).__init__()
assert d_model % h == 0
self.d_head = d_model // h
self.h = h
self.proj_q = nn.Linear(d_model, d_model)
self.proj_k = nn.Linear(d_model, d_model)
self.proj_v = nn.Linear(d_model, d_model)
self.ret_proj = nn.Linear(d_model, d_model)
self.dropout = nn.Dropout(p=dropout)
def forward(self, query: 'torch.Tensor', key: 'torch.Tensor', value:
'torch.Tensor', mask: 'Optional[torch.Tensor]'=None):
"""
query: (batch_size, seq_len, dmodel)
key: (batch_size, seq_len, dmodel)
value: (batch_size, seq_len, dmodel)
mask: (batch_size, seq_len)
"""
if mask is not None:
mask = mask.unsqueeze(1)
batch_size = query.size(0)
seq_len = query.size(1)
query = self.proj_q(query).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
key = self.proj_k(key).view(batch_size, seq_len, self.h, self.d_head
).transpose(1, 2)
value = self.proj_v(value).view(batch_size, seq_len, self.h, self.
d_head).transpose(1, 2)
x, _ = attention(query, key, value, mask=mask, dropout=self.dropout)
x = x.transpose(1, 2).contiguous().view(batch_size, seq_len, self.h *
self.d_head)
return self.ret_proj(x)
class PositionwiseFeedForward(nn.Module):
"""Implements FFN equation."""
def __init__(self, d_model, d_ff, dropout=0.1):
super(PositionwiseFeedForward, self).__init__()
self.w_1 = nn.Linear(d_model, d_ff)
self.w_2 = nn.Linear(d_ff, d_model)
self.dropout = nn.Dropout(dropout)
def forward(self, x):
return self.w_2(self.dropout(F.relu(self.w_1(x))))
class BlockNew(nn.Module):
"""A standard Decoder block for GPT."""
def __init__(self, d_model: 'int', n_heads: 'int', dropout: 'float'=0.1):
super(BlockNew, self).__init__()
self.d_model = d_model
self.d_inner = 4 * self.d_model
self.n_heads = n_heads
self.dropout = dropout
self.layer_norm1 = LayerNorm(self.d_model)
self.layer_norm2 = LayerNorm(self.d_model)
self.multi_head_attn = MultiHeadedAttention(self.n_heads, self.
d_model, self.dropout)
self.feed_fwd = PositionwiseFeedForward(d_model, self.d_inner, self
.dropout)
def forward(self, input_0):
primals_2 = self.layer_norm1.a_2
primals_3 = self.layer_norm1.b_2
primals_5 = self.layer_norm2.a_2
primals_7 = self.layer_norm2.b_2
primals_4 = self.multi_head_attn.proj_q.weight
primals_9 = self.multi_head_attn.proj_q.bias
primals_6 = self.multi_head_attn.proj_k.weight
primals_11 = self.multi_head_attn.proj_k.bias
primals_8 = self.multi_head_attn.proj_v.weight
primals_12 = self.multi_head_attn.proj_v.bias
primals_10 = self.multi_head_attn.ret_proj.weight
primals_13 = self.multi_head_attn.ret_proj.bias
primals_14 = self.feed_fwd.w_1.weight
primals_15 = self.feed_fwd.w_1.bias
primals_16 = self.feed_fwd.w_2.weight
primals_17 = self.feed_fwd.w_2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17])
return output[0]
|
malhotraa/transformer-experiments
|
Block
| false | 10,472 |
[
"MIT"
] | 0 |
82931b89b14d26dbd6e4ffef8d6f2fd8b7279c0f
|
https://github.com/malhotraa/transformer-experiments/tree/82931b89b14d26dbd6e4ffef8d6f2fd8b7279c0f
|
BinaryLoss
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/td/ctdj5kazgiki6gdaadhqtp2x7tq2ee5ey5hqqdcoqmp54jyhf74f.py
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
triton_poi_fused__log_softmax_0 = async_compile.triton('triton_poi_fused__log_softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ha/chaw4j4cijilv4rurwmzvs3nnol6fmwmio6yneprgmynztfqb3kg.py
# Topologically Sorted Source Nodes: [pos_loss, sum_1, neg_loss, sum_2, add, loss], Original ATen: [aten.neg, aten.sum, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# loss => div
# neg_loss => neg_1
# pos_loss => neg
# sum_1 => sum_3
# sum_2 => sum_4
# Graph fragment:
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%select,), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%neg,), kwargs = {})
# %neg_1 : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%select_1,), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%neg_1,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_3, %sum_4), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, 8), kwargs = {})
triton_per_fused_add_div_neg_sum_1 = async_compile.triton('triton_per_fused_add_div_neg_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_neg_sum_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_div_neg_sum_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex % 16
r1 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (16 + r0 + (64*r1)), None)
tmp1 = tl.load(in_ptr0 + (r0 + (64*r1)), None)
tmp5 = tl.load(in_ptr0 + (32 + r0 + (64*r1)), None)
tmp8 = tl.load(in_ptr0 + (48 + r0 + (64*r1)), None)
tmp17 = tl.load(in_ptr1 + (r0 + (64*r1)), None)
tmp19 = tl.load(in_ptr1 + (16 + r0 + (64*r1)), None)
tmp22 = tl.load(in_ptr1 + (32 + r0 + (64*r1)), None)
tmp25 = tl.load(in_ptr1 + (48 + r0 + (64*r1)), None)
tmp2 = tl_math.exp(tmp1)
tmp3 = tl_math.exp(tmp0)
tmp4 = tmp2 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp13 = -tmp12
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, RBLOCK])
tmp16 = tl.sum(tmp14, 1)[:, None]
tmp18 = tl_math.exp(tmp17)
tmp20 = tl_math.exp(tmp19)
tmp21 = tmp18 + tmp20
tmp23 = tl_math.exp(tmp22)
tmp24 = tmp21 + tmp23
tmp26 = tl_math.exp(tmp25)
tmp27 = tmp24 + tmp26
tmp28 = tl_math.log(tmp27)
tmp29 = tmp17 - tmp28
tmp30 = -tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tmp34 = tmp16 + tmp33
tmp35 = 0.125
tmp36 = tmp34 * tmp35
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp36, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax_1], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_0.run(arg1_1, buf2, 256, grid=grid(256), stream=stream0)
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf4 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [pos_loss, sum_1, neg_loss, sum_2, add, loss], Original ATen: [aten.neg, aten.sum, aten.add, aten.div]
triton_per_fused_add_div_neg_sum_1.run(buf4, buf0, buf2, 1, 64, grid=grid(1), stream=stream0)
del buf0
del buf2
return (buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class BinaryLoss(nn.Module):
def __init__(self):
super(BinaryLoss, self).__init__()
def forward(self, pos_score, neg_score):
pos_loss = -F.log_softmax(pos_score)[:, 1]
neg_loss = -F.log_softmax(neg_score)[:, 0]
loss = (pos_loss.sum() + neg_loss.sum()) / (pos_loss.size(0) +
neg_loss.size(0))
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_per_fused_add_div_neg_sum_1(in_out_ptr0, in_ptr0, in_ptr1,
xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex % 16
r1 = rindex // 16
tmp0 = tl.load(in_ptr0 + (16 + r0 + 64 * r1), None)
tmp1 = tl.load(in_ptr0 + (r0 + 64 * r1), None)
tmp5 = tl.load(in_ptr0 + (32 + r0 + 64 * r1), None)
tmp8 = tl.load(in_ptr0 + (48 + r0 + 64 * r1), None)
tmp17 = tl.load(in_ptr1 + (r0 + 64 * r1), None)
tmp19 = tl.load(in_ptr1 + (16 + r0 + 64 * r1), None)
tmp22 = tl.load(in_ptr1 + (32 + r0 + 64 * r1), None)
tmp25 = tl.load(in_ptr1 + (48 + r0 + 64 * r1), None)
tmp2 = tl_math.exp(tmp1)
tmp3 = tl_math.exp(tmp0)
tmp4 = tmp2 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp13 = -tmp12
tmp14 = tl.broadcast_to(tmp13, [XBLOCK, RBLOCK])
tmp16 = tl.sum(tmp14, 1)[:, None]
tmp18 = tl_math.exp(tmp17)
tmp20 = tl_math.exp(tmp19)
tmp21 = tmp18 + tmp20
tmp23 = tl_math.exp(tmp22)
tmp24 = tmp21 + tmp23
tmp26 = tl_math.exp(tmp25)
tmp27 = tmp24 + tmp26
tmp28 = tl_math.log(tmp27)
tmp29 = tmp17 - tmp28
tmp30 = -tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tmp34 = tmp16 + tmp33
tmp35 = 0.125
tmp36 = tmp34 * tmp35
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp36, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__log_softmax_0[grid(256)](arg1_1, buf2, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf4 = buf1
del buf1
triton_per_fused_add_div_neg_sum_1[grid(1)](buf4, buf0, buf2, 1, 64,
XBLOCK=1, num_warps=2, num_stages=1)
del buf0
del buf2
return buf4,
class BinaryLossNew(nn.Module):
def __init__(self):
super(BinaryLossNew, self).__init__()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
melster1010/VIAME
|
BinaryLoss
| false | 10,473 |
[
"BSD-3-Clause"
] | 0 |
0062265088aae65effbfcd130bfb874c343c785f
|
https://github.com/melster1010/VIAME/tree/0062265088aae65effbfcd130bfb874c343c785f
|
Successor
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yk/cykarwk3qy3pmem22bdbqxnd35nfbs5r3u5357r5uoderrncltwa.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[524288],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 288000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 3600) % 20
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3k/c3k57g5yhtujuoxgs7jwv6kipspabfze46wndcbnl5hbd4s7xuz5.py
# Topologically Sorted Source Nodes: [max_pool2d, conv_out], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
# Source node to ATen node mapping:
# conv_out => relu
# max_pool2d => _low_memory_max_pool2d_with_offsets, getitem_1
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem,), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_1 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[131072],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 72000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = (xindex // 30)
x2 = (xindex // 18000)
x4 = xindex % 18000
x5 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x4 + (18048*x2)), tmp15, xmask)
tl.store(out_ptr1 + (x5), tmp18, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7m/c7mejpb4ik5p2rmp2v4icum4owyptzcyjkwyolwpj6vuvfnck64b.py
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 135200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 676) % 50
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/4l/c4l22elswhi7fd5thcrckd6rdgebe6e6s2sw2sfhhavnlajf5vkq.py
# Topologically Sorted Source Nodes: [max_pool2d_1, conv_out_1], Original ATen: [aten.max_pool2d_with_indices, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# conv_out_1 => relu_1
# max_pool2d_1 => _low_memory_max_pool2d_with_offsets_1, getitem_3
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets_1 : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution_1, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_3 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_1, 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem_2,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3(in_ptr0, out_ptr0, out_ptr1, out_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 33800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 13
x3 = (xindex // 13)
x2 = (xindex // 8450)
x4 = xindex % 8450
x5 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (52*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (52*x3)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (26 + (2*x0) + (52*x3)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (27 + (2*x0) + (52*x3)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp19 = 0.0
tmp20 = tmp18 <= tmp19
tl.store(out_ptr0 + (x4 + (8576*x2)), tmp15, xmask)
tl.store(out_ptr1 + (x5), tmp18, xmask)
tl.store(out_ptr2 + (x4 + (8576*x2)), tmp20, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (20, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (20, ), (1, ))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (50, 20, 5, 5), (500, 25, 5, 1))
assert_size_stride(primals_5, (50, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 20, 60, 60), (72000, 3600, 60, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_2, 288000, grid=grid(288000), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1), torch.int8)
buf3 = empty_strided_cuda((4, 20, 30, 30), (18000, 900, 30, 1), torch.float32)
# Topologically Sorted Source Nodes: [max_pool2d, conv_out], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
triton_poi_fused_max_pool2d_with_indices_relu_1.run(buf1, buf2, buf3, 72000, grid=grid(72000), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 50, 26, 26), (33800, 676, 26, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf5, primals_5, 135200, grid=grid(135200), stream=stream0)
del primals_5
buf6 = empty_strided_cuda((4, 50, 13, 13), (8576, 169, 13, 1), torch.int8)
buf7 = empty_strided_cuda((4, 50, 13, 13), (8450, 169, 13, 1), torch.float32)
buf8 = empty_strided_cuda((4, 50, 13, 13), (8576, 169, 13, 1), torch.bool)
# Topologically Sorted Source Nodes: [max_pool2d_1, conv_out_1], Original ATen: [aten.max_pool2d_with_indices, aten.relu, aten.threshold_backward]
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3.run(buf5, buf6, buf7, buf8, 33800, grid=grid(33800), stream=stream0)
return (buf7, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5, buf6, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((20, 1, 5, 5), (25, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((20, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 1, 64, 64), (4096, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((50, 20, 5, 5), (500, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((50, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
import torch.nn.functional as F
class Successor(nn.Module):
"""Successor successorEncoder model for ADDA."""
def __init__(self):
"""Init Successor successorEncoder."""
super(Successor, self).__init__()
self.restored = False
self.conv1 = nn.Conv2d(1, 20, kernel_size=5)
self.pool1 = nn.MaxPool2d(kernel_size=2)
self.conv2 = nn.Conv2d(20, 50, kernel_size=5)
self.dropout2 = nn.Dropout2d()
self.pool2 = nn.MaxPool2d(kernel_size=2)
def forward(self, input):
"""Forward the Successor."""
conv_out = F.relu(self.pool1(self.conv1(input)))
conv_out = F.relu(self.pool2(self.dropout2(self.conv2(conv_out))))
out = conv_out
return out
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 288000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 3600 % 20
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_1(in_ptr0, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 72000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = xindex // 30
x2 = xindex // 18000
x4 = xindex % 18000
x5 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 120 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x4 + 18048 * x2), tmp15, xmask)
tl.store(out_ptr1 + x5, tmp18, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 135200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 676 % 50
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3(in_ptr0,
out_ptr0, out_ptr1, out_ptr2, xnumel, XBLOCK: tl.constexpr):
xnumel = 33800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 13
x3 = xindex // 13
x2 = xindex // 8450
x4 = xindex % 8450
x5 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 52 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 52 * x3), xmask, eviction_policy
='evict_last')
tmp7 = tl.load(in_ptr0 + (26 + 2 * x0 + 52 * x3), xmask,
eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (27 + 2 * x0 + 52 * x3), xmask,
eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp19 = 0.0
tmp20 = tmp18 <= tmp19
tl.store(out_ptr0 + (x4 + 8576 * x2), tmp15, xmask)
tl.store(out_ptr1 + x5, tmp18, xmask)
tl.store(out_ptr2 + (x4 + 8576 * x2), tmp20, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (20, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (20,), (1,))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (50, 20, 5, 5), (500, 25, 5, 1))
assert_size_stride(primals_5, (50,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 20, 60, 60), (72000, 3600, 60, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(288000)](buf1, primals_2,
288000, XBLOCK=512, num_warps=8, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 20, 30, 30), (18048, 900, 30, 1),
torch.int8)
buf3 = empty_strided_cuda((4, 20, 30, 30), (18000, 900, 30, 1),
torch.float32)
triton_poi_fused_max_pool2d_with_indices_relu_1[grid(72000)](buf1,
buf2, buf3, 72000, XBLOCK=512, num_warps=8, num_stages=1)
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 50, 26, 26), (33800, 676, 26, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_2[grid(135200)](buf5, primals_5,
135200, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_5
buf6 = empty_strided_cuda((4, 50, 13, 13), (8576, 169, 13, 1),
torch.int8)
buf7 = empty_strided_cuda((4, 50, 13, 13), (8450, 169, 13, 1),
torch.float32)
buf8 = empty_strided_cuda((4, 50, 13, 13), (8576, 169, 13, 1),
torch.bool)
triton_poi_fused_max_pool2d_with_indices_relu_threshold_backward_3[grid
(33800)](buf5, buf6, buf7, buf8, 33800, XBLOCK=256, num_warps=4,
num_stages=1)
return (buf7, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5,
buf6, buf8)
class SuccessorNew(nn.Module):
"""Successor successorEncoder model for ADDA."""
def __init__(self):
"""Init Successor successorEncoder."""
super(SuccessorNew, self).__init__()
self.restored = False
self.conv1 = nn.Conv2d(1, 20, kernel_size=5)
self.pool1 = nn.MaxPool2d(kernel_size=2)
self.conv2 = nn.Conv2d(20, 50, kernel_size=5)
self.dropout2 = nn.Dropout2d()
self.pool2 = nn.MaxPool2d(kernel_size=2)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
lindagaw/Kadara
|
Successor
| false | 10,474 |
[
"MIT"
] | 0 |
f1059b69a581344ca460c8df02ac3f73f3fbcba1
|
https://github.com/lindagaw/Kadara/tree/f1059b69a581344ca460c8df02ac3f73f3fbcba1
|
MHSA
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/pw/cpw5jgywzg5ntkknxkt5orxsrrr5zq7a6eoteboi3ba7zrcxj2p7.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/4j/c4jcxsis7ek5ymkow2uiwddmhsmsyxijcvzvkbfjbavvo7na3asf.py
# Topologically Sorted Source Nodes: [content_position], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# content_position => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_1,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16) % 4
x3 = xindex % 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + ((4*x1) + (x0 % 4)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + ((x3 // 4)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x4), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/yz/cyzrujblrx7okcdxhnceg335n3kw4ymde6htxmd352eiillyccf2.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax, div, exp, sub, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_10, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_10, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_per_fused__softmax_2 = async_compile.triton('triton_per_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 1024],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__softmax_2(in_ptr0, in_ptr1, out_ptr2, xnumel, rnumel):
xnumel = 4
XBLOCK: tl.constexpr = 1
rnumel = 1024
RBLOCK: tl.constexpr = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (1024*x0)), None)
tmp1 = tl.load(in_ptr1 + (r1 + (1024*x0)), None)
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp3, 0))
tmp6 = tmp2 - tmp5
tmp7 = tl_math.exp(tmp6)
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = tmp7 / tmp10
tl.store(out_ptr2 + (r1 + (1024*x0)), tmp11, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/le/clegwyiweed5ksuxpqumnnpzduznhg57pkc32gtm37gqnyk2iaoq.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.avg_pool2d]
# Source node to ATen node mapping:
# out => avg_pool2d
# Graph fragment:
# %avg_pool2d : [num_users=1] = call_function[target=torch.ops.aten.avg_pool2d.default](args = (%view_15, [2, 2], [2, 2]), kwargs = {})
triton_poi_fused_avg_pool2d_3 = async_compile.triton('triton_poi_fused_avg_pool2d_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_avg_pool2d_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_avg_pool2d_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 2
x1 = (xindex // 2)
x2 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (8*x1)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (8*x1)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (4 + (2*x0) + (8*x1)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (5 + (2*x0) + (8*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 + tmp0
tmp4 = tmp3 + tmp2
tmp6 = tmp5 + tmp4
tmp7 = 0.25
tmp8 = tmp6 * tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (1, 4, 1, 1, 4), (16, 4, 4, 4, 1))
assert_size_stride(primals_9, (1, 4, 1, 4, 1), (16, 4, 4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_1, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf3, primals_3, 256, grid=grid(256), stream=stream0)
del primals_3
buf4 = empty_strided_cuda((4, 4, 1, 16), (64, 16, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [content_position], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(primals_8, primals_9, buf4, 256, grid=grid(256), stream=stream0)
del primals_8
del primals_9
buf5 = empty_strided_cuda((16, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [content_position], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 16, 1), (16, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 16), (16, 0, 1), 0), out=buf5)
buf6 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf6, primals_5, 256, grid=grid(256), stream=stream0)
del primals_5
buf7 = empty_strided_cuda((16, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [content_content], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 16, 1), (16, 1, 0), 0), reinterpret_tensor(buf6, (16, 1, 16), (16, 0, 1), 0), out=buf7)
buf10 = empty_strided_cuda((4, 1024), (1024, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_per_fused__softmax_2.run(buf7, buf5, buf10, 4, 1024, grid=grid(4), stream=stream0)
del buf5
del buf7
buf11 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf11, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
buf12 = empty_strided_cuda((16, 1, 16), (16, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul_2], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf11, (16, 1, 16), (16, 0, 1), 0), reinterpret_tensor(buf10, (16, 16, 16), (256, 16, 1), 0), out=buf12)
buf13 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.avg_pool2d]
triton_poi_fused_avg_pool2d_3.run(buf12, buf13, 64, grid=grid(64), stream=stream0)
return (buf13, primals_1, primals_2, primals_4, primals_6, buf10, reinterpret_tensor(buf12, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf11, (16, 16, 1), (16, 1, 16), 0), reinterpret_tensor(buf3, (16, 1, 16), (16, 16, 1), 0), reinterpret_tensor(buf6, (16, 16, 1), (16, 1, 16), 0), reinterpret_tensor(buf4, (16, 16, 1), (16, 1, 16), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 4, 1, 1, 4), (16, 4, 4, 4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, 4, 1, 4, 1), (16, 4, 4, 1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class MHSA(nn.Module):
def __init__(self, height, width, dim, head):
super(MHSA, self).__init__()
self.head = head
self.r_h = nn.Parameter(data=torch.randn(1, head, dim // head, 1,
height), requires_grad=True)
self.r_w = nn.Parameter(data=torch.randn(1, head, dim // head,
width, 1), requires_grad=True)
self.w_q = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.w_k = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.w_v = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.softmax = nn.Softmax(dim=-1)
self.pool = nn.AvgPool2d(kernel_size=2, stride=2, padding=0)
def forward(self, x):
batch, dim, height, width = x.size()
q = self.w_q(x).view(batch, self.head, dim // self.head, -1).permute(
0, 1, 3, 2)
k = self.w_k(x).view(batch, self.head, dim // self.head, -1)
v = self.w_v(x).view(batch, self.head, dim // self.head, -1)
r = (self.r_h + self.r_w).view(1, self.head, dim // self.head, -1)
content_position = torch.matmul(q, r)
content_content = torch.matmul(q, k)
energy = (content_content + content_position).view(batch, -1)
attention = self.softmax(energy).view(batch, self.head, height *
width, height * width)
feature = torch.matmul(v, attention).view(batch, dim, height, width)
out = self.pool(feature)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'height': 4, 'width': 4, 'dim': 4, 'head': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16 % 4
x3 = xindex % 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + (4 * x1 + x0 % 4), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + x3 // 4, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + x4, tmp2, xmask)
@triton.jit
def triton_per_fused__softmax_2(in_ptr0, in_ptr1, out_ptr2, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 1024 * x0), None)
tmp1 = tl.load(in_ptr1 + (r1 + 1024 * x0), None)
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp3, 0))
tmp6 = tmp2 - tmp5
tmp7 = tl_math.exp(tmp6)
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = tmp7 / tmp10
tl.store(out_ptr2 + (r1 + 1024 * x0), tmp11, None)
@triton.jit
def triton_poi_fused_avg_pool2d_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 2
x1 = xindex // 2
x2 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 8 * x1), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 8 * x1), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (4 + 2 * x0 + 8 * x1), xmask, eviction_policy=
'evict_last')
tmp5 = tl.load(in_ptr0 + (5 + 2 * x0 + 8 * x1), xmask, eviction_policy=
'evict_last')
tmp2 = tmp1 + tmp0
tmp4 = tmp3 + tmp2
tmp6 = tmp5 + tmp4
tmp7 = 0.25
tmp8 = tmp6 * tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (1, 4, 1, 1, 4), (16, 4, 4, 4, 1))
assert_size_stride(primals_9, (1, 4, 1, 4, 1), (16, 4, 4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = extern_kernels.convolution(primals_1, primals_6, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(256)](buf3, primals_3, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
buf4 = empty_strided_cuda((4, 4, 1, 16), (64, 16, 16, 1), torch.float32
)
triton_poi_fused_clone_1[grid(256)](primals_8, primals_9, buf4, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_8
del primals_9
buf5 = empty_strided_cuda((16, 16, 16), (256, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 16, 1), (16, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 16), (16, 0, 1), 0), out=buf5)
buf6 = buf1
del buf1
triton_poi_fused_convolution_0[grid(256)](buf6, primals_5, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf7 = empty_strided_cuda((16, 16, 16), (256, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 16, 1), (16, 1, 0),
0), reinterpret_tensor(buf6, (16, 1, 16), (16, 0, 1), 0), out=buf7)
buf10 = empty_strided_cuda((4, 1024), (1024, 1), torch.float32)
triton_per_fused__softmax_2[grid(4)](buf7, buf5, buf10, 4, 1024,
num_warps=8, num_stages=1)
del buf5
del buf7
buf11 = buf2
del buf2
triton_poi_fused_convolution_0[grid(256)](buf11, primals_7, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_7
buf12 = empty_strided_cuda((16, 1, 16), (16, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf11, (16, 1, 16), (16, 0, 1
), 0), reinterpret_tensor(buf10, (16, 16, 16), (256, 16, 1), 0),
out=buf12)
buf13 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
triton_poi_fused_avg_pool2d_3[grid(64)](buf12, buf13, 64, XBLOCK=64,
num_warps=1, num_stages=1)
return (buf13, primals_1, primals_2, primals_4, primals_6, buf10,
reinterpret_tensor(buf12, (4, 4, 4, 4), (64, 16, 4, 1), 0),
reinterpret_tensor(buf11, (16, 16, 1), (16, 1, 16), 0),
reinterpret_tensor(buf3, (16, 1, 16), (16, 16, 1), 0),
reinterpret_tensor(buf6, (16, 16, 1), (16, 1, 16), 0),
reinterpret_tensor(buf4, (16, 16, 1), (16, 1, 16), 0))
class MHSANew(nn.Module):
def __init__(self, height, width, dim, head):
super(MHSANew, self).__init__()
self.head = head
self.r_h = nn.Parameter(data=torch.randn(1, head, dim // head, 1,
height), requires_grad=True)
self.r_w = nn.Parameter(data=torch.randn(1, head, dim // head,
width, 1), requires_grad=True)
self.w_q = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.w_k = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.w_v = nn.Conv2d(in_channels=dim, out_channels=dim, kernel_size
=1, stride=1, bias=True)
self.softmax = nn.Softmax(dim=-1)
self.pool = nn.AvgPool2d(kernel_size=2, stride=2, padding=0)
def forward(self, input_0):
primals_8 = self.r_h
primals_9 = self.r_w
primals_2 = self.w_q.weight
primals_3 = self.w_q.bias
primals_4 = self.w_k.weight
primals_5 = self.w_k.bias
primals_6 = self.w_v.weight
primals_7 = self.w_v.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
|
lzu-zhanghr/vision-transformer-zoo
|
MHSA
| false | 10,475 |
[
"MIT"
] | 0 |
2cc6e3551c39816acc95ade040bbf9bd115a6b03
|
https://github.com/lzu-zhanghr/vision-transformer-zoo/tree/2cc6e3551c39816acc95ade040bbf9bd115a6b03
|
IdentityPadding
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/i5/ci5jtsingp2qnrodx6wlbyak3tjanrom5rpybyruspgrcwq6flhn.py
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# out_1 => getitem
# Graph fragment:
# %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 0), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_0 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.max_pool2d_with_indices]
stream0 = get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.functional as F
class IdentityPadding(nn.Module):
def __init__(self, in_channels, out_channels, stride):
super(IdentityPadding, self).__init__()
self.pooling = nn.MaxPool2d(1, stride=stride)
self.add_channels = out_channels - in_channels
def forward(self, x):
out = F.pad(x, (0, 0, 0, 0, 0, self.add_channels))
out = self.pooling(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'stride': 1}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tl.store(out_ptr0 + x0, tmp0, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0[grid(256)](arg0_1, buf0,
256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class IdentityPaddingNew(nn.Module):
def __init__(self, in_channels, out_channels, stride):
super(IdentityPaddingNew, self).__init__()
self.pooling = nn.MaxPool2d(1, stride=stride)
self.add_channels = out_channels - in_channels
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
moerashidi/deep_ensemble
|
IdentityPadding
| false | 10,476 |
[
"MIT"
] | 0 |
51cd890643b0f01849583e6585eef241776b0ef4
|
https://github.com/moerashidi/deep_ensemble/tree/51cd890643b0f01849583e6585eef241776b0ef4
|
MutliClassNN
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/r3/cr3dj3iqo4mxp3jap6oqywrnhpoxhtxzi373qtxftjspbz3o3pb3.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 1000
x2 = xindex % 4000
x3 = (xindex // 4000)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (4096*x3)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/q5/cq52p2qap7uob2ddnn4qeh67r3muutkp3yhbkqpu4eqaemol3idl.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# x_1 => sigmoid
# Graph fragment:
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%view_3,), kwargs = {})
triton_poi_fused_sigmoid_1 = async_compile.triton('triton_poi_fused_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (1000, 4), (4, 1))
assert_size_stride(primals_2, (1000, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 1000), (1000, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 1000), (1000, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 1000), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 1000), (16000, 4000, 1000, 1), 0); del buf0 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 1000), (16384, 4096, 1000, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf4, 64000, grid=grid(64000), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 1000), (1000, 1), 0), reinterpret_tensor(primals_4, (1000, 4), (1, 1000), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_1.run(buf3, primals_5, 256, grid=grid(256), stream=stream0)
del primals_5
return (buf3, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 1000), (1000, 1), 0), buf3, primals_4, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((1000, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1000, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 1000), (1000, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class MutliClassNN(nn.Module):
def __init__(self, num_features, num_labels):
super(MutliClassNN, self).__init__()
self.fc1 = torch.nn.Linear(num_features, 1000)
self.fc3 = torch.nn.Linear(1000, num_labels)
def forward(self, x):
x = torch.relu(self.fc1(x))
x = torch.sigmoid(self.fc3(x))
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_features': 4, 'num_labels': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 1000
x2 = xindex % 4000
x3 = xindex // 4000
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 4096 * x3), tmp6, xmask)
@triton.jit
def triton_poi_fused_sigmoid_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (1000, 4), (4, 1))
assert_size_stride(primals_2, (1000,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 1000), (1000, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 1000), (1000, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 1000), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 1000), (16000, 4000, 1000,
1), 0)
del buf0
buf4 = empty_strided_cuda((4, 4, 4, 1000), (16384, 4096, 1000, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(64000)](buf1,
primals_2, buf4, 64000, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 1000), (1000, 1), 0
), reinterpret_tensor(primals_4, (1000, 4), (1, 1000), 0), out=buf2
)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
triton_poi_fused_sigmoid_1[grid(256)](buf3, primals_5, 256, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_5
return buf3, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 1000), (1000, 1), 0
), buf3, primals_4, buf4
class MutliClassNNNew(nn.Module):
def __init__(self, num_features, num_labels):
super(MutliClassNNNew, self).__init__()
self.fc1 = torch.nn.Linear(num_features, 1000)
self.fc3 = torch.nn.Linear(1000, num_labels)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc3.weight
primals_5 = self.fc3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
mhagenow01/ECE532ClassifierComparison
|
MutliClassNN
| false | 10,477 |
[
"MIT"
] | 0 |
5066931d97aae2c25c8b9451fe3d12021f5748a1
|
https://github.com/mhagenow01/ECE532ClassifierComparison/tree/5066931d97aae2c25c8b9451fe3d12021f5748a1
|
SymKlCriterion
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wv/cwvti54lsojjpkh6f73xsvv55wjtno2rrmtjgaznlsjst37yn74a.py
# Topologically Sorted Source Nodes: [softmax, log_softmax_1], Original ATen: [aten._softmax, aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax_1 => amax_2, sub_4
# softmax => amax_1, exp_1, sub_2
# Graph fragment:
# %amax_1 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg1_1, [-1], True), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %amax_1), kwargs = {})
# %exp_1 : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_2,), kwargs = {})
# %amax_2 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg1_1, [-1], True), kwargs = {})
# %sub_4 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %amax_2), kwargs = {})
triton_poi_fused__log_softmax__softmax_0 = async_compile.triton('triton_poi_fused__log_softmax__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax__softmax_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
tl.store(out_ptr1 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/7m/c7m62lqyfgv73n2cb7qnzr2xodmurvv7yy6z2m7nhu5m3pfdjjxe.py
# Topologically Sorted Source Nodes: [log_softmax, softmax_1], Original ATen: [aten._log_softmax, aten._softmax]
# Source node to ATen node mapping:
# log_softmax => amax, sub
# softmax_1 => amax_3, exp_3, sub_6
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
# %amax_3 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [-1], True), kwargs = {})
# %sub_6 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax_3), kwargs = {})
# %exp_3 : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_6,), kwargs = {})
triton_poi_fused__log_softmax__softmax_1 = async_compile.triton('triton_poi_fused__log_softmax__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax__softmax_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp8, xmask)
tl.store(out_ptr1 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/t4/ct4y23pxxoeimylfghb3zgxs2sgexranitqqpowxqihbyqj4rh4q.py
# Topologically Sorted Source Nodes: [softmax, kl_div, log_softmax, softmax_1, kl_div_1, log_softmax_1, loss, loss_1], Original ATen: [aten._softmax, aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.mean, aten.add]
# Source node to ATen node mapping:
# kl_div => eq, full_default, full_default_1, isnan, log_1, mean, mul, mul_1, sub_3, where, where_1
# kl_div_1 => eq_1, full_default_2, full_default_3, isnan_1, log_3, mean_1, mul_2, mul_3, sub_7, where_2, where_3
# log_softmax => exp, log, sub_1, sum_1
# log_softmax_1 => exp_2, log_2, sub_5, sum_3
# loss => add
# loss_1 => mul_4
# softmax => div, sum_2
# softmax_1 => div_1, sum_4
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_1, [-1], True), kwargs = {})
# %div : [num_users=5] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_1, %sum_2), kwargs = {})
# %isnan : [num_users=1] = call_function[target=torch.ops.aten.isnan.default](args = (%div,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], nan), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%div, 0), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %log_1 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%div,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %log_1), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %mul_1), kwargs = {})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%isnan, %full_default_1, %where), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %sub_1), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where_1, %mul), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_3,), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_3, [-1], True), kwargs = {})
# %div_1 : [num_users=5] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_3, %sum_4), kwargs = {})
# %isnan_1 : [num_users=1] = call_function[target=torch.ops.aten.isnan.default](args = (%div_1,), kwargs = {})
# %full_default_3 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], nan), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq_1 : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%div_1, 0), kwargs = {})
# %full_default_2 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %log_3 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%div_1,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %log_3), kwargs = {})
# %where_2 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq_1, %full_default_2, %mul_3), kwargs = {})
# %where_3 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%isnan_1, %full_default_3, %where_2), kwargs = {})
# %exp_2 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub_4,), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_2, [-1], True), kwargs = {})
# %log_2 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_3,), kwargs = {})
# %sub_5 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_4, %log_2), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %sub_5), kwargs = {})
# %sub_7 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where_3, %mul_2), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_7,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, %mean_1), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add, 1.0), kwargs = {})
triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2 = async_compile.triton('triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 6), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 20, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, xnumel, rnumel, XBLOCK : tl.constexpr, RBLOCK : tl.constexpr):
xnumel = 1
rnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rbase = tl.arange(0, RBLOCK)[None, :]
_tmp34 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
_tmp68 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
for roffset in range(0, rnumel, RBLOCK):
rindex = roffset + rbase
rmask = rindex < rnumel
r2 = rindex
r1 = (rindex // 4)
tmp0 = tl.load(in_ptr0 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp1 = tl.load(in_ptr0 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp2 = tl.load(in_ptr0 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp4 = tl.load(in_ptr0 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp17 = tl.load(in_ptr1 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp18 = tl.load(in_ptr1 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp20 = tl.load(in_ptr1 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp23 = tl.load(in_ptr1 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp26 = tl.load(in_ptr1 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp36 = tl.load(in_ptr2 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp37 = tl.load(in_ptr2 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp38 = tl.load(in_ptr2 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp40 = tl.load(in_ptr2 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp42 = tl.load(in_ptr2 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp51 = tl.load(in_ptr3 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp52 = tl.load(in_ptr3 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp54 = tl.load(in_ptr3 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp57 = tl.load(in_ptr3 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp60 = tl.load(in_ptr3 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp9 = libdevice.isnan(tmp8).to(tl.int1)
tmp10 = 0.0
tmp11 = tmp8 == tmp10
tmp12 = tl_math.log(tmp8)
tmp13 = tmp8 * tmp12
tmp14 = tl.where(tmp11, tmp10, tmp13)
tmp15 = float("nan")
tmp16 = tl.where(tmp9, tmp15, tmp14)
tmp19 = tl_math.exp(tmp18)
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp19 + tmp21
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tl_math.log(tmp28)
tmp30 = tmp17 - tmp29
tmp31 = tmp8 * tmp30
tmp32 = tmp16 - tmp31
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = _tmp34 + tmp33
_tmp34 = tl.where(rmask, tmp35, _tmp34)
tmp39 = tmp37 + tmp38
tmp41 = tmp39 + tmp40
tmp43 = tmp41 + tmp42
tmp44 = tmp36 / tmp43
tmp45 = libdevice.isnan(tmp44).to(tl.int1)
tmp46 = tmp44 == tmp10
tmp47 = tl_math.log(tmp44)
tmp48 = tmp44 * tmp47
tmp49 = tl.where(tmp46, tmp10, tmp48)
tmp50 = tl.where(tmp45, tmp15, tmp49)
tmp53 = tl_math.exp(tmp52)
tmp55 = tl_math.exp(tmp54)
tmp56 = tmp53 + tmp55
tmp58 = tl_math.exp(tmp57)
tmp59 = tmp56 + tmp58
tmp61 = tl_math.exp(tmp60)
tmp62 = tmp59 + tmp61
tmp63 = tl_math.log(tmp62)
tmp64 = tmp51 - tmp63
tmp65 = tmp44 * tmp64
tmp66 = tmp50 - tmp65
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = _tmp68 + tmp67
_tmp68 = tl.where(rmask, tmp69, _tmp68)
tmp34 = tl.sum(_tmp34, 1)[:, None]
tmp68 = tl.sum(_tmp68, 1)[:, None]
tmp70 = 256.0
tmp71 = tmp34 / tmp70
tmp72 = tmp68 / tmp70
tmp73 = tmp71 + tmp72
tmp74 = 1.0
tmp75 = tmp73 * tmp74
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp75, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax, log_softmax_1], Original ATen: [aten._softmax, aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax__softmax_0.run(arg1_1, buf0, buf6, 256, grid=grid(256), stream=stream0)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax, softmax_1], Original ATen: [aten._log_softmax, aten._softmax]
triton_poi_fused__log_softmax__softmax_1.run(arg0_1, buf2, buf4, 256, grid=grid(256), stream=stream0)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf8 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [softmax, kl_div, log_softmax, softmax_1, kl_div_1, log_softmax_1, loss, loss_1], Original ATen: [aten._softmax, aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.mean, aten.add]
triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2.run(buf8, buf0, buf2, buf4, buf6, 1, 256, grid=grid(1), stream=stream0)
del buf0
del buf2
del buf4
del buf6
return (buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
from torch.nn.modules.loss import _Loss
from torch.optim.lr_scheduler import *
class Criterion(_Loss):
def __init__(self, alpha=1.0, name='criterion'):
super().__init__()
"""Alpha is used to weight each loss term
"""
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""weight: sample weight
"""
return
class SymKlCriterion(Criterion):
def __init__(self, alpha=1.0, name='KL Div Criterion'):
super().__init__()
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""input/target: logits
"""
input = input.float()
target = target.float()
loss = F.kl_div(F.log_softmax(input, dim=-1), F.softmax(target.
detach(), dim=-1)) + F.kl_div(F.log_softmax(target, dim=-1), F.
softmax(input.detach(), dim=-1))
loss = loss * self.alpha
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch.nn.modules.loss import _Loss
from torch.optim.lr_scheduler import *
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__log_softmax__softmax_0(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
tl.store(out_ptr1 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused__log_softmax__softmax_1(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp8, xmask)
tl.store(out_ptr1 + x2, tmp9, xmask)
@triton.jit
def triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2(in_out_ptr0
, in_ptr0, in_ptr1, in_ptr2, in_ptr3, xnumel, rnumel, XBLOCK: tl.
constexpr, RBLOCK: tl.constexpr):
rnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rbase = tl.arange(0, RBLOCK)[None, :]
_tmp34 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
_tmp68 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
for roffset in range(0, rnumel, RBLOCK):
rindex = roffset + rbase
rmask = rindex < rnumel
r2 = rindex
r1 = rindex // 4
tmp0 = tl.load(in_ptr0 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp1 = tl.load(in_ptr0 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp2 = tl.load(in_ptr0 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp4 = tl.load(in_ptr0 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp17 = tl.load(in_ptr1 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp18 = tl.load(in_ptr1 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp20 = tl.load(in_ptr1 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp23 = tl.load(in_ptr1 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp26 = tl.load(in_ptr1 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp36 = tl.load(in_ptr2 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp37 = tl.load(in_ptr2 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp38 = tl.load(in_ptr2 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp40 = tl.load(in_ptr2 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp42 = tl.load(in_ptr2 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp51 = tl.load(in_ptr3 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp52 = tl.load(in_ptr3 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp54 = tl.load(in_ptr3 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp57 = tl.load(in_ptr3 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp60 = tl.load(in_ptr3 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp9 = libdevice.isnan(tmp8).to(tl.int1)
tmp10 = 0.0
tmp11 = tmp8 == tmp10
tmp12 = tl_math.log(tmp8)
tmp13 = tmp8 * tmp12
tmp14 = tl.where(tmp11, tmp10, tmp13)
tmp15 = float('nan')
tmp16 = tl.where(tmp9, tmp15, tmp14)
tmp19 = tl_math.exp(tmp18)
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp19 + tmp21
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tl_math.log(tmp28)
tmp30 = tmp17 - tmp29
tmp31 = tmp8 * tmp30
tmp32 = tmp16 - tmp31
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = _tmp34 + tmp33
_tmp34 = tl.where(rmask, tmp35, _tmp34)
tmp39 = tmp37 + tmp38
tmp41 = tmp39 + tmp40
tmp43 = tmp41 + tmp42
tmp44 = tmp36 / tmp43
tmp45 = libdevice.isnan(tmp44).to(tl.int1)
tmp46 = tmp44 == tmp10
tmp47 = tl_math.log(tmp44)
tmp48 = tmp44 * tmp47
tmp49 = tl.where(tmp46, tmp10, tmp48)
tmp50 = tl.where(tmp45, tmp15, tmp49)
tmp53 = tl_math.exp(tmp52)
tmp55 = tl_math.exp(tmp54)
tmp56 = tmp53 + tmp55
tmp58 = tl_math.exp(tmp57)
tmp59 = tmp56 + tmp58
tmp61 = tl_math.exp(tmp60)
tmp62 = tmp59 + tmp61
tmp63 = tl_math.log(tmp62)
tmp64 = tmp51 - tmp63
tmp65 = tmp44 * tmp64
tmp66 = tmp50 - tmp65
tmp67 = tl.broadcast_to(tmp66, [XBLOCK, RBLOCK])
tmp69 = _tmp68 + tmp67
_tmp68 = tl.where(rmask, tmp69, _tmp68)
tmp34 = tl.sum(_tmp34, 1)[:, None]
tmp68 = tl.sum(_tmp68, 1)[:, None]
tmp70 = 256.0
tmp71 = tmp34 / tmp70
tmp72 = tmp68 / tmp70
tmp73 = tmp71 + tmp72
tmp74 = 1.0
tmp75 = tmp73 * tmp74
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp75, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax__softmax_0[grid(256)](arg1_1, buf0,
buf6, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__log_softmax__softmax_1[grid(256)](arg0_1, buf2,
buf4, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf8 = buf3
del buf3
triton_red_fused__log_softmax__softmax_add_mean_mul_sub_xlogy_2[grid(1)
](buf8, buf0, buf2, buf4, buf6, 1, 256, XBLOCK=1, RBLOCK=256,
num_warps=8, num_stages=1)
del buf0
del buf2
del buf4
del buf6
return buf8,
class Criterion(_Loss):
def __init__(self, alpha=1.0, name='criterion'):
super().__init__()
"""Alpha is used to weight each loss term
"""
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""weight: sample weight
"""
return
class SymKlCriterionNew(Criterion):
def __init__(self, alpha=1.0, name='KL Div Criterion'):
super().__init__()
self.alpha = alpha
self.name = name
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
mahartmann/mt-dnn
|
SymKlCriterion
| false | 10,479 |
[
"MIT"
] | 0 |
c9aa3379dc255fd8fc40f24b6cd508f6a645b32f
|
https://github.com/mahartmann/mt-dnn/tree/c9aa3379dc255fd8fc40f24b6cd508f6a645b32f
|
KlCriterion
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/wx/cwxwvlntewdrqi2r4caciy5ht4jdvafnhtiqncr4lo4aegcb4imz.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax_1, exp_1, sub_2
# Graph fragment:
# %amax_1 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg1_1, [-1], True), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %amax_1), kwargs = {})
# %exp_1 : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_2,), kwargs = {})
triton_poi_fused__softmax_0 = async_compile.triton('triton_poi_fused__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/g5/cg5f2rptqnpi2mrqpqc4tujqpbrrrjrse6plhgftx425znsffpfv.py
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
triton_poi_fused__log_softmax_1 = async_compile.triton('triton_poi_fused__log_softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/i2/ci2yamwiz72oxetvudslk7ph4kozejsdj65prr6khbbmwns6aynp.py
# Topologically Sorted Source Nodes: [softmax, loss, log_softmax, loss_1], Original ATen: [aten._softmax, aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.mean]
# Source node to ATen node mapping:
# log_softmax => exp, log, sub_1, sum_1
# loss => eq, full_default, full_default_1, isnan, log_1, mean, mul, mul_1, sub_3, where, where_1
# loss_1 => mul_2
# softmax => div, sum_2
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_1, [-1], True), kwargs = {})
# %div : [num_users=5] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_1, %sum_2), kwargs = {})
# %isnan : [num_users=1] = call_function[target=torch.ops.aten.isnan.default](args = (%div,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], nan), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%div, 0), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %log_1 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%div,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %log_1), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %mul_1), kwargs = {})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%isnan, %full_default_1, %where), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %sub_1), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where_1, %mul), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_3,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 1.0), kwargs = {})
triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2 = async_compile.triton('triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr, RBLOCK : tl.constexpr):
xnumel = 1
rnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rbase = tl.arange(0, RBLOCK)[None, :]
_tmp34 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
for roffset in range(0, rnumel, RBLOCK):
rindex = roffset + rbase
rmask = rindex < rnumel
r2 = rindex
r1 = (rindex // 4)
tmp0 = tl.load(in_ptr0 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp1 = tl.load(in_ptr0 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp2 = tl.load(in_ptr0 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp4 = tl.load(in_ptr0 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp17 = tl.load(in_ptr1 + (r2), rmask, eviction_policy='evict_first', other=0.0)
tmp18 = tl.load(in_ptr1 + (4*r1), rmask, eviction_policy='evict_last', other=0.0)
tmp20 = tl.load(in_ptr1 + (1 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp23 = tl.load(in_ptr1 + (2 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp26 = tl.load(in_ptr1 + (3 + (4*r1)), rmask, eviction_policy='evict_last', other=0.0)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp9 = libdevice.isnan(tmp8).to(tl.int1)
tmp10 = 0.0
tmp11 = tmp8 == tmp10
tmp12 = tl_math.log(tmp8)
tmp13 = tmp8 * tmp12
tmp14 = tl.where(tmp11, tmp10, tmp13)
tmp15 = float("nan")
tmp16 = tl.where(tmp9, tmp15, tmp14)
tmp19 = tl_math.exp(tmp18)
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp19 + tmp21
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tl_math.log(tmp28)
tmp30 = tmp17 - tmp29
tmp31 = tmp8 * tmp30
tmp32 = tmp16 - tmp31
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = _tmp34 + tmp33
_tmp34 = tl.where(rmask, tmp35, _tmp34)
tmp34 = tl.sum(_tmp34, 1)[:, None]
tmp36 = 256.0
tmp37 = tmp34 / tmp36
tmp38 = 1.0
tmp39 = tmp37 * tmp38
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp39, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(arg1_1, buf0, 256, grid=grid(256), stream=stream0)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_1.run(arg0_1, buf2, 256, grid=grid(256), stream=stream0)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [softmax, loss, log_softmax, loss_1], Original ATen: [aten._softmax, aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.mean]
triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2.run(buf4, buf0, buf2, 1, 256, grid=grid(1), stream=stream0)
del buf0
del buf2
return (buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn.functional as F
from torch.nn.modules.loss import _Loss
from torch.optim.lr_scheduler import *
class Criterion(_Loss):
def __init__(self, alpha=1.0, name='criterion'):
super().__init__()
"""Alpha is used to weight each loss term
"""
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""weight: sample weight
"""
return
class KlCriterion(Criterion):
def __init__(self, alpha=1.0, name='KL Div Criterion'):
super().__init__()
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""input/target: logits
"""
input = input.float()
target = target.float()
loss = F.kl_div(F.log_softmax(input, dim=-1), F.softmax(target, dim=-1)
)
loss = loss * self.alpha
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch.nn.modules.loss import _Loss
from torch.optim.lr_scheduler import *
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__log_softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2(in_out_ptr0,
in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr, RBLOCK: tl.
constexpr):
rnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rbase = tl.arange(0, RBLOCK)[None, :]
_tmp34 = tl.full([XBLOCK, RBLOCK], 0, tl.float32)
for roffset in range(0, rnumel, RBLOCK):
rindex = roffset + rbase
rmask = rindex < rnumel
r2 = rindex
r1 = rindex // 4
tmp0 = tl.load(in_ptr0 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp1 = tl.load(in_ptr0 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp2 = tl.load(in_ptr0 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp4 = tl.load(in_ptr0 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp17 = tl.load(in_ptr1 + r2, rmask, eviction_policy='evict_first',
other=0.0)
tmp18 = tl.load(in_ptr1 + 4 * r1, rmask, eviction_policy=
'evict_last', other=0.0)
tmp20 = tl.load(in_ptr1 + (1 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp23 = tl.load(in_ptr1 + (2 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp26 = tl.load(in_ptr1 + (3 + 4 * r1), rmask, eviction_policy=
'evict_last', other=0.0)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp9 = libdevice.isnan(tmp8).to(tl.int1)
tmp10 = 0.0
tmp11 = tmp8 == tmp10
tmp12 = tl_math.log(tmp8)
tmp13 = tmp8 * tmp12
tmp14 = tl.where(tmp11, tmp10, tmp13)
tmp15 = float('nan')
tmp16 = tl.where(tmp9, tmp15, tmp14)
tmp19 = tl_math.exp(tmp18)
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp19 + tmp21
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tl_math.log(tmp28)
tmp30 = tmp17 - tmp29
tmp31 = tmp8 * tmp30
tmp32 = tmp16 - tmp31
tmp33 = tl.broadcast_to(tmp32, [XBLOCK, RBLOCK])
tmp35 = _tmp34 + tmp33
_tmp34 = tl.where(rmask, tmp35, _tmp34)
tmp34 = tl.sum(_tmp34, 1)[:, None]
tmp36 = 256.0
tmp37 = tmp34 / tmp36
tmp38 = 1.0
tmp39 = tmp37 * tmp38
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp39, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(256)](arg1_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__log_softmax_1[grid(256)](arg0_1, buf2, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3
del buf3
triton_red_fused__log_softmax__softmax_mean_mul_sub_xlogy_2[grid(1)](
buf4, buf0, buf2, 1, 256, XBLOCK=1, RBLOCK=256, num_warps=8,
num_stages=1)
del buf0
del buf2
return buf4,
class Criterion(_Loss):
def __init__(self, alpha=1.0, name='criterion'):
super().__init__()
"""Alpha is used to weight each loss term
"""
self.alpha = alpha
self.name = name
def forward(self, input, target, weight=None, ignore_index=-1):
"""weight: sample weight
"""
return
class KlCriterionNew(Criterion):
def __init__(self, alpha=1.0, name='KL Div Criterion'):
super().__init__()
self.alpha = alpha
self.name = name
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
|
mahartmann/mt-dnn
|
KlCriterion
| false | 10,480 |
[
"MIT"
] | 0 |
c9aa3379dc255fd8fc40f24b6cd508f6a645b32f
|
https://github.com/mahartmann/mt-dnn/tree/c9aa3379dc255fd8fc40f24b6cd508f6a645b32f
|
FlowHead
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/3p/c3pccxgnuu4ndvejdrgnnzzkvckhydfsbcaf7lwd5g3lofpww4cc.py
# Topologically Sorted Source Nodes: [conv2d, relu], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d => convolution
# relu => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_0 = async_compile.triton('triton_poi_fused_convolution_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4194304],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4194304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 4096) % 256
tmp0 = tl.load(in_out_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sc/cscsiwn4jzs35kmdkiqiai55z42bpakheiazewur3x5beq7teiv3.py
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32768
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 4096) % 2
tmp0 = tl.load(in_out_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 128, 64, 64), (524288, 4096, 64, 1))
assert_size_stride(primals_4, (2, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (2, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d, relu], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 4194304, grid=grid(4194304), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 2, 64, 64), (8192, 4096, 64, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf3, primals_5, 32768, grid=grid(32768), stream=stream0)
del primals_5
return (buf3, primals_1, primals_3, primals_4, buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((256, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 128, 64, 64), (524288, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((2, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((2, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class FlowHead(nn.Module):
def __init__(self, input_dim=128, hidden_dim=256):
super(FlowHead, self).__init__()
self.conv1 = nn.Conv2d(input_dim, hidden_dim, 3, padding=1)
self.conv2 = nn.Conv2d(hidden_dim, 2, 3, padding=1)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
return self.conv2(self.relu(self.conv1(x)))
def get_inputs():
return [torch.rand([4, 128, 64, 64])]
def get_init_inputs():
return [[], {}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 4096 % 256
tmp0 = tl.load(in_out_ptr0 + x3, None)
tmp1 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, None)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 4096 % 2
tmp0 = tl.load(in_out_ptr0 + x3, None)
tmp1 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, None)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 128, 64, 64), (524288, 4096, 64, 1))
assert_size_stride(primals_4, (2, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (2,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(4194304)](buf1, primals_2,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 2, 64, 64), (8192, 4096, 64, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_1[grid(32768)](buf3, primals_5, 32768,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
return buf3, primals_1, primals_3, primals_4, buf1
class FlowHeadNew(nn.Module):
def __init__(self, input_dim=128, hidden_dim=256):
super(FlowHeadNew, self).__init__()
self.conv1 = nn.Conv2d(input_dim, hidden_dim, 3, padding=1)
self.conv2 = nn.Conv2d(hidden_dim, 2, 3, padding=1)
self.relu = nn.ReLU(inplace=True)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
luyu94/RAFT
|
FlowHead
| false | 10,481 |
[
"BSD-3-Clause"
] | 0 |
d0a37db031af49a5d0d9b524d214acc989becf5b
|
https://github.com/luyu94/RAFT/tree/d0a37db031af49a5d0d9b524d214acc989becf5b
|
DPLSTMCell
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/cm/ccmo7fnacrzz2bpb3kjhjwznydxc2ydumd6dycmnumtbvfyp23ld.py
# Topologically Sorted Source Nodes: [i_t, f_t, g_t, o_t, mul, mul_1, c_t, tanh_1, h_t], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add, aten.sigmoid_backward]
# Source node to ATen node mapping:
# c_t => add_1
# f_t => sigmoid_1
# g_t => tanh
# h_t => mul_2
# i_t => sigmoid
# mul => mul
# mul_1 => mul_1
# o_t => sigmoid_2
# tanh_1 => tanh_1
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem,), kwargs = {})
# %sigmoid_1 : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_1,), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%getitem_2,), kwargs = {})
# %sigmoid_2 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_3,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %primals_7), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %tanh), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
# %tanh_1 : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_1,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_2, %tanh_1), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid_1), kwargs = {})
# %mul_14 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %sub_3), kwargs = {})
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 17, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask)
tmp9 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (12 + x0 + (16*x1)), xmask)
tmp12 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + (16*x1)), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + (16*x1)), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask)
tmp25 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (4 + x0 + (16*x1)), xmask)
tmp28 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = tl.sigmoid(tmp30)
tmp33 = tmp31 * tmp32
tmp34 = tmp7 * tmp23
tmp35 = tmp33 + tmp34
tmp36 = 1.0
tmp37 = tmp36 - tmp31
tmp38 = tmp31 * tmp37
tmp39 = libdevice.tanh(tmp35)
tmp40 = tmp15 * tmp39
tl.store(out_ptr0 + (x2), tmp7, xmask)
tl.store(out_ptr1 + (x2), tmp15, xmask)
tl.store(out_ptr2 + (x2), tmp23, xmask)
tl.store(out_ptr3 + (x2), tmp35, xmask)
tl.store(out_ptr4 + (x2), tmp38, xmask)
tl.store(out_ptr5 + (x2), tmp40, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (16, 4), (4, 1))
assert_size_stride(primals_2, (16, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (16, 4), (4, 1))
assert_size_stride(primals_5, (16, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 16), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_6, reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_t, f_t, g_t, o_t, mul, mul_1, c_t, tanh_1, h_t], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add, aten.sigmoid_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0.run(buf0, primals_2, buf1, primals_5, primals_7, buf2, buf4, buf3, buf5, buf7, buf6, 16, grid=grid(16), stream=stream0)
del buf0
del buf1
del primals_2
del primals_5
return (buf6, buf5, primals_3, primals_6, primals_7, buf2, buf3, buf4, buf5, buf7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.parallel
from typing import Tuple
class LSTMLinear(nn.Linear):
"""
This function is the same as a nn.Linear layer, except that in the backward pass
the grad_samples get accumulated (instead of being concatenated as in the standard
nn.Linear)
"""
def __init__(self, in_features: 'int', out_features: 'int', bias:
'bool'=True):
super().__init__(in_features, out_features, bias)
class DPLSTMCell(nn.Module):
"""
Internal-only class. Implements *one* step of LSTM so that a LSTM layer can be seen as repeated
applications of this class.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.ih = LSTMLinear(input_size, 4 * hidden_size, bias=self.bias)
self.hh = LSTMLinear(hidden_size, 4 * hidden_size, bias=self.bias)
self.reset_parameters()
def reset_parameters(self):
"""
Resets parameters by initializing them from an uniform distribution.
"""
stdv = 1.0 / math.sqrt(self.hidden_size)
for weight in self.parameters():
nn.init.uniform_(weight, -stdv, stdv)
def forward(self, x: 'torch.Tensor', h_prev: 'torch.Tensor', c_prev:
'torch.Tensor') ->Tuple[torch.Tensor, torch.Tensor]:
gates = self.ih(x) + self.hh(h_prev)
i_t_input, f_t_input, g_t_input, o_t_input = torch.split(gates,
self.hidden_size, 1)
i_t = torch.sigmoid(i_t_input)
f_t = torch.sigmoid(f_t_input)
g_t = torch.tanh(g_t_input)
o_t = torch.sigmoid(o_t_input)
c_t = f_t * c_prev + i_t * g_t
h_t = o_t * torch.tanh(c_t)
return h_t, c_t
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4, 'bias': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import math
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.parallel
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0(in_ptr0,
in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2,
out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask)
tmp9 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (12 + x0 + 16 * x1), xmask)
tmp12 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + 16 * x1), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + 16 * x1), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask)
tmp25 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (4 + x0 + 16 * x1), xmask)
tmp28 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = tl.sigmoid(tmp30)
tmp33 = tmp31 * tmp32
tmp34 = tmp7 * tmp23
tmp35 = tmp33 + tmp34
tmp36 = 1.0
tmp37 = tmp36 - tmp31
tmp38 = tmp31 * tmp37
tmp39 = libdevice.tanh(tmp35)
tmp40 = tmp15 * tmp39
tl.store(out_ptr0 + x2, tmp7, xmask)
tl.store(out_ptr1 + x2, tmp15, xmask)
tl.store(out_ptr2 + x2, tmp23, xmask)
tl.store(out_ptr3 + x2, tmp35, xmask)
tl.store(out_ptr4 + x2, tmp38, xmask)
tl.store(out_ptr5 + x2, tmp40, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (16, 4), (4, 1))
assert_size_stride(primals_2, (16,), (1,))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (16, 4), (4, 1))
assert_size_stride(primals_5, (16,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 16),
(1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
extern_kernels.mm(primals_6, reinterpret_tensor(primals_4, (4, 16),
(1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_0[grid(16)](buf0
, primals_2, buf1, primals_5, primals_7, buf2, buf4, buf3, buf5,
buf7, buf6, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf0
del buf1
del primals_2
del primals_5
return (buf6, buf5, primals_3, primals_6, primals_7, buf2, buf3, buf4,
buf5, buf7)
class LSTMLinear(nn.Linear):
"""
This function is the same as a nn.Linear layer, except that in the backward pass
the grad_samples get accumulated (instead of being concatenated as in the standard
nn.Linear)
"""
def __init__(self, in_features: 'int', out_features: 'int', bias:
'bool'=True):
super().__init__(in_features, out_features, bias)
class DPLSTMCellNew(nn.Module):
"""
Internal-only class. Implements *one* step of LSTM so that a LSTM layer can be seen as repeated
applications of this class.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.ih = LSTMLinear(input_size, 4 * hidden_size, bias=self.bias)
self.hh = LSTMLinear(hidden_size, 4 * hidden_size, bias=self.bias)
self.reset_parameters()
def reset_parameters(self):
"""
Resets parameters by initializing them from an uniform distribution.
"""
stdv = 1.0 / math.sqrt(self.hidden_size)
for weight in self.parameters():
nn.init.uniform_(weight, -stdv, stdv)
def forward(self, input_0, input_1, input_2):
primals_1 = self.ih.weight
primals_2 = self.ih.bias
primals_4 = self.hh.weight
primals_5 = self.hh.bias
primals_3 = input_0
primals_6 = input_1
primals_7 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0], output[1]
|
madhavajay/opacus
|
DPLSTMCell
| false | 10,482 |
[
"Apache-2.0"
] | 0 |
7ae098764b4cf2388c66e263dd8d56bca0a290d0
|
https://github.com/madhavajay/opacus/tree/7ae098764b4cf2388c66e263dd8d56bca0a290d0
|
PositionWiseFeedForward
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/6s/c6shmuvjmq6zc4ifvdsynorwri47ra63qxa7jg3e7p6lw6xlqj5q.py
# Topologically Sorted Source Nodes: [mul, truediv, erf, add, mul_1], Original ATen: [aten.mul, aten.div, aten.erf, aten.add]
# Source node to ATen node mapping:
# add => add
# erf => erf
# mul => mul
# mul_1 => mul_1
# truediv => div
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_1, 1.4142135623730951), kwargs = {})
# %erf : [num_users=1] = call_function[target=torch.ops.aten.erf.default](args = (%div,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%erf, 1.0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %add), kwargs = {})
triton_poi_fused_add_div_erf_mul_0 = async_compile.triton('triton_poi_fused_add_div_erf_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_erf_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_erf_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp0 * tmp3
tmp5 = libdevice.erf(tmp4)
tmp6 = 1.0
tmp7 = tmp5 + tmp6
tmp8 = tmp2 * tmp7
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, truediv, erf, add, mul_1], Original ATen: [aten.mul, aten.div, aten.erf, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_erf_mul_0.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_5
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf0, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
from _paritybench_helpers import _mock_config
import math
import torch
import torch.nn as nn
def gelu(x):
"""Implementation of the gelu activation function by Hugging Face"""
return x * 0.5 * (1.0 + torch.erf(x / math.sqrt(2.0)))
class PositionWiseFeedForward(nn.Module):
""" FeedForward Neural Networks for each position """
def __init__(self, cfg):
super().__init__()
self.fc1 = nn.Linear(cfg.dim, cfg.dim_ff)
self.fc2 = nn.Linear(cfg.dim_ff, cfg.dim)
def forward(self, x):
return self.fc2(gelu(self.fc1(x)))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'cfg': _mock_config(dim=4, dim_ff=4)}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_div_erf_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp3 = 0.7071067811865475
tmp4 = tmp0 * tmp3
tmp5 = libdevice.erf(tmp4)
tmp6 = 1.0
tmp7 = tmp5 + tmp6
tmp8 = tmp2 * tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_erf_mul_0[grid(256)](buf0, buf1, 256,
XBLOCK=256, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf2)
del primals_5
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf0, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_4
def gelu(x):
"""Implementation of the gelu activation function by Hugging Face"""
return x * 0.5 * (1.0 + torch.erf(x / math.sqrt(2.0)))
class PositionWiseFeedForwardNew(nn.Module):
""" FeedForward Neural Networks for each position """
def __init__(self, cfg):
super().__init__()
self.fc1 = nn.Linear(cfg.dim, cfg.dim_ff)
self.fc2 = nn.Linear(cfg.dim_ff, cfg.dim)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
FengMingquan-sjtu/pytorchic-bert
|
PositionWiseFeedForward
| false | 10,483 |
[
"Apache-2.0"
] | 0 |
83d616fb9c7e1d5c3646f9b6267ca912e2616d65
|
https://github.com/FengMingquan-sjtu/pytorchic-bert/tree/83d616fb9c7e1d5c3646f9b6267ca912e2616d65
|
AttentionUnit
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/4u/c4uzfozwsgh3le5rfytdcr7fatpzxslwn6zrlgczuib4kxua4g25.py
# Topologically Sorted Source Nodes: [add, sumTanh], Original ATen: [aten.add, aten.tanh]
# Source node to ATen node mapping:
# add => add
# sumTanh => tanh
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%expand, %view_1), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add,), kwargs = {})
triton_poi_fused_add_tanh_0 = async_compile.triton('triton_poi_fused_add_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_tanh_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = (xindex // 16)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr0 + (x3), xmask)
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + (x3), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ts/ctscnzvbagjv4t25zui245b3recij5udu7nvujnr5rixcyo7elc6.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/k6/ck6fz3qsfeqgn5jtm4ugikmu7cwvvlq3jpttijbb5kdniicwtyz6.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (1, 4), (4, 1))
assert_size_stride(primals_8, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_4, reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf1)
del primals_5
buf2 = reinterpret_tensor(buf0, (4, 4, 4), (16, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [add, sumTanh], Original ATen: [aten.add, aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_add_tanh_0.run(buf2, buf1, primals_6, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
del primals_6
buf4 = reinterpret_tensor(buf1, (16, 1), (1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [vProj], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_8, reinterpret_tensor(buf2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf4)
del primals_8
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf4, buf5, 16, grid=grid(16), stream=stream0)
buf6 = reinterpret_tensor(buf4, (4, 4), (4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf5, buf6, 16, grid=grid(16), stream=stream0)
del buf5
return (buf6, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), primals_4, buf2, buf6, primals_7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
import torch.nn.functional as F
from torch.nn import init
class AttentionUnit(nn.Module):
def __init__(self, sDim, xDim, attDim):
super(AttentionUnit, self).__init__()
self.sDim = sDim
self.xDim = xDim
self.attDim = attDim
self.sEmbed = nn.Linear(sDim, attDim)
self.xEmbed = nn.Linear(xDim, attDim)
self.wEmbed = nn.Linear(attDim, 1)
def init_weights(self):
init.normal_(self.sEmbed.weight, std=0.01)
init.constant_(self.sEmbed.bias, 0)
init.normal_(self.xEmbed.weight, std=0.01)
init.constant_(self.xEmbed.bias, 0)
init.normal_(self.wEmbed.weight, std=0.01)
init.constant_(self.wEmbed.bias, 0)
def forward(self, x, sPrev):
batch_size, T, _ = x.size()
x = x.view(-1, self.xDim)
xProj = self.xEmbed(x)
xProj = xProj.view(batch_size, T, -1)
sPrev = sPrev.squeeze(0)
sProj = self.sEmbed(sPrev)
sProj = torch.unsqueeze(sProj, 1)
sProj = sProj.expand(batch_size, T, self.attDim)
sumTanh = torch.tanh(sProj + xProj)
sumTanh = sumTanh.view(-1, self.attDim)
vProj = self.wEmbed(sumTanh)
vProj = vProj.view(batch_size, T)
alpha = F.softmax(vProj, dim=1)
return alpha
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'sDim': 4, 'xDim': 4, 'attDim': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch import nn
from torch.nn import init
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex // 16
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr0 + x3, xmask)
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + x3, tmp7, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (1, 4), (4, 1))
assert_size_stride(primals_8, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_4, reinterpret_tensor(primals_5, (4, 4),
(1, 4), 0), out=buf1)
del primals_5
buf2 = reinterpret_tensor(buf0, (4, 4, 4), (16, 4, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_add_tanh_0[grid(64)](buf2, buf1, primals_6,
primals_3, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_3
del primals_6
buf4 = reinterpret_tensor(buf1, (16, 1), (1, 1), 0)
del buf1
extern_kernels.addmm(primals_8, reinterpret_tensor(buf2, (16, 4), (
4, 1), 0), reinterpret_tensor(primals_7, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf4)
del primals_8
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(16)](buf4, buf5, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf6 = reinterpret_tensor(buf4, (4, 4), (4, 1), 0)
del buf4
triton_poi_fused__softmax_2[grid(16)](buf5, buf6, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del buf5
return buf6, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), primals_4, buf2, buf6, primals_7
class AttentionUnitNew(nn.Module):
def __init__(self, sDim, xDim, attDim):
super(AttentionUnitNew, self).__init__()
self.sDim = sDim
self.xDim = xDim
self.attDim = attDim
self.sEmbed = nn.Linear(sDim, attDim)
self.xEmbed = nn.Linear(xDim, attDim)
self.wEmbed = nn.Linear(attDim, 1)
def init_weights(self):
init.normal_(self.sEmbed.weight, std=0.01)
init.constant_(self.sEmbed.bias, 0)
init.normal_(self.xEmbed.weight, std=0.01)
init.constant_(self.xEmbed.bias, 0)
init.normal_(self.wEmbed.weight, std=0.01)
init.constant_(self.wEmbed.bias, 0)
def forward(self, input_0, input_1):
primals_2 = self.sEmbed.weight
primals_3 = self.sEmbed.bias
primals_4 = self.xEmbed.weight
primals_6 = self.xEmbed.bias
primals_7 = self.wEmbed.weight
primals_8 = self.wEmbed.bias
primals_1 = input_0
primals_5 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8])
return output[0]
|
lohzhunyewcs/aster.pytorch
|
AttentionUnit
| false | 10,484 |
[
"MIT"
] | 0 |
9441d386135a73b1baa3ec8c505f5eba99c26905
|
https://github.com/lohzhunyewcs/aster.pytorch/tree/9441d386135a73b1baa3ec8c505f5eba99c26905
|
FeatureResizer
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/66/c666qvr725wwogti7syalhhjsndtv2n5sxzt6zi4wlesyjxocpic.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# x_1 => add, rsqrt, var_mean
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_1, [3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-12), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
triton_poi_fused_native_layer_norm_0 = async_compile.triton('triton_poi_fused_native_layer_norm_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-12
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + (x0), tmp8, xmask)
tl.store(out_ptr1 + (x0), tmp23, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/lh/clhh73owbiuj4adasmetdqsot2nlmw2ljupnw2q4yt3du76mikww.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# x_1 => add, add_1, mul, mul_1, rsqrt, sub, var_mean
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_1, [3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-12), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_1, %getitem_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %primals_4), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_5), kwargs = {})
triton_poi_fused_native_layer_norm_1 = async_compile.triton('triton_poi_fused_native_layer_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
stream0 = get_raw_stream(0)
triton_poi_fused_native_layer_norm_0.run(buf0, buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_1.run(buf0, buf1, buf2, primals_4, primals_5, buf3, 256, grid=grid(256), stream=stream0)
del buf1
del buf2
del primals_5
return (buf3, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.utils.data
import torch
import torch.nn
import torch.optim
import torch.utils
from torch import nn
import torch.distributed
class FeatureResizer(nn.Module):
"""
This class takes as input a set of embeddings of dimension C1 and outputs a set of
embedding of dimension C2, after a linear transformation, dropout and normalization (LN).
"""
def __init__(self, input_feat_size, output_feat_size, dropout, do_ln=True):
super().__init__()
self.do_ln = do_ln
self.fc = nn.Linear(input_feat_size, output_feat_size, bias=True)
self.layer_norm = nn.LayerNorm(output_feat_size, eps=1e-12)
self.dropout = nn.Dropout(dropout)
def forward(self, encoder_features):
x = self.fc(encoder_features)
if self.do_ln:
x = self.layer_norm(x)
output = self.dropout(x)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_feat_size': 4, 'output_feat_size': 4, 'dropout': 0.5}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.utils.data
import torch
import torch.nn
import torch.optim
import torch.utils
from torch import nn
import torch.distributed
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_native_layer_norm_0(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-12
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + x0, tmp8, xmask)
tl.store(out_ptr1 + x0, tmp23, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
get_raw_stream(0)
triton_poi_fused_native_layer_norm_0[grid(64)](buf0, buf1, buf2, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_native_layer_norm_1[grid(256)](buf0, buf1, buf2,
primals_4, primals_5, buf3, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del buf1
del buf2
del primals_5
return buf3, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf0
class FeatureResizerNew(nn.Module):
"""
This class takes as input a set of embeddings of dimension C1 and outputs a set of
embedding of dimension C2, after a linear transformation, dropout and normalization (LN).
"""
def __init__(self, input_feat_size, output_feat_size, dropout, do_ln=True):
super().__init__()
self.do_ln = do_ln
self.fc = nn.Linear(input_feat_size, output_feat_size, bias=True)
self.layer_norm = nn.LayerNorm(output_feat_size, eps=1e-12)
self.dropout = nn.Dropout(dropout)
def forward(self, input_0):
primals_1 = self.fc.weight
primals_2 = self.fc.bias
primals_4 = self.layer_norm.weight
primals_5 = self.layer_norm.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
mmaaz60/mdetr
|
FeatureResizer
| false | 10,485 |
[
"Apache-2.0"
] | 0 |
fe1394c67e76a6c7e521bbda77d8294714038a3a
|
https://github.com/mmaaz60/mdetr/tree/fe1394c67e76a6c7e521bbda77d8294714038a3a
|
PairwiseRankerModel
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/b6/cb6j3uzxk3hlq74h24e2ofv66auocp2fsayzgs46c5z7xwnji5sg.py
# Topologically Sorted Source Nodes: [query_doc_1_rep, query_doc_2_rep], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# query_doc_1_rep => cat
# query_doc_2_rep => cat_1
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %primals_2], 1), kwargs = {})
# %cat_1 : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %primals_5], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr1 + ((4*x1) + ((-4) + x0)), tmp6 & xmask, eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp11 = tl.load(in_ptr2 + ((4*x1) + ((-4) + x0)), tmp6 & xmask, eviction_policy='evict_last', other=0.0)
tmp12 = tl.where(tmp4, tmp5, tmp11)
tl.store(out_ptr0 + (x2), tmp10, xmask)
tl.store(out_ptr1 + (x2), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/2v/c2vcxjglklzwx6o2kqa6tmbd6f33y5rn3si52kju3aeqb5iwawxx.py
# Topologically Sorted Source Nodes: [compare], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# compare => cat_2
# Graph fragment:
# %cat_2 : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%sigmoid, %sigmoid_1], 1), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.sigmoid(tmp5)
tmp7 = tl.full(tmp6.shape, 0.0, tmp6.dtype)
tmp8 = tl.where(tmp4, tmp6, tmp7)
tmp9 = tmp0 >= tmp3
tmp10 = tl.full([1], 8, tl.int64)
tmp11 = tmp0 < tmp10
tmp12 = tl.load(in_ptr1 + ((4*x1) + ((-4) + x0)), tmp9 & xmask, eviction_policy='evict_last', other=0.0)
tmp13 = tl.sigmoid(tmp12)
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp9, tmp13, tmp14)
tmp16 = tl.where(tmp4, tmp8, tmp15)
tl.store(out_ptr0 + (x2), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/b7/cb7iq44xucvx4o4uio3etz5hrrkllxx5igr3vjyglpwcku6mi232.py
# Topologically Sorted Source Nodes: [sigmoid_2], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid_2 => sigmoid_2
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_7), kwargs = {})
# %sigmoid_2 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_sigmoid_2 = async_compile.triton('triton_poi_fused_sigmoid_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 8), (8, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (1, 8), (8, 1))
assert_size_stride(primals_7, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
buf2 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [query_doc_1_rep, query_doc_2_rep], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, primals_2, primals_5, buf0, buf2, 32, grid=grid(32), stream=stream0)
del primals_1
del primals_2
del primals_5
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3, (8, 4), (1, 8), 0), alpha=1, beta=1, out=buf1)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, buf2, reinterpret_tensor(primals_3, (8, 4), (1, 8), 0), alpha=1, beta=1, out=buf3)
del primals_3
del primals_4
buf4 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [compare], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf1, buf3, buf4, 32, grid=grid(32), stream=stream0)
buf5 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf4, reinterpret_tensor(primals_6, (8, 1), (1, 8), 0), out=buf5)
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [sigmoid_2], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_2.run(buf6, primals_7, 4, grid=grid(4), stream=stream0)
del primals_7
return (buf6, buf0, buf1, buf2, buf3, buf4, buf6, primals_6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.onnx
import torch.nn as nn
class PairwiseRankerModel(nn.Module):
def __init__(self, embedding_size):
super(PairwiseRankerModel, self).__init__()
self.query_doc_transform = torch.nn.Linear(in_features=
embedding_size * 2, out_features=embedding_size)
self.compare_transform = torch.nn.Linear(in_features=embedding_size *
2, out_features=1)
def forward(self, query_embedding, doc_1_embedding, doc_2_embedding):
query_doc_1_rep = torch.cat((query_embedding, doc_1_embedding), 1)
query_doc_1_rep = torch.sigmoid(self.query_doc_transform(
query_doc_1_rep))
query_doc_2_rep = torch.cat((query_embedding, doc_2_embedding), 1)
query_doc_2_rep = torch.sigmoid(self.query_doc_transform(
query_doc_2_rep))
compare = torch.cat((query_doc_1_rep, query_doc_2_rep), 1)
compare = self.compare_transform(compare)
return torch.sigmoid(compare)
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'embedding_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.onnx
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp9 = tl.load(in_ptr1 + (4 * x1 + (-4 + x0)), tmp6 & xmask,
eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp11 = tl.load(in_ptr2 + (4 * x1 + (-4 + x0)), tmp6 & xmask,
eviction_policy='evict_last', other=0.0)
tmp12 = tl.where(tmp4, tmp5, tmp11)
tl.store(out_ptr0 + x2, tmp10, xmask)
tl.store(out_ptr1 + x2, tmp12, xmask)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.sigmoid(tmp5)
tmp7 = tl.full(tmp6.shape, 0.0, tmp6.dtype)
tmp8 = tl.where(tmp4, tmp6, tmp7)
tmp9 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp12 = tl.load(in_ptr1 + (4 * x1 + (-4 + x0)), tmp9 & xmask,
eviction_policy='evict_last', other=0.0)
tmp13 = tl.sigmoid(tmp12)
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp9, tmp13, tmp14)
tmp16 = tl.where(tmp4, tmp8, tmp15)
tl.store(out_ptr0 + x2, tmp16, xmask)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + x0, tmp4, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 8), (8, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (1, 8), (8, 1))
assert_size_stride(primals_7, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
buf2 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(32)](primals_1, primals_2, primals_5,
buf0, buf2, 32, XBLOCK=32, num_warps=1, num_stages=1)
del primals_1
del primals_2
del primals_5
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3,
(8, 4), (1, 8), 0), alpha=1, beta=1, out=buf1)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_4, buf2, reinterpret_tensor(primals_3,
(8, 4), (1, 8), 0), alpha=1, beta=1, out=buf3)
del primals_3
del primals_4
buf4 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
triton_poi_fused_cat_1[grid(32)](buf1, buf3, buf4, 32, XBLOCK=32,
num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
extern_kernels.mm(buf4, reinterpret_tensor(primals_6, (8, 1), (1, 8
), 0), out=buf5)
buf6 = buf5
del buf5
triton_poi_fused_sigmoid_2[grid(4)](buf6, primals_7, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_7
return buf6, buf0, buf1, buf2, buf3, buf4, buf6, primals_6
class PairwiseRankerModelNew(nn.Module):
def __init__(self, embedding_size):
super(PairwiseRankerModelNew, self).__init__()
self.query_doc_transform = torch.nn.Linear(in_features=
embedding_size * 2, out_features=embedding_size)
self.compare_transform = torch.nn.Linear(in_features=embedding_size *
2, out_features=1)
def forward(self, input_0, input_1, input_2):
primals_3 = self.query_doc_transform.weight
primals_4 = self.query_doc_transform.bias
primals_6 = self.compare_transform.weight
primals_7 = self.compare_transform.bias
primals_1 = input_0
primals_2 = input_1
primals_5 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
|
mikhail-tsir/vespa-exloration
|
PairwiseRankerModel
| false | 10,486 |
[
"Apache-2.0"
] | 0 |
9bebc00acb43021fa60c6e144fe4f1fa1d7719fc
|
https://github.com/mikhail-tsir/vespa-exloration/tree/9bebc00acb43021fa60c6e144fe4f1fa1d7719fc
|
DNN_Classifier
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/5y/c5yq7wkgmmcygrawripwacy566sggsmh2mzk5izw35wk7ferohhu.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x2 = xindex % 1600
x3 = (xindex // 1600)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (1664*x3)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/xk/cxkugsynlmnyrjhah42fewrhwovuvurnuv2qimo2qhxq27wjmq7q.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# x_1 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x3), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/jf/cjfzp64ny4hf7wdw5wptah3hqv5fcsh5rrw4brz7uxcy6ad57n7h.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# x_1 => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (100, 4), (4, 1))
assert_size_stride(primals_2, (100, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 100), (100, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 100), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 100), (1600, 400, 100, 1), 0); del buf0 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf5, 6400, grid=grid(6400), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 100), (100, 1), 0), reinterpret_tensor(primals_4, (100, 4), (1, 100), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf3
return (buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 100), (100, 1), 0), buf4, primals_4, buf5, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((100, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((100, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 100), (100, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
class DNN_Classifier(torch.nn.Module):
def __init__(self, input_dim, nb_categories, hidden_dim=100):
super(DNN_Classifier, self).__init__()
self.fc_1 = nn.Linear(input_dim, hidden_dim)
self.fc_2 = nn.Linear(hidden_dim, nb_categories)
self.softmax = nn.Softmax(dim=1)
def forward(self, x_in):
x = torch.relu(self.fc_1(x_in))
x = self.softmax(self.fc_2(x))
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'nb_categories': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
x2 = xindex % 1600
x3 = xindex // 1600
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 1664 * x3), tmp6, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x3, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (100, 4), (4, 1))
assert_size_stride(primals_2, (100,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 100), (100, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 100), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 100), (1600, 400, 100, 1), 0)
del buf0
buf5 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(6400)](buf1,
primals_2, buf5, 6400, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 100),
(100, 1), 0), reinterpret_tensor(primals_4, (100, 4), (1, 100),
0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf2, buf3, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
triton_poi_fused__softmax_2[grid(256)](buf3, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf3
return buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 100), (100, 1), 0
), buf4, primals_4, buf5
class DNN_ClassifierNew(torch.nn.Module):
def __init__(self, input_dim, nb_categories, hidden_dim=100):
super(DNN_ClassifierNew, self).__init__()
self.fc_1 = nn.Linear(input_dim, hidden_dim)
self.fc_2 = nn.Linear(hidden_dim, nb_categories)
self.softmax = nn.Softmax(dim=1)
def forward(self, input_0):
primals_1 = self.fc_1.weight
primals_2 = self.fc_1.bias
primals_4 = self.fc_2.weight
primals_5 = self.fc_2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
mleila/AGNews_Document_Classifcation
|
DNN_Classifier
| false | 10,487 |
[
"MIT"
] | 0 |
1ff44edf1fcaaee582b79141a419d61df62da56e
|
https://github.com/mleila/AGNews_Document_Classifcation/tree/1ff44edf1fcaaee582b79141a419d61df62da56e
|
PositionEmbs
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/v4/cv47lmemhm4jgxapntmy6avnx2r4yydlpdjkih5r27hnltgiweqs.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.add]
# Source node to ATen node mapping:
# out => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_2, %primals_1), kwargs = {})
triton_poi_fused_add_0 = async_compile.triton('triton_poi_fused_add_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 320
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 20
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (1, 5, 4), (20, 4, 1))
assert_size_stride(primals_2, (4, 4, 5, 4), (80, 20, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 5, 4), (80, 20, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_0.run(primals_2, primals_1, buf0, 320, grid=grid(320), stream=stream0)
del primals_1
del primals_2
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((1, 5, 4), (20, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 5, 4), (80, 20, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class PositionEmbs(nn.Module):
def __init__(self, num_patches, emb_dim, dropout_rate=0.1):
super(PositionEmbs, self).__init__()
self.pos_embedding = nn.Parameter(torch.randn(1, num_patches + 1,
emb_dim))
if dropout_rate > 0:
self.dropout = nn.Dropout(dropout_rate)
else:
self.dropout = None
def forward(self, x):
out = x + self.pos_embedding
if self.dropout:
out = self.dropout(out)
return out
def get_inputs():
return [torch.rand([4, 4, 5, 4])]
def get_init_inputs():
return [[], {'num_patches': 4, 'emb_dim': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 320
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 20
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + x2, tmp2, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (1, 5, 4), (20, 4, 1))
assert_size_stride(primals_2, (4, 4, 5, 4), (80, 20, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 5, 4), (80, 20, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_0[grid(320)](primals_2, primals_1, buf0, 320,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_1
del primals_2
return buf0,
class PositionEmbsNew(nn.Module):
def __init__(self, num_patches, emb_dim, dropout_rate=0.1):
super(PositionEmbsNew, self).__init__()
self.pos_embedding = nn.Parameter(torch.randn(1, num_patches + 1,
emb_dim))
if dropout_rate > 0:
self.dropout = nn.Dropout(dropout_rate)
else:
self.dropout = None
def forward(self, input_0):
primals_1 = self.pos_embedding
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
|
longxianlei/UtilsTools
|
PositionEmbs
| false | 10,488 |
[
"MIT"
] | 0 |
f45c648eb679ed59bb512b61a1af52938e326ac3
|
https://github.com/longxianlei/UtilsTools/tree/f45c648eb679ed59bb512b61a1af52938e326ac3
|
FPNHead
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/yw/cywcz4pxnzyvlsoydzxcj5pzlu3i5g7qgj7guhgyvlrzkngzehmv.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/w3/cw3zrmvca7grv74jw3rs72gt6ae2mq5prncci372h3zwksqmyouw.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x_1 => relu_1
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_out_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_2, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
return (buf3, primals_1, primals_2, primals_3, buf1, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
class FPNHead(nn.Module):
def __init__(self, num_in, num_mid, num_out):
super().__init__()
self.block0 = nn.Conv2d(num_in, num_mid, kernel_size=3, padding=1,
bias=False)
self.block1 = nn.Conv2d(num_mid, num_out, kernel_size=3, padding=1,
bias=False)
def forward(self, x):
x = nn.functional.relu(self.block0(x), inplace=True)
x = nn.functional.relu(self.block1(x), inplace=True)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_in': 4, 'num_mid': 4, 'num_out': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(in_out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_out_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr0 + x0, tmp4, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_2, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(256)](buf1, 256, XBLOCK=256, num_warps
=4, num_stages=1)
buf2 = extern_kernels.convolution(buf1, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2
del buf2
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(256)](buf3, buf4,
256, XBLOCK=256, num_warps=4, num_stages=1)
return buf3, primals_1, primals_2, primals_3, buf1, buf4
class FPNHeadNew(nn.Module):
def __init__(self, num_in, num_mid, num_out):
super().__init__()
self.block0 = nn.Conv2d(num_in, num_mid, kernel_size=3, padding=1,
bias=False)
self.block1 = nn.Conv2d(num_mid, num_out, kernel_size=3, padding=1,
bias=False)
def forward(self, input_0):
primals_1 = self.block0.weight
primals_3 = self.block1.weight
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
|
lePossum/DeblurGANv2
|
FPNHead
| false | 10,489 |
[
"BSD-2-Clause"
] | 0 |
b02c86de98f98604e2416a3a6121110ede7a2de9
|
https://github.com/lePossum/DeblurGANv2/tree/b02c86de98f98604e2416a3a6121110ede7a2de9
|
ClipLayer
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/l4/cl4tubtxyw5ee5xzznmwlohggffmq2we62eygecmxpgmibijhpes.py
# Topologically Sorted Source Nodes: [norms, data], Original ATen: [aten.linalg_vector_norm, aten.mul]
# Source node to ATen node mapping:
# data => mul_1
# norms => pow_1, sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [-1]), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %view_1), kwargs = {})
# %copy_ : [num_users=1] = call_function[target=torch.ops.aten.copy_.default](args = (%arg0_1, %mul_1), kwargs = {})
triton_per_fused_linalg_vector_norm_mul_0 = async_compile.triton('triton_per_fused_linalg_vector_norm_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_linalg_vector_norm_mul_0', 'mutated_arg_names': ['in_ptr0', 'out_ptr2'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_linalg_vector_norm_mul_0(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp7 = tl.full([1, 1], 1, tl.int32)
tmp8 = tmp7 / tmp6
tmp9 = 4.0
tmp10 = tmp8 * tmp9
tmp11 = 1.0
tmp12 = triton_helpers.minimum(tmp10, tmp11)
tmp13 = tmp0 * tmp12
tl.store(out_ptr2 + (r1 + (64*x0)), tmp13, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [norms, data], Original ATen: [aten.linalg_vector_norm, aten.mul]
stream0 = get_raw_stream(0)
triton_per_fused_linalg_vector_norm_mul_0.run(arg0_1, arg0_1, 4, 64, grid=grid(4), stream=stream0)
return (arg0_1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
def clip_data(data, max_norm):
norms = torch.norm(data.reshape(data.shape[0], -1), dim=-1)
scale = (max_norm / norms).clamp(max=1.0)
data *= scale.reshape(-1, 1, 1, 1)
return data
class ClipLayer(nn.Module):
def __init__(self, max_norm):
super(ClipLayer, self).__init__()
self.max_norm = max_norm
def forward(self, x):
return clip_data(x, self.max_norm)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'max_norm': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_per_fused_linalg_vector_norm_mul_0(in_ptr0, out_ptr2, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp7 = tl.full([1, 1], 1, tl.int32)
tmp8 = tmp7 / tmp6
tmp9 = 4.0
tmp10 = tmp8 * tmp9
tmp11 = 1.0
tmp12 = triton_helpers.minimum(tmp10, tmp11)
tmp13 = tmp0 * tmp12
tl.store(out_ptr2 + (r1 + 64 * x0), tmp13, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
get_raw_stream(0)
triton_per_fused_linalg_vector_norm_mul_0[grid(4)](arg0_1, arg0_1,
4, 64, XBLOCK=1, num_warps=2, num_stages=1)
return arg0_1,
def clip_data(data, max_norm):
norms = torch.norm(data.reshape(data.shape[0], -1), dim=-1)
scale = (max_norm / norms).clamp(max=1.0)
data *= scale.reshape(-1, 1, 1, 1)
return data
class ClipLayerNew(nn.Module):
def __init__(self, max_norm):
super(ClipLayerNew, self).__init__()
self.max_norm = max_norm
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
lxuechen/Handcrafted-DP
|
ClipLayer
| false | 10,490 |
[
"MIT"
] | 0 |
64ca4759238027e307d8e88215a0a86fc8f3b395
|
https://github.com/lxuechen/Handcrafted-DP/tree/64ca4759238027e307d8e88215a0a86fc8f3b395
|
PolynomialEnvelope
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/hz/chzlrhhvdiw3cqqyi2kylodpwhcjuju7rbif4i23erkdoy4vsy3m.py
# Topologically Sorted Source Nodes: [lt, pow_1, mul, add, pow_2, mul_1, add_1, pow_3, mul_2, env_val, zeros_like, where], Original ATen: [aten.lt, aten.pow, aten.mul, aten.add, aten.zeros_like, aten.where]
# Source node to ATen node mapping:
# add => add
# add_1 => add_1
# env_val => add_2
# lt => lt
# mul => mul
# mul_1 => mul_1
# mul_2 => mul_2
# pow_1 => pow_1
# pow_2 => pow_2
# pow_3 => pow_3
# where => where
# zeros_like => full_default
# Graph fragment:
# %lt : [num_users=1] = call_function[target=torch.ops.aten.lt.Scalar](args = (%arg0_1, 1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 4), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_1, -15.0), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 1), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 5), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_2, 24), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %mul_1), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 6), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_3, -10.0), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %mul_2), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%lt, %add_2, %full_default), kwargs = {})
triton_poi_fused_add_lt_mul_pow_where_zeros_like_0 = async_compile.triton('triton_poi_fused_add_lt_mul_pow_where_zeros_like_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_lt_mul_pow_where_zeros_like_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_lt_mul_pow_where_zeros_like_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 < tmp1
tmp3 = tmp0 * tmp0
tmp4 = tmp3 * tmp3
tmp5 = -15.0
tmp6 = tmp4 * tmp5
tmp7 = tmp6 + tmp1
tmp8 = tmp4 * tmp0
tmp9 = 24.0
tmp10 = tmp8 * tmp9
tmp11 = tmp7 + tmp10
tmp12 = tmp3 * tmp0
tmp13 = tmp12 * tmp12
tmp14 = -10.0
tmp15 = tmp13 * tmp14
tmp16 = tmp11 + tmp15
tmp17 = 0.0
tmp18 = tl.where(tmp2, tmp16, tmp17)
tl.store(out_ptr0 + (x0), tmp18, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [lt, pow_1, mul, add, pow_2, mul_1, add_1, pow_3, mul_2, env_val, zeros_like, where], Original ATen: [aten.lt, aten.pow, aten.mul, aten.add, aten.zeros_like, aten.where]
stream0 = get_raw_stream(0)
triton_poi_fused_add_lt_mul_pow_where_zeros_like_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
class PolynomialEnvelope(torch.nn.Module):
"""
Polynomial envelope function that ensures a smooth cutoff.
Parameters
----------
exponent: int
Exponent of the envelope function.
"""
def __init__(self, exponent):
super().__init__()
assert exponent > 0
self.p = exponent
self.a = -(self.p + 1) * (self.p + 2) / 2
self.b = self.p * (self.p + 2)
self.c = -self.p * (self.p + 1) / 2
def forward(self, d_scaled):
env_val = 1 + self.a * d_scaled ** self.p + self.b * d_scaled ** (self
.p + 1) + self.c * d_scaled ** (self.p + 2)
return torch.where(d_scaled < 1, env_val, torch.zeros_like(d_scaled))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'exponent': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_lt_mul_pow_where_zeros_like_0(in_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 < tmp1
tmp3 = tmp0 * tmp0
tmp4 = tmp3 * tmp3
tmp5 = -15.0
tmp6 = tmp4 * tmp5
tmp7 = tmp6 + tmp1
tmp8 = tmp4 * tmp0
tmp9 = 24.0
tmp10 = tmp8 * tmp9
tmp11 = tmp7 + tmp10
tmp12 = tmp3 * tmp0
tmp13 = tmp12 * tmp12
tmp14 = -10.0
tmp15 = tmp13 * tmp14
tmp16 = tmp11 + tmp15
tmp17 = 0.0
tmp18 = tl.where(tmp2, tmp16, tmp17)
tl.store(out_ptr0 + x0, tmp18, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_lt_mul_pow_where_zeros_like_0[grid(256)](arg0_1,
buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class PolynomialEnvelopeNew(torch.nn.Module):
"""
Polynomial envelope function that ensures a smooth cutoff.
Parameters
----------
exponent: int
Exponent of the envelope function.
"""
def __init__(self, exponent):
super().__init__()
assert exponent > 0
self.p = exponent
self.a = -(self.p + 1) * (self.p + 2) / 2
self.b = self.p * (self.p + 2)
self.c = -self.p * (self.p + 1) / 2
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
krylea/ocp
|
PolynomialEnvelope
| false | 10,491 |
[
"MIT"
] | 0 |
00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
https://github.com/krylea/ocp/tree/00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
ScaledSiLU
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/tb/ctba7iaosb7jgdxzd2uonnb4zpebevarpcjx6lt3orlpglpmb64v.py
# Topologically Sorted Source Nodes: [silu, mul], Original ATen: [aten.silu, aten.mul]
# Source node to ATen node mapping:
# mul => mul_1
# silu => mul, sigmoid
# Graph fragment:
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%arg0_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %sigmoid), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, 1.6666666666666667), kwargs = {})
triton_poi_fused_mul_silu_0 = async_compile.triton('triton_poi_fused_mul_silu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_silu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_silu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tmp3 = 1.6666666666666667
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [silu, mul], Original ATen: [aten.silu, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_silu_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
class ScaledSiLU(torch.nn.Module):
def __init__(self):
super().__init__()
self.scale_factor = 1 / 0.6
self._activation = torch.nn.SiLU()
def forward(self, x):
return self._activation(x) * self.scale_factor
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_silu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tmp3 = 1.6666666666666667
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + x0, tmp4, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_silu_0[grid(256)](arg0_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class ScaledSiLUNew(torch.nn.Module):
def __init__(self):
super().__init__()
self.scale_factor = 1 / 0.6
self._activation = torch.nn.SiLU()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
krylea/ocp
|
ScaledSiLU
| false | 10,492 |
[
"MIT"
] | 0 |
00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
https://github.com/krylea/ocp/tree/00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
SiQU
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/e3/ce3flp2f3766ehq6qmatay7ybft362x5j23kxeaedseuzqtovc4g.py
# Topologically Sorted Source Nodes: [silu, mul], Original ATen: [aten.silu, aten.mul]
# Source node to ATen node mapping:
# mul => mul_1
# silu => mul, sigmoid
# Graph fragment:
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%arg0_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %sigmoid), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %mul), kwargs = {})
triton_poi_fused_mul_silu_0 = async_compile.triton('triton_poi_fused_mul_silu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_silu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_silu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [silu, mul], Original ATen: [aten.silu, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_silu_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
class SiQU(torch.nn.Module):
def __init__(self):
super().__init__()
self._activation = torch.nn.SiLU()
def forward(self, x):
return x * self._activation(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_silu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + x0, tmp3, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_silu_0[grid(256)](arg0_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class SiQUNew(torch.nn.Module):
def __init__(self):
super().__init__()
self._activation = torch.nn.SiLU()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
krylea/ocp
|
SiQU
| false | 10,493 |
[
"MIT"
] | 0 |
00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
https://github.com/krylea/ocp/tree/00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
DPSLTMAdapter
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/l4/cl4boort6vfsvh6h6bfd4lck36kbmtipkqcrnhckuuxer6sfib77.py
# Topologically Sorted Source Nodes: [h_0s], Original ATen: [aten.zeros]
# Source node to ATen node mapping:
# h_0s => full_default
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([1, 1, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
triton_poi_fused_zeros_0 = async_compile.triton('triton_poi_fused_zeros_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_zeros_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_zeros_0(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = 0.0
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/sw/cswlmqxq57h53l6axfd6psb5uckvpd32zawfihpwp3s4owvqpcsb.py
# Topologically Sorted Source Nodes: [i_t, f_t, g_t, o_t, mul, mul_1, c_t, tanh_1, h_t], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add, aten.sigmoid_backward]
# Source node to ATen node mapping:
# c_t => add_1
# f_t => sigmoid_1
# g_t => tanh
# h_t => mul_2
# i_t => sigmoid
# mul => mul
# mul_1 => mul_1
# o_t => sigmoid_2
# tanh_1 => tanh_1
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_4,), kwargs = {})
# %sigmoid_1 : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_5,), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%getitem_6,), kwargs = {})
# %sigmoid_2 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_7,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %squeeze), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %tanh), kwargs = {})
# %add_1 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
# %tanh_1 : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_1,), kwargs = {})
# %mul_2 : [num_users=3] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_2, %tanh_1), kwargs = {})
# %sub_18 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid_1), kwargs = {})
# %mul_71 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %sub_18), kwargs = {})
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 17, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask)
tmp9 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (12 + x0 + (16*x1)), xmask)
tmp12 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + (16*x1)), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + (16*x1)), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask)
tmp25 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (4 + x0 + (16*x1)), xmask)
tmp28 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = tl.sigmoid(tmp30)
tmp33 = tmp31 * tmp32
tmp34 = tmp7 * tmp23
tmp35 = tmp33 + tmp34
tmp36 = 1.0
tmp37 = tmp36 - tmp31
tmp38 = tmp31 * tmp37
tmp39 = libdevice.tanh(tmp35)
tmp40 = tmp15 * tmp39
tl.store(out_ptr0 + (x2), tmp7, xmask)
tl.store(out_ptr1 + (x2), tmp15, xmask)
tl.store(out_ptr2 + (x2), tmp23, xmask)
tl.store(out_ptr3 + (x2), tmp35, xmask)
tl.store(out_ptr4 + (x2), tmp38, xmask)
tl.store(out_ptr5 + (x2), tmp40, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/tl/ctlt2p573ni7yx4xldsz64sjwqkudoxkn6d6yvb2r54qxlmxds5p.py
# Topologically Sorted Source Nodes: [i_t_1, f_t_1, g_t_1, o_t_1, mul_3, mul_4, c_t_1, tanh_3, h_t_1], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add]
# Source node to ATen node mapping:
# c_t_1 => add_3
# f_t_1 => sigmoid_4
# g_t_1 => tanh_2
# h_t_1 => mul_5
# i_t_1 => sigmoid_3
# mul_3 => mul_3
# mul_4 => mul_4
# o_t_1 => sigmoid_5
# tanh_3 => tanh_3
# Graph fragment:
# %sigmoid_3 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_8,), kwargs = {})
# %sigmoid_4 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_9,), kwargs = {})
# %tanh_2 : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%getitem_10,), kwargs = {})
# %sigmoid_5 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_11,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_4, %add_1), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_3, %tanh_2), kwargs = {})
# %add_3 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_3, %mul_4), kwargs = {})
# %tanh_3 : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_3,), kwargs = {})
# %mul_5 : [num_users=3] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_5, %tanh_3), kwargs = {})
triton_poi_fused_add_mul_sigmoid_tanh_2 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_tanh_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_tanh_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 17, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (4 + x0 + (16*x1)), xmask)
tmp12 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask)
tmp17 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (12 + x0 + (16*x1)), xmask)
tmp20 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (8 + x0 + (16*x1)), xmask)
tmp25 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (8 + x0 + (16*x1)), xmask)
tmp28 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = tl.sigmoid(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = libdevice.tanh(tmp30)
tmp33 = tmp15 * tmp32
tmp34 = tmp7 * tmp31
tmp35 = tmp33 + tmp34
tmp36 = libdevice.tanh(tmp35)
tmp37 = tmp23 * tmp36
tl.store(out_ptr0 + (x2), tmp7, xmask)
tl.store(out_ptr1 + (x2), tmp15, xmask)
tl.store(out_ptr2 + (x2), tmp23, xmask)
tl.store(out_ptr3 + (x2), tmp31, xmask)
tl.store(out_ptr4 + (x2), tmp35, xmask)
tl.store(out_ptr5 + (x2), tmp37, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/ks/cksi7m4m7yxvsvd2ce7ftnos6exaibxyudnraey6nxajdfw246hj.py
# Topologically Sorted Source Nodes: [i_t_3, f_t_3, g_t_3, mul_9, mul_10, c_t_3, tanh_7], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add]
# Source node to ATen node mapping:
# c_t_3 => add_7
# f_t_3 => sigmoid_10
# g_t_3 => tanh_6
# i_t_3 => sigmoid_9
# mul_10 => mul_10
# mul_9 => mul_9
# tanh_7 => tanh_7
# Graph fragment:
# %sigmoid_9 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_16,), kwargs = {})
# %sigmoid_10 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_17,), kwargs = {})
# %tanh_6 : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%getitem_18,), kwargs = {})
# %mul_9 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_10, %add_5), kwargs = {})
# %mul_10 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_9, %tanh_6), kwargs = {})
# %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_9, %mul_10), kwargs = {})
# %tanh_7 : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_7,), kwargs = {})
triton_poi_fused_add_mul_sigmoid_tanh_3 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_tanh_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_tanh_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 13, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (4 + x0 + (16*x1)), xmask)
tmp12 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + (16*x1)), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + (16*x1)), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp25 = tmp15 * tmp24
tmp26 = tmp7 * tmp23
tmp27 = tmp25 + tmp26
tmp28 = libdevice.tanh(tmp27)
tl.store(out_ptr0 + (x2), tmp7, xmask)
tl.store(out_ptr1 + (x2), tmp15, xmask)
tl.store(out_ptr2 + (x2), tmp23, xmask)
tl.store(out_ptr3 + (x2), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/o5/co5z4awunowwwrh5of536l26ozrjuwdhph5f2zpgdrfapcc3dduk.py
# Topologically Sorted Source Nodes: [o_t_3], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# o_t_3 => sigmoid_11
# Graph fragment:
# %sigmoid_11 : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_19,), kwargs = {})
triton_poi_fused_sigmoid_4 = async_compile.triton('triton_poi_fused_sigmoid_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (12 + x0 + (16*x1)), xmask)
tmp4 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tl.store(out_ptr0 + (x2), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/vm/cvmcreohmmmgtc2sngri4exy6k5t5hutelgjbv763z3d5cnripkm.py
# Topologically Sorted Source Nodes: [h_n], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# h_n => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mul_2, %mul_5, %mul_8, %mul_11],), kwargs = {})
triton_poi_fused_stack_5 = async_compile.triton('triton_poi_fused_stack_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_stack_5(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x0 = xindex % 4
x2 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (4*x1)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + (4*((-4) + x1))), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + (4*((-8) + x1))), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tl.load(in_ptr3 + (x0 + (4*((-12) + x1))), tmp16 & xmask, other=0.0)
tmp20 = tl.load(in_ptr4 + (x0 + (4*((-12) + x1))), tmp16 & xmask, other=0.0)
tmp21 = tmp19 * tmp20
tmp22 = tl.full(tmp21.shape, 0.0, tmp21.dtype)
tmp23 = tl.where(tmp16, tmp21, tmp22)
tmp24 = tl.where(tmp14, tmp15, tmp23)
tmp25 = tl.where(tmp9, tmp10, tmp24)
tmp26 = tl.where(tmp4, tmp5, tmp25)
tl.store(out_ptr0 + (x2), tmp26, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (16, 4), (4, 1))
assert_size_stride(primals_3, (16, ), (1, ))
assert_size_stride(primals_4, (16, 4), (4, 1))
assert_size_stride(primals_5, (16, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((1, 1, 4, 4), (16, 1, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [h_0s], Original ATen: [aten.zeros]
stream0 = get_raw_stream(0)
triton_poi_fused_zeros_0.run(buf0, 16, grid=grid(16), stream=stream0)
buf1 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (4, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf32 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_t, f_t, g_t, o_t, mul, mul_1, c_t, tanh_1, h_t], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add, aten.sigmoid_backward]
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1.run(buf1, primals_3, buf2, primals_5, buf0, buf3, buf5, buf4, buf6, buf32, buf7, 16, grid=grid(16), stream=stream0)
buf8 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 16), reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf8)
buf9 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf7, reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf11 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf13 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf14 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_t_1, f_t_1, g_t_1, o_t_1, mul_3, mul_4, c_t_1, tanh_3, h_t_1], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add]
triton_poi_fused_add_mul_sigmoid_tanh_2.run(buf8, primals_3, buf9, primals_5, buf6, buf10, buf11, buf13, buf12, buf14, buf15, 16, grid=grid(16), stream=stream0)
buf16 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 32), reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf16)
buf17 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf15, reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf17)
buf18 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf19 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf21 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf20 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf22 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf23 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_t_2, f_t_2, g_t_2, o_t_2, mul_6, mul_7, c_t_2, tanh_5, h_t_2], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add]
triton_poi_fused_add_mul_sigmoid_tanh_2.run(buf16, primals_3, buf17, primals_5, buf14, buf18, buf19, buf21, buf20, buf22, buf23, 16, grid=grid(16), stream=stream0)
buf24 = buf17; del buf17 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 48), reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf24)
del primals_2
buf25 = buf16; del buf16 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf23, reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf25)
buf26 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf27 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf28 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf30 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [i_t_3, f_t_3, g_t_3, mul_9, mul_10, c_t_3, tanh_7], Original ATen: [aten.sigmoid, aten.tanh, aten.mul, aten.add]
triton_poi_fused_add_mul_sigmoid_tanh_3.run(buf24, primals_3, buf25, primals_5, buf22, buf26, buf27, buf28, buf30, 16, grid=grid(16), stream=stream0)
buf29 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [o_t_3], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_4.run(buf24, primals_3, buf25, primals_5, buf29, 16, grid=grid(16), stream=stream0)
del buf24
del primals_3
del primals_5
buf31 = reinterpret_tensor(buf25, (16, 4), (4, 1), 0); del buf25 # reuse
# Topologically Sorted Source Nodes: [h_n], Original ATen: [aten.stack]
triton_poi_fused_stack_5.run(buf7, buf15, buf23, buf29, buf30, buf31, 64, grid=grid(64), stream=stream0)
return (reinterpret_tensor(buf31, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf0, (4, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (4, 1), 16), reinterpret_tensor(primals_1, (4, 4), (4, 1), 32), reinterpret_tensor(primals_1, (4, 4), (4, 1), 48), buf3, buf4, buf5, buf6, buf7, buf10, buf11, buf12, buf13, buf14, buf15, buf18, buf19, buf20, buf21, buf22, buf23, buf26, buf27, buf28, buf29, buf30, primals_4, buf32, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
from torch import Tensor
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.parallel
from typing import Tuple
from typing import List
from typing import Optional
from typing import Dict
from typing import Union
from torch.nn.modules.module import _IncompatibleKeys
def filter_out_old_keys(self, state_dict, prefix, local_metadata):
new_state_dict = {param_name: param_value for param_name, param_value in
state_dict.items() if param_name not in self.old_to_new}
return new_state_dict
class LSTMLinear(nn.Linear):
"""
This function is the same as a nn.Linear layer, except that in the backward pass
the grad_samples get accumulated (instead of being concatenated as in the standard
nn.Linear)
"""
def __init__(self, in_features: 'int', out_features: 'int', bias:
'bool'=True):
super().__init__(in_features, out_features, bias)
class DPLSTMCell(nn.Module):
"""
Internal-only class. Implements *one* step of LSTM so that a LSTM layer can be seen as repeated
applications of this class.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.ih = LSTMLinear(input_size, 4 * hidden_size, bias=self.bias)
self.hh = LSTMLinear(hidden_size, 4 * hidden_size, bias=self.bias)
self.reset_parameters()
def reset_parameters(self):
"""
Resets parameters by initializing them from an uniform distribution.
"""
stdv = 1.0 / math.sqrt(self.hidden_size)
for weight in self.parameters():
nn.init.uniform_(weight, -stdv, stdv)
def forward(self, x: 'torch.Tensor', h_prev: 'torch.Tensor', c_prev:
'torch.Tensor') ->Tuple[torch.Tensor, torch.Tensor]:
gates = self.ih(x) + self.hh(h_prev)
i_t_input, f_t_input, g_t_input, o_t_input = torch.split(gates,
self.hidden_size, 1)
i_t = torch.sigmoid(i_t_input)
f_t = torch.sigmoid(f_t_input)
g_t = torch.tanh(g_t_input)
o_t = torch.sigmoid(o_t_input)
c_t = f_t * c_prev + i_t * g_t
h_t = o_t * torch.tanh(c_t)
return h_t, c_t
class DPLSTMLayer(nn.Module):
"""
Implements *one* layer of LSTM in a way amenable to differential privacy.
We don't expect you to use this directly: use DPLSTM instead :)
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool',
dropout: 'float', reverse: 'bool'=False):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.dropout = dropout
self.reverse = reverse
self.cell = DPLSTMCell(input_size=input_size, hidden_size=
hidden_size, bias=bias)
self.dropout_layer = nn.Dropout(dropout) if dropout > 0 else None
def forward(self, x: 'torch.Tensor', state_init:
'Tuple[torch.Tensor, torch.Tensor]') ->Tuple[torch.Tensor, Tuple[
torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTMLayer when a sequence is given in input.
Args:
x: Input sequence to the DPLSTMCell of shape ``[T, B, D]``.
state_init: Initial state of the LSTMCell as a tuple ``(h_0, c_0)``
where ``h_0`` is the initial hidden state and ``c_0`` is the
initial cell state of the DPLSTMCell
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTMCell for each timestep ``t``.
- ``h_n`` is of shape ``[B, H]`` and is a tensor containing the hidden state for ``t = T``.
- ``c_n`` is of shape ``[B, H]`` tensor containing the cell state for ``t = T``.
"""
seq_length, _batch_sz, _ = x.shape
if self.reverse:
x = x.flip(0)
x = torch.unbind(x, dim=0)
h_0, c_0 = state_init
h_n = [h_0]
c_n = [c_0]
for t in range(seq_length):
h_next, c_next = self.cell(x[t], h_n[t], c_n[t])
if self.dropout:
h_next = self.dropout_layer(h_next)
h_n.append(h_next)
c_n.append(c_next)
h_n = torch.stack(h_n[1:], dim=0)
return h_n.flip(0) if self.reverse else h_n, (h_n[-1], c_n[-1])
class BidirectionalDPLSTMLayer(nn.Module):
"""
Implements *one* layer of Bidirectional LSTM in a way amenable to differential privacy.
We don't expect you to use this directly: use DPLSTM instead :)
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool',
dropout: 'float'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.dropout = dropout
self.forward_layer = DPLSTMLayer(input_size=input_size, hidden_size
=hidden_size, bias=bias, dropout=dropout, reverse=False)
self.reverse_layer = DPLSTMLayer(input_size=input_size, hidden_size
=hidden_size, bias=bias, dropout=dropout, reverse=True)
def forward(self, x: 'torch.Tensor', state_init:
'Tuple[torch.Tensor, torch.Tensor]') ->Tuple[torch.Tensor, Tuple[
torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTM when a sequence is input.
Dimensions as follows:
- B: Batch size
- T: Sequence length
- D: LSTM input hidden size (eg from a word embedding)
- H: LSTM output hidden size
- P: number of directions (2 if bidirectional, else 1)
Args:
x: Input sequence to the DPLSTM of shape ``[T, B, D]``
state_init: Initial state of the LSTM as a tuple ``(h_0, c_0)``, where:
- h_0 of shape ``[P, B, H] contains the initial hidden state
- c_0 of shape ``[P, B, H] contains the initial cell state
This argument can be (and defaults to) None, in which case zero tensors will be used.
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H * P]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTM for each timestep ``t``.
- ``h_n`` is of shape ``[P, B, H]`` and contains the hidden state for ``t = T``.
- ``c_n`` is of shape ``[P, B, H]`` and contains the cell state for ``t = T``.
"""
h0, c0 = state_init
h0_f, h0_r = h0.unbind(0)
c0_f, c0_r = c0.unbind(0)
out_f, (h_f, c_f) = self.forward_layer(x, (h0_f, c0_f))
out_r, (h_r, c_r) = self.reverse_layer(x, (h0_r, c0_r))
out = torch.cat([out_f, out_r], dim=-1)
h = torch.stack([h_f, h_r], dim=0)
c = torch.stack([c_f, c_r], dim=0)
return out, (h, c)
class ParamRenamedModule(nn.Module):
"""
This class defines a nn.Module whose parameters are renamed. This is useful when you want to
reimplement a layer but make sure its state_dict and list of parameters are exactly the same
as another reference layer so that you can have a drop-in replacement that does not depend on
how your layer is actually implemented. In Opacus, this is used for DPLSTM, where our
implementation leverages submodules and requires alignment to the state_dict of nn.LSTM.
"""
def __init__(self, rename_map: 'Dict[str, str]'):
"""
Initializes internal state. Subclass this instead of ``torch.nn.Module`` whenever you need
to rename your model's state.
Args:
rename_map: mapping from old name -> new name for each parameter you want renamed.
Note that this must be a 1:1 mapping!
"""
super().__init__()
self.old_to_new = rename_map
self.new_to_old = {v: k for k, v in rename_map.items()}
self._register_state_dict_hook(filter_out_old_keys)
def _register_renamed_parameters(self):
"""
Internal function. This function simply registers parameters under their new name. They will
automatically mask their duplicates coming from submodules. This trick works because
self.parameters() proceeds recursively from the top, going into submodules after processing
items at the current level, and will not return duplicates.
"""
for old_name, param in super().named_parameters():
if old_name in self.old_to_new:
new_name = self.old_to_new[old_name]
self.register_parameter(new_name, param)
def __setattr__(self, name: 'str', value: 'Union[Tensor, nn.Module]'
) ->None:
"""
Whenever you set an attribute, eg `self.linear`, this is called to actually register it in
any nn.Module. We rely on the masking trick explained in the docs for
``_register_renamed_parameters`` to make sure we replace things only once. If a new parameter
in the rename list is detected, we rename and mask it so next time this is called we will
no longer find it.
"""
super().__setattr__(name, value)
try:
self._register_renamed_parameters()
except ModuleAttributeError:
pass
def load_state_dict(self, state_dict: 'Dict[str, Tensor]', strict:
'bool'=True):
"""
Identical to ``torch.nn.Module.load_state_dict()`` but handles the renamed keys.
"""
missing_keys, unexpected_keys = super().load_state_dict(state_dict,
strict=False)
missing_keys = [k for k in missing_keys if k not in self.old_to_new]
if strict:
error_msgs = []
if len(unexpected_keys) > 0:
error_msgs.insert(0,
'Unexpected key(s) in state_dict: {}. '.format(', '.
join('"{}"'.format(k) for k in unexpected_keys)))
if len(missing_keys) > 0:
error_msgs.insert(0, 'Missing key(s) in state_dict: {}. '.
format(', '.join('"{}"'.format(k) for k in missing_keys)))
if len(error_msgs) > 0:
raise RuntimeError(
'Error(s) in loading state_dict for {}:\n\t{}'.format(
self.__class__.__name__, '\n\t'.join(error_msgs)))
return _IncompatibleKeys(missing_keys, unexpected_keys)
class DPLSTM(ParamRenamedModule):
"""
DP-friendly drop-in replacement of the ``torch.nn.LSTM`` module.
Its state_dict matches that of nn.LSTM exactly, so that after training it can be exported
and loaded by an nn.LSTM for inference.
Refer to nn.LSTM's documentation for all parameters and inputs.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', num_layers:
'int'=1, bias: 'bool'=True, batch_first: 'bool'=False, dropout:
'float'=0, bidirectional: 'bool'=False):
rename_dict = self._make_rename_dict(num_layers, bias, bidirectional)
super().__init__(rename_dict)
self.input_size = input_size
self.hidden_size = hidden_size
self.num_layers = num_layers
self.bias = bias
self.batch_first = batch_first
self.dropout = dropout
self.bidirectional = bidirectional
self.num_directions = 2 if self.bidirectional else 1
LayerClass = BidirectionalDPLSTMLayer if bidirectional else DPLSTMLayer
self.layers = nn.ModuleList([LayerClass(input_size=self.input_size if
i == 0 else self.hidden_size * self.num_directions, hidden_size
=self.hidden_size, bias=self.bias, dropout=self.dropout if i <
self.num_layers - 1 else 0) for i in range(num_layers)])
def forward(self, x: 'torch.Tensor', state_init:
'Optional[Tuple[torch.Tensor, torch.Tensor]]'=None) ->Tuple[torch.
Tensor, Tuple[torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTM when a sequence is input.
Dimensions as follows:
- B: Batch size
- T: Sequence length
- D: LSTM input hidden size (eg from a word embedding)
- H: LSTM output hidden size
- L: number of layers in the LSTM
- P: number of directions (2 if bidirectional, else 1)
Args:
x: Input sequence to the DPLSTM of shape ``[T, B, D]``
state_init: Initial state of the LSTM as a tuple ``(h_0, c_0)``, where:
- h_0 of shape ``[L*P, B, H] contains the initial hidden state
- c_0 of shape ``[L*P, B, H] contains the initial cell state
This argument can be (and defaults to) None, in which case zero tensors will be used.
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H * P]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTM for each timestep ``t``.
- ``h_n`` is of shape ``[L * P, B, H]`` and contains the hidden state for ``t = T``.
- ``c_n`` is of shape ``[L * P, B, H]`` and contains the cell state for ``t = T``.
"""
x = self._rearrange_batch_dim(x)
_T, B, _D = x.shape
L = self.num_layers
P = 2 if self.bidirectional else 1
H = self.hidden_size
h_0s, c_0s = state_init or (None, None)
if h_0s is None:
h_0s = torch.zeros(L, P, B, self.hidden_size, dtype=x[0].dtype,
device=x[0].device)
else:
h_0s = h_0s.reshape([L, P, B, H])
if c_0s is None:
c_0s = torch.zeros(L, P, B, self.hidden_size, dtype=x[0].dtype,
device=x[0].device)
else:
c_0s = c_0s.reshape([L, P, B, H])
hs: 'List[torch.Tensor]' = []
cs: 'List[torch.Tensor]' = []
for layer, h0, c0 in zip(self.layers, h_0s, c_0s):
if not self.bidirectional:
h0 = h0.squeeze()
c0 = c0.squeeze()
x, (h, c) = layer(x, (h0, c0))
if not self.bidirectional:
h = h.unsqueeze(0)
c = c.unsqueeze(0)
hs.append(h)
cs.append(c)
hs = torch.cat(hs, dim=0)
cs = torch.cat(cs, dim=0)
out = self._rearrange_batch_dim(x)
return out, (hs, cs)
def _rearrange_batch_dim(self, x: 'torch.Tensor') ->torch.Tensor:
if self.batch_first:
x = x.transpose(0, 1)
return x
def __repr__(self):
s = f'DPLSTM({self.input_size}, {self.hidden_size}, bias={self.bias}'
if self.batch_first:
s += f', batch_first={self.batch_first}'
if self.num_layers > 1:
s += f', num_layers={self.num_layers}'
if self.dropout:
s += f', dropout={self.dropout}'
if self.bidirectional:
s += f', bidirectional={self.bidirectional}'
return s
def _make_rename_dict(self, num_layers, bias, bidirectional):
"""
Programmatically constructs a dictionary old_name -> new_name to align with the param
names used in ``torch.nn.LSTM``.
"""
d = {}
components = ['weight'] + ['bias' if bias else []]
matrices = ['ih', 'hh']
for i in range(num_layers):
for c in components:
for m in matrices:
nn_name = f'{c}_{m}_l{i}'
if bidirectional:
d[f'layers.{i}.forward_layer.cell.{m}.{c}'] = nn_name
d[f'layers.{i}.reverse_layer.cell.{m}.{c}'
] = nn_name + '_reverse'
else:
d[f'layers.{i}.cell.{m}.{c}'] = nn_name
return d
class DPSLTMAdapter(nn.Module):
"""
Adapter for DPLSTM.
LSTM returns a tuple, but our testing tools need the model to return a single tensor in output.
We do this adaption here.
"""
def __init__(self, *args, **kwargs):
super().__init__()
self.dplstm = DPLSTM(*args, **kwargs)
def forward(self, x):
out, _rest = self.dplstm(x)
return out
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
|
import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import math
from torch import Tensor
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.parallel
from typing import Tuple
from typing import List
from typing import Optional
from typing import Dict
from typing import Union
from torch.nn.modules.module import _IncompatibleKeys
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_zeros_0(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = 0.0
tl.store(out_ptr0 + x0, tmp0, xmask)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1(in_ptr0,
in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2,
out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask)
tmp9 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (12 + x0 + 16 * x1), xmask)
tmp12 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + 16 * x1), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + 16 * x1), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask)
tmp25 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (4 + x0 + 16 * x1), xmask)
tmp28 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = tl.sigmoid(tmp30)
tmp33 = tmp31 * tmp32
tmp34 = tmp7 * tmp23
tmp35 = tmp33 + tmp34
tmp36 = 1.0
tmp37 = tmp36 - tmp31
tmp38 = tmp31 * tmp37
tmp39 = libdevice.tanh(tmp35)
tmp40 = tmp15 * tmp39
tl.store(out_ptr0 + x2, tmp7, xmask)
tl.store(out_ptr1 + x2, tmp15, xmask)
tl.store(out_ptr2 + x2, tmp23, xmask)
tl.store(out_ptr3 + x2, tmp35, xmask)
tl.store(out_ptr4 + x2, tmp38, xmask)
tl.store(out_ptr5 + x2, tmp40, xmask)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4,
out_ptr5, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (4 + x0 + 16 * x1), xmask)
tmp12 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask)
tmp17 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (12 + x0 + 16 * x1), xmask)
tmp20 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (8 + x0 + 16 * x1), xmask)
tmp25 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (8 + x0 + 16 * x1), xmask)
tmp28 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = tl.sigmoid(tmp22)
tmp26 = tmp24 + tmp25
tmp29 = tmp27 + tmp28
tmp30 = tmp26 + tmp29
tmp31 = libdevice.tanh(tmp30)
tmp33 = tmp15 * tmp32
tmp34 = tmp7 * tmp31
tmp35 = tmp33 + tmp34
tmp36 = libdevice.tanh(tmp35)
tmp37 = tmp23 * tmp36
tl.store(out_ptr0 + x2, tmp7, xmask)
tl.store(out_ptr1 + x2, tmp15, xmask)
tl.store(out_ptr2 + x2, tmp23, xmask)
tl.store(out_ptr3 + x2, tmp31, xmask)
tl.store(out_ptr4 + x2, tmp35, xmask)
tl.store(out_ptr5 + x2, tmp37, xmask)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_3(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, out_ptr0, out_ptr1, out_ptr2, out_ptr3, xnumel,
XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (4 + x0 + 16 * x1), xmask)
tmp12 = tl.load(in_ptr3 + (4 + x0), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (8 + x0 + 16 * x1), xmask)
tmp17 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (8 + x0 + 16 * x1), xmask)
tmp20 = tl.load(in_ptr3 + (8 + x0), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp10 = tmp8 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = tl.sigmoid(tmp14)
tmp18 = tmp16 + tmp17
tmp21 = tmp19 + tmp20
tmp22 = tmp18 + tmp21
tmp23 = libdevice.tanh(tmp22)
tmp25 = tmp15 * tmp24
tmp26 = tmp7 * tmp23
tmp27 = tmp25 + tmp26
tmp28 = libdevice.tanh(tmp27)
tl.store(out_ptr0 + x2, tmp7, xmask)
tl.store(out_ptr1 + x2, tmp15, xmask)
tl.store(out_ptr2 + x2, tmp23, xmask)
tl.store(out_ptr3 + x2, tmp28, xmask)
@triton.jit
def triton_poi_fused_sigmoid_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (12 + x0 + 16 * x1), xmask)
tmp4 = tl.load(in_ptr3 + (12 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tl.store(out_ptr0 + x2, tmp7, xmask)
@triton.jit
def triton_poi_fused_stack_5(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x0 = xindex % 4
x2 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 4 * x1), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + 4 * (-4 + x1)), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + 4 * (-8 + x1)), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tl.full([1], 16, tl.int64)
tmp19 = tl.load(in_ptr3 + (x0 + 4 * (-12 + x1)), tmp16 & xmask, other=0.0)
tmp20 = tl.load(in_ptr4 + (x0 + 4 * (-12 + x1)), tmp16 & xmask, other=0.0)
tmp21 = tmp19 * tmp20
tmp22 = tl.full(tmp21.shape, 0.0, tmp21.dtype)
tmp23 = tl.where(tmp16, tmp21, tmp22)
tmp24 = tl.where(tmp14, tmp15, tmp23)
tmp25 = tl.where(tmp9, tmp10, tmp24)
tmp26 = tl.where(tmp4, tmp5, tmp25)
tl.store(out_ptr0 + x2, tmp26, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (16, 4), (4, 1))
assert_size_stride(primals_3, (16,), (1,))
assert_size_stride(primals_4, (16, 4), (4, 1))
assert_size_stride(primals_5, (16,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((1, 1, 4, 4), (16, 1, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_zeros_0[grid(16)](buf0, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf1 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf1)
buf2 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (4, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 16), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf32 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_tanh_1[grid(16)](buf1
, primals_3, buf2, primals_5, buf0, buf3, buf5, buf4, buf6,
buf32, buf7, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf8 = buf2
del buf2
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 16),
reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf8)
buf9 = buf1
del buf1
extern_kernels.mm(buf7, reinterpret_tensor(primals_4, (4, 16), (1,
4), 0), out=buf9)
buf10 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf11 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf13 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf14 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf15 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_add_mul_sigmoid_tanh_2[grid(16)](buf8, primals_3,
buf9, primals_5, buf6, buf10, buf11, buf13, buf12, buf14, buf15,
16, XBLOCK=16, num_warps=1, num_stages=1)
buf16 = buf9
del buf9
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 32),
reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf16)
buf17 = buf8
del buf8
extern_kernels.mm(buf15, reinterpret_tensor(primals_4, (4, 16), (1,
4), 0), out=buf17)
buf18 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf19 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf21 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf20 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf22 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf23 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_add_mul_sigmoid_tanh_2[grid(16)](buf16, primals_3,
buf17, primals_5, buf14, buf18, buf19, buf21, buf20, buf22,
buf23, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf24 = buf17
del buf17
extern_kernels.mm(reinterpret_tensor(primals_1, (4, 4), (4, 1), 48),
reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), out=buf24)
del primals_2
buf25 = buf16
del buf16
extern_kernels.mm(buf23, reinterpret_tensor(primals_4, (4, 16), (1,
4), 0), out=buf25)
buf26 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf27 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf28 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf30 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_add_mul_sigmoid_tanh_3[grid(16)](buf24, primals_3,
buf25, primals_5, buf22, buf26, buf27, buf28, buf30, 16, XBLOCK
=16, num_warps=1, num_stages=1)
buf29 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_sigmoid_4[grid(16)](buf24, primals_3, buf25,
primals_5, buf29, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf24
del primals_3
del primals_5
buf31 = reinterpret_tensor(buf25, (16, 4), (4, 1), 0)
del buf25
triton_poi_fused_stack_5[grid(64)](buf7, buf15, buf23, buf29, buf30,
buf31, 64, XBLOCK=64, num_warps=1, num_stages=1)
return (reinterpret_tensor(buf31, (4, 4, 4), (16, 4, 1), 0),
reinterpret_tensor(buf0, (4, 4), (4, 1), 0), reinterpret_tensor(
primals_1, (4, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4),
(4, 1), 16), reinterpret_tensor(primals_1, (4, 4), (4, 1), 32),
reinterpret_tensor(primals_1, (4, 4), (4, 1), 48), buf3, buf4, buf5,
buf6, buf7, buf10, buf11, buf12, buf13, buf14, buf15, buf18, buf19,
buf20, buf21, buf22, buf23, buf26, buf27, buf28, buf29, buf30,
primals_4, buf32)
def filter_out_old_keys(self, state_dict, prefix, local_metadata):
new_state_dict = {param_name: param_value for param_name, param_value in
state_dict.items() if param_name not in self.old_to_new}
return new_state_dict
class LSTMLinear(nn.Linear):
"""
This function is the same as a nn.Linear layer, except that in the backward pass
the grad_samples get accumulated (instead of being concatenated as in the standard
nn.Linear)
"""
def __init__(self, in_features: 'int', out_features: 'int', bias:
'bool'=True):
super().__init__(in_features, out_features, bias)
class DPLSTMCell(nn.Module):
"""
Internal-only class. Implements *one* step of LSTM so that a LSTM layer can be seen as repeated
applications of this class.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.ih = LSTMLinear(input_size, 4 * hidden_size, bias=self.bias)
self.hh = LSTMLinear(hidden_size, 4 * hidden_size, bias=self.bias)
self.reset_parameters()
def reset_parameters(self):
"""
Resets parameters by initializing them from an uniform distribution.
"""
stdv = 1.0 / math.sqrt(self.hidden_size)
for weight in self.parameters():
nn.init.uniform_(weight, -stdv, stdv)
def forward(self, x: 'torch.Tensor', h_prev: 'torch.Tensor', c_prev:
'torch.Tensor') ->Tuple[torch.Tensor, torch.Tensor]:
gates = self.ih(x) + self.hh(h_prev)
i_t_input, f_t_input, g_t_input, o_t_input = torch.split(gates,
self.hidden_size, 1)
i_t = torch.sigmoid(i_t_input)
f_t = torch.sigmoid(f_t_input)
g_t = torch.tanh(g_t_input)
o_t = torch.sigmoid(o_t_input)
c_t = f_t * c_prev + i_t * g_t
h_t = o_t * torch.tanh(c_t)
return h_t, c_t
class DPLSTMLayer(nn.Module):
"""
Implements *one* layer of LSTM in a way amenable to differential privacy.
We don't expect you to use this directly: use DPLSTM instead :)
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool',
dropout: 'float', reverse: 'bool'=False):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.dropout = dropout
self.reverse = reverse
self.cell = DPLSTMCell(input_size=input_size, hidden_size=
hidden_size, bias=bias)
self.dropout_layer = nn.Dropout(dropout) if dropout > 0 else None
def forward(self, x: 'torch.Tensor', state_init:
'Tuple[torch.Tensor, torch.Tensor]') ->Tuple[torch.Tensor, Tuple[
torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTMLayer when a sequence is given in input.
Args:
x: Input sequence to the DPLSTMCell of shape ``[T, B, D]``.
state_init: Initial state of the LSTMCell as a tuple ``(h_0, c_0)``
where ``h_0`` is the initial hidden state and ``c_0`` is the
initial cell state of the DPLSTMCell
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTMCell for each timestep ``t``.
- ``h_n`` is of shape ``[B, H]`` and is a tensor containing the hidden state for ``t = T``.
- ``c_n`` is of shape ``[B, H]`` tensor containing the cell state for ``t = T``.
"""
seq_length, _batch_sz, _ = x.shape
if self.reverse:
x = x.flip(0)
x = torch.unbind(x, dim=0)
h_0, c_0 = state_init
h_n = [h_0]
c_n = [c_0]
for t in range(seq_length):
h_next, c_next = self.cell(x[t], h_n[t], c_n[t])
if self.dropout:
h_next = self.dropout_layer(h_next)
h_n.append(h_next)
c_n.append(c_next)
h_n = torch.stack(h_n[1:], dim=0)
return h_n.flip(0) if self.reverse else h_n, (h_n[-1], c_n[-1])
class BidirectionalDPLSTMLayer(nn.Module):
"""
Implements *one* layer of Bidirectional LSTM in a way amenable to differential privacy.
We don't expect you to use this directly: use DPLSTM instead :)
"""
def __init__(self, input_size: 'int', hidden_size: 'int', bias: 'bool',
dropout: 'float'):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.bias = bias
self.dropout = dropout
self.forward_layer = DPLSTMLayer(input_size=input_size, hidden_size
=hidden_size, bias=bias, dropout=dropout, reverse=False)
self.reverse_layer = DPLSTMLayer(input_size=input_size, hidden_size
=hidden_size, bias=bias, dropout=dropout, reverse=True)
def forward(self, x: 'torch.Tensor', state_init:
'Tuple[torch.Tensor, torch.Tensor]') ->Tuple[torch.Tensor, Tuple[
torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTM when a sequence is input.
Dimensions as follows:
- B: Batch size
- T: Sequence length
- D: LSTM input hidden size (eg from a word embedding)
- H: LSTM output hidden size
- P: number of directions (2 if bidirectional, else 1)
Args:
x: Input sequence to the DPLSTM of shape ``[T, B, D]``
state_init: Initial state of the LSTM as a tuple ``(h_0, c_0)``, where:
- h_0 of shape ``[P, B, H] contains the initial hidden state
- c_0 of shape ``[P, B, H] contains the initial cell state
This argument can be (and defaults to) None, in which case zero tensors will be used.
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H * P]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTM for each timestep ``t``.
- ``h_n`` is of shape ``[P, B, H]`` and contains the hidden state for ``t = T``.
- ``c_n`` is of shape ``[P, B, H]`` and contains the cell state for ``t = T``.
"""
h0, c0 = state_init
h0_f, h0_r = h0.unbind(0)
c0_f, c0_r = c0.unbind(0)
out_f, (h_f, c_f) = self.forward_layer(x, (h0_f, c0_f))
out_r, (h_r, c_r) = self.reverse_layer(x, (h0_r, c0_r))
out = torch.cat([out_f, out_r], dim=-1)
h = torch.stack([h_f, h_r], dim=0)
c = torch.stack([c_f, c_r], dim=0)
return out, (h, c)
class ParamRenamedModule(nn.Module):
"""
This class defines a nn.Module whose parameters are renamed. This is useful when you want to
reimplement a layer but make sure its state_dict and list of parameters are exactly the same
as another reference layer so that you can have a drop-in replacement that does not depend on
how your layer is actually implemented. In Opacus, this is used for DPLSTM, where our
implementation leverages submodules and requires alignment to the state_dict of nn.LSTM.
"""
def __init__(self, rename_map: 'Dict[str, str]'):
"""
Initializes internal state. Subclass this instead of ``torch.nn.Module`` whenever you need
to rename your model's state.
Args:
rename_map: mapping from old name -> new name for each parameter you want renamed.
Note that this must be a 1:1 mapping!
"""
super().__init__()
self.old_to_new = rename_map
self.new_to_old = {v: k for k, v in rename_map.items()}
self._register_state_dict_hook(filter_out_old_keys)
def _register_renamed_parameters(self):
"""
Internal function. This function simply registers parameters under their new name. They will
automatically mask their duplicates coming from submodules. This trick works because
self.parameters() proceeds recursively from the top, going into submodules after processing
items at the current level, and will not return duplicates.
"""
for old_name, param in super().named_parameters():
if old_name in self.old_to_new:
new_name = self.old_to_new[old_name]
self.register_parameter(new_name, param)
def __setattr__(self, name: 'str', value: 'Union[Tensor, nn.Module]'
) ->None:
"""
Whenever you set an attribute, eg `self.linear`, this is called to actually register it in
any nn.Module. We rely on the masking trick explained in the docs for
``_register_renamed_parameters`` to make sure we replace things only once. If a new parameter
in the rename list is detected, we rename and mask it so next time this is called we will
no longer find it.
"""
super().__setattr__(name, value)
try:
self._register_renamed_parameters()
except ModuleAttributeError:
pass
def load_state_dict(self, state_dict: 'Dict[str, Tensor]', strict:
'bool'=True):
"""
Identical to ``torch.nn.Module.load_state_dict()`` but handles the renamed keys.
"""
missing_keys, unexpected_keys = super().load_state_dict(state_dict,
strict=False)
missing_keys = [k for k in missing_keys if k not in self.old_to_new]
if strict:
error_msgs = []
if len(unexpected_keys) > 0:
error_msgs.insert(0,
'Unexpected key(s) in state_dict: {}. '.format(', '.
join('"{}"'.format(k) for k in unexpected_keys)))
if len(missing_keys) > 0:
error_msgs.insert(0, 'Missing key(s) in state_dict: {}. '.
format(', '.join('"{}"'.format(k) for k in missing_keys)))
if len(error_msgs) > 0:
raise RuntimeError(
'Error(s) in loading state_dict for {}:\n\t{}'.format(
self.__class__.__name__, '\n\t'.join(error_msgs)))
return _IncompatibleKeys(missing_keys, unexpected_keys)
class DPLSTM(ParamRenamedModule):
"""
DP-friendly drop-in replacement of the ``torch.nn.LSTM`` module.
Its state_dict matches that of nn.LSTM exactly, so that after training it can be exported
and loaded by an nn.LSTM for inference.
Refer to nn.LSTM's documentation for all parameters and inputs.
"""
def __init__(self, input_size: 'int', hidden_size: 'int', num_layers:
'int'=1, bias: 'bool'=True, batch_first: 'bool'=False, dropout:
'float'=0, bidirectional: 'bool'=False):
rename_dict = self._make_rename_dict(num_layers, bias, bidirectional)
super().__init__(rename_dict)
self.input_size = input_size
self.hidden_size = hidden_size
self.num_layers = num_layers
self.bias = bias
self.batch_first = batch_first
self.dropout = dropout
self.bidirectional = bidirectional
self.num_directions = 2 if self.bidirectional else 1
LayerClass = BidirectionalDPLSTMLayer if bidirectional else DPLSTMLayer
self.layers = nn.ModuleList([LayerClass(input_size=self.input_size if
i == 0 else self.hidden_size * self.num_directions, hidden_size
=self.hidden_size, bias=self.bias, dropout=self.dropout if i <
self.num_layers - 1 else 0) for i in range(num_layers)])
def forward(self, x: 'torch.Tensor', state_init:
'Optional[Tuple[torch.Tensor, torch.Tensor]]'=None) ->Tuple[torch.
Tensor, Tuple[torch.Tensor, torch.Tensor]]:
"""
Implements the forward pass of the DPLSTM when a sequence is input.
Dimensions as follows:
- B: Batch size
- T: Sequence length
- D: LSTM input hidden size (eg from a word embedding)
- H: LSTM output hidden size
- L: number of layers in the LSTM
- P: number of directions (2 if bidirectional, else 1)
Args:
x: Input sequence to the DPLSTM of shape ``[T, B, D]``
state_init: Initial state of the LSTM as a tuple ``(h_0, c_0)``, where:
- h_0 of shape ``[L*P, B, H] contains the initial hidden state
- c_0 of shape ``[L*P, B, H] contains the initial cell state
This argument can be (and defaults to) None, in which case zero tensors will be used.
Returns:
``output, (h_n, c_n)`` where:
- ``output`` is of shape ``[T, B, H * P]`` and is a tensor containing the output
features (``h_t``) from the last layer of the DPLSTM for each timestep ``t``.
- ``h_n`` is of shape ``[L * P, B, H]`` and contains the hidden state for ``t = T``.
- ``c_n`` is of shape ``[L * P, B, H]`` and contains the cell state for ``t = T``.
"""
x = self._rearrange_batch_dim(x)
_T, B, _D = x.shape
L = self.num_layers
P = 2 if self.bidirectional else 1
H = self.hidden_size
h_0s, c_0s = state_init or (None, None)
if h_0s is None:
h_0s = torch.zeros(L, P, B, self.hidden_size, dtype=x[0].dtype,
device=x[0].device)
else:
h_0s = h_0s.reshape([L, P, B, H])
if c_0s is None:
c_0s = torch.zeros(L, P, B, self.hidden_size, dtype=x[0].dtype,
device=x[0].device)
else:
c_0s = c_0s.reshape([L, P, B, H])
hs: 'List[torch.Tensor]' = []
cs: 'List[torch.Tensor]' = []
for layer, h0, c0 in zip(self.layers, h_0s, c_0s):
if not self.bidirectional:
h0 = h0.squeeze()
c0 = c0.squeeze()
x, (h, c) = layer(x, (h0, c0))
if not self.bidirectional:
h = h.unsqueeze(0)
c = c.unsqueeze(0)
hs.append(h)
cs.append(c)
hs = torch.cat(hs, dim=0)
cs = torch.cat(cs, dim=0)
out = self._rearrange_batch_dim(x)
return out, (hs, cs)
def _rearrange_batch_dim(self, x: 'torch.Tensor') ->torch.Tensor:
if self.batch_first:
x = x.transpose(0, 1)
return x
def __repr__(self):
s = f'DPLSTM({self.input_size}, {self.hidden_size}, bias={self.bias}'
if self.batch_first:
s += f', batch_first={self.batch_first}'
if self.num_layers > 1:
s += f', num_layers={self.num_layers}'
if self.dropout:
s += f', dropout={self.dropout}'
if self.bidirectional:
s += f', bidirectional={self.bidirectional}'
return s
def _make_rename_dict(self, num_layers, bias, bidirectional):
"""
Programmatically constructs a dictionary old_name -> new_name to align with the param
names used in ``torch.nn.LSTM``.
"""
d = {}
components = ['weight'] + ['bias' if bias else []]
matrices = ['ih', 'hh']
for i in range(num_layers):
for c in components:
for m in matrices:
nn_name = f'{c}_{m}_l{i}'
if bidirectional:
d[f'layers.{i}.forward_layer.cell.{m}.{c}'] = nn_name
d[f'layers.{i}.reverse_layer.cell.{m}.{c}'
] = nn_name + '_reverse'
else:
d[f'layers.{i}.cell.{m}.{c}'] = nn_name
return d
class DPSLTMAdapterNew(nn.Module):
"""
Adapter for DPLSTM.
LSTM returns a tuple, but our testing tools need the model to return a single tensor in output.
We do this adaption here.
"""
def __init__(self, *args, **kwargs):
super().__init__()
self.dplstm = DPLSTM(*args, **kwargs)
def forward(self, input_0):
primals_2 = self.dplstm.weight_ih_l0
primals_3 = self.dplstm.bias_ih_l0
primals_4 = self.dplstm.weight_hh_l0
primals_5 = self.dplstm.bias_hh_l0
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
madhavajay/opacus
|
DPSLTMAdapter
| false | 10,494 |
[
"Apache-2.0"
] | 0 |
7ae098764b4cf2388c66e263dd8d56bca0a290d0
|
https://github.com/madhavajay/opacus/tree/7ae098764b4cf2388c66e263dd8d56bca0a290d0
|
CombineSlices
|
# AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/xj/cxjyiylkxhhw7puycpyh4rhisqin43l5kgucj4lyyniez2ypunvk.py
# Topologically Sorted Source Nodes: [index_select], Original ATen: [aten.index_select]
# Source node to ATen node mapping:
# index_select => index
# Graph fragment:
# %index : [num_users=1] = call_function[target=torch.ops.aten.index.Tensor](args = (%arg0_1, [None, None, %full_default]), kwargs = {})
triton_poi_fused_index_select_0 = async_compile.triton('triton_poi_fused_index_select_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_index_select_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_index_select_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 4), (16, 4, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [index_select], Original ATen: [aten.index_select]
stream0 = get_raw_stream(0)
triton_poi_fused_index_select_0.run(arg0_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
from torch import nn
import torch.utils.data
import torch.utils.data.distributed
import torch.optim
class CombineSlices(nn.Module):
def __init__(self, slice_dim=2):
super().__init__()
self.slice_dim = slice_dim
def forward(self, x):
return torch.index_select(x, dim=self.slice_dim, index=torch.tensor
(0, device=x.device))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch import nn
import torch.utils.data
import torch.utils.data.distributed
import torch.optim
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_index_select_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 4), (16, 4, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_index_select_0[grid(64)](arg0_1, buf0, 64, XBLOCK=
64, num_warps=1, num_stages=1)
del arg0_1
return buf0,
class CombineSlicesNew(nn.Module):
def __init__(self, slice_dim=2):
super().__init__()
self.slice_dim = slice_dim
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
|
kapoor1992/fastMRI
|
CombineSlices
| false | 10,495 |
[
"MIT"
] | 0 |
6b0af94663faa55a2dd901a6a5cbb7d7b5f4cf6d
|
https://github.com/kapoor1992/fastMRI/tree/6b0af94663faa55a2dd901a6a5cbb7d7b5f4cf6d
|
Discriminator
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/4s/c4szhh35rbtilv4zlnanegnq2hofrkvv7yac3nsynw6qjxjbg3tg.py
# Topologically Sorted Source Nodes: [bilinear], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# bilinear => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4) % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_8/inductor_cache/3a/c3a7jkhbifqenis2g23pmqmidmnwe5he2pwgbrunksukoz44fhmm.py
# Topologically Sorted Source Nodes: [logits], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# logits => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%squeeze, %squeeze_1], 1), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp6 = tl.load(in_ptr1 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp8 = tmp5 + tmp7
tmp9 = tl.full(tmp8.shape, 0.0, tmp8.dtype)
tmp10 = tl.where(tmp4, tmp8, tmp9)
tmp11 = tmp0 >= tmp3
tmp12 = tl.full([1], 8, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tl.load(in_ptr2 + ((4*x1) + ((-4) + x0)), tmp11 & xmask, eviction_policy='evict_last', other=0.0)
tmp15 = tmp14 + tmp7
tmp16 = tl.full(tmp15.shape, 0.0, tmp15.dtype)
tmp17 = tl.where(tmp11, tmp15, tmp16)
tmp18 = tl.where(tmp4, tmp10, tmp17)
tl.store(out_ptr0 + (x2), tmp18, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, ), (1, ))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (1, ), (1, ))
assert_size_stride(primals_5, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [bilinear], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_1, buf0, 64, grid=grid(64), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [bilinear], Original ATen: [aten._trilinear]
buf1 = torch.ops.aten._trilinear.default(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), primals_3, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), [1, 3], [0], [1, 2], [2, 3])
buf2 = buf1
del buf1
# Topologically Sorted Source Nodes: [bilinear_1], Original ATen: [aten._trilinear]
buf3 = torch.ops.aten._trilinear.default(reinterpret_tensor(primals_5, (16, 4), (4, 1), 0), primals_3, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), [1, 3], [0], [1, 2], [2, 3])
del primals_3
buf4 = buf3
del buf3
buf5 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [logits], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf2, primals_4, buf4, buf5, 32, grid=grid(32), stream=stream0)
del buf2
del buf4
del primals_4
return (buf5, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (16, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import torch
import torch.nn as nn
class Discriminator(nn.Module):
def __init__(self, n_h):
super(Discriminator, self).__init__()
self.f_k = nn.Bilinear(n_h, n_h, 1)
for m in self.modules():
self.weights_init(m)
def weights_init(self, m):
if isinstance(m, nn.Bilinear):
torch.nn.init.xavier_uniform_(m.weight.data)
if m.bias is not None:
m.bias.data.fill_(0.0)
def forward(self, c, h_pl, h_mi, s_bias1=None, s_bias2=None):
c_x = torch.unsqueeze(c, 1)
c_x = c_x.expand_as(h_pl)
sc_1 = torch.squeeze(self.f_k(h_pl, c_x), 2)
sc_2 = torch.squeeze(self.f_k(h_mi, c_x), 2)
if s_bias1 is not None:
sc_1 += s_bias1
if s_bias2 is not None:
sc_2 += s_bias2
logits = torch.cat((sc_1, sc_2), 1)
return logits
def get_inputs():
return [torch.rand([4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'n_h': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4 % 4
x3 = xindex
tmp0 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp6 = tl.load(in_ptr1 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp8 = tmp5 + tmp7
tmp9 = tl.full(tmp8.shape, 0.0, tmp8.dtype)
tmp10 = tl.where(tmp4, tmp8, tmp9)
tmp11 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp14 = tl.load(in_ptr2 + (4 * x1 + (-4 + x0)), tmp11 & xmask,
eviction_policy='evict_last', other=0.0)
tmp15 = tmp14 + tmp7
tmp16 = tl.full(tmp15.shape, 0.0, tmp15.dtype)
tmp17 = tl.where(tmp11, tmp15, tmp16)
tmp18 = tl.where(tmp4, tmp10, tmp17)
tl.store(out_ptr0 + x2, tmp18, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4,), (1,))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (1, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (1,), (1,))
assert_size_stride(primals_5, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(64)](primals_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_1
buf1 = torch.ops.aten._trilinear.default(reinterpret_tensor(
primals_2, (16, 4), (4, 1), 0), primals_3, reinterpret_tensor(
buf0, (16, 4), (4, 1), 0), [1, 3], [0], [1, 2], [2, 3])
buf2 = buf1
del buf1
buf3 = torch.ops.aten._trilinear.default(reinterpret_tensor(
primals_5, (16, 4), (4, 1), 0), primals_3, reinterpret_tensor(
buf0, (16, 4), (4, 1), 0), [1, 3], [0], [1, 2], [2, 3])
del primals_3
buf4 = buf3
del buf3
buf5 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
triton_poi_fused_cat_1[grid(32)](buf2, primals_4, buf4, buf5, 32,
XBLOCK=32, num_warps=1, num_stages=1)
del buf2
del buf4
del primals_4
return buf5, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0
), reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(
primals_5, (16, 4), (4, 1), 0)
class DiscriminatorNew(nn.Module):
def __init__(self, n_h):
super(DiscriminatorNew, self).__init__()
self.f_k = nn.Bilinear(n_h, n_h, 1)
for m in self.modules():
self.weights_init(m)
def weights_init(self, m):
if isinstance(m, nn.Bilinear):
torch.nn.init.xavier_uniform_(m.weight.data)
if m.bias is not None:
m.bias.data.fill_(0.0)
def forward(self, input_0, input_1, input_2):
primals_3 = self.f_k.weight
primals_4 = self.f_k.bias
primals_1 = input_0
primals_2 = input_1
primals_5 = input_2
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
|
mess-clarifai/DGI
|
Discriminator
| false | 10,496 |
[
"MIT"
] | 0 |
3a7c96d59991d448b84d709916d1d5f256e5b9be
|
https://github.com/mess-clarifai/DGI/tree/3a7c96d59991d448b84d709916d1d5f256e5b9be
|
SphericalBesselBasis
|
# AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_8/inductor_cache/mh/cmhudqgyaffvfgzwidcjby2k4225p53z7fnky2nqvz37ykelzacp.py
# Topologically Sorted Source Nodes: [truediv, mul, sin, mul_1], Original ATen: [aten.reciprocal, aten.mul, aten.sin]
# Source node to ATen node mapping:
# mul => mul_1
# mul_1 => mul_2
# sin => sin
# truediv => mul, reciprocal
# Graph fragment:
# %reciprocal : [num_users=1] = call_function[target=torch.ops.aten.reciprocal.default](args = (%unsqueeze,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%reciprocal, 0.1767766952966369), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %unsqueeze), kwargs = {})
# %sin : [num_users=1] = call_function[target=torch.ops.aten.sin.default](args = (%mul_1,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %sin), kwargs = {})
triton_poi_fused_mul_reciprocal_sin_0 = async_compile.triton('triton_poi_fused_mul_reciprocal_sin_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_reciprocal_sin_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_reciprocal_sin_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp5 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 1, tl.int32)
tmp2 = tmp1 / tmp0
tmp3 = 0.1767766952966369
tmp4 = tmp2 * tmp3
tmp6 = tmp5 * tmp0
tmp7 = tl_math.sin(tmp6)
tmp8 = tmp4 * tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 4, 4, 4), (64, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [truediv, mul, sin, mul_1], Original ATen: [aten.reciprocal, aten.mul, aten.sin]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_reciprocal_sin_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
return (buf0, primals_1, primals_2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
|
import math
import torch
import numpy as np
class SphericalBesselBasis(torch.nn.Module):
"""
1D spherical Bessel basis
Parameters
----------
num_radial: int
Controls maximum frequency.
cutoff: float
Cutoff distance in Angstrom.
"""
def __init__(self, num_radial: 'int', cutoff: 'float'):
super().__init__()
self.norm_const = math.sqrt(2 / cutoff ** 3)
self.frequencies = torch.nn.Parameter(data=torch.tensor(np.pi * np.
arange(1, num_radial + 1, dtype=np.float32)), requires_grad=True)
def forward(self, d_scaled):
return self.norm_const / d_scaled[:, None] * torch.sin(self.
frequencies * d_scaled[:, None])
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_radial': 4, 'cutoff': 4}]
|
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_reciprocal_sin_0(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp5 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 1, tl.int32)
tmp2 = tmp1 / tmp0
tmp3 = 0.1767766952966369
tmp4 = tmp2 * tmp3
tmp6 = tmp5 * tmp0
tmp7 = tl_math.sin(tmp6)
tmp8 = tmp4 * tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 4, 4, 4), (64, 64, 16, 4, 1),
torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_reciprocal_sin_0[grid(256)](primals_1,
primals_2, buf0, 256, XBLOCK=128, num_warps=4, num_stages=1)
return buf0, primals_1, primals_2
class SphericalBesselBasisNew(torch.nn.Module):
"""
1D spherical Bessel basis
Parameters
----------
num_radial: int
Controls maximum frequency.
cutoff: float
Cutoff distance in Angstrom.
"""
def __init__(self, num_radial: 'int', cutoff: 'float'):
super().__init__()
self.norm_const = math.sqrt(2 / cutoff ** 3)
self.frequencies = torch.nn.Parameter(data=torch.tensor(np.pi * np.
arange(1, num_radial + 1, dtype=np.float32)), requires_grad=True)
def forward(self, input_0):
primals_2 = self.frequencies
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
|
krylea/ocp
|
SphericalBesselBasis
| false | 10,497 |
[
"MIT"
] | 0 |
00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
https://github.com/krylea/ocp/tree/00fc1df29731d70ff1b5cf8e9323d1d2f1f8e540
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.