text
stringlengths 96
319k
| id
stringlengths 14
178
| metadata
dict |
|---|---|---|
# Testing mixed int8 quantization
The following is the recipe on how to effectively debug `bitsandbytes` integration on Hugging Face `transformers`.
## Library requirements
+ `transformers>=4.22.0`
+ `accelerate>=0.12.0`
+ `bitsandbytes>=0.31.5`.
## Hardware requirements
The following instructions are tested with 2 NVIDIA-Tesla T4 GPUs. To run successfully `bitsandbytes` you would need a 8-bit core tensor supported GPU. Note that Turing, Ampere or newer architectures - e.g. T4, RTX20s RTX30s, A40-A100, A6000 should be supported.
## Virutal envs
```bash
conda create --name int8-testing python==3.8
pip install bitsandbytes>=0.31.5
pip install accelerate>=0.12.0
pip install transformers>=4.23.0
```
if `transformers>=4.23.0` is not released yet, then use:
```bash
pip install git+https://github.com/huggingface/transformers.git
```
## Troubleshooting
A list of common errors:
### Torch does not correctly do the operations on GPU
First check that:
```py
import torch
vec = torch.randn(1, 2, 3).to(0)
```
Works without any error. If not, install torch using `conda` like:
```bash
conda create --name int8-testing python==3.8
conda install pytorch torchvision torchaudio cudatoolkit=11.6 -c pytorch -c conda-forge
pip install bitsandbytes>=0.31.5
pip install accelerate>=0.12.0
pip install transformers>=4.23.0
```
For the latest pytorch instructions please see [this](https://pytorch.org/get-started/locally/)
and the snippet above should work.
### ` bitsandbytes operations are not supported under CPU!`
This happens when some Linear weights are set to the CPU when using `accelerate`. Please check carefully `model.hf_device_map` and make sure that there is no `Linear` module that is assigned to CPU. It is fine to have the last module (usually the Lm_head) set on CPU.
### `To use the type as a Parameter, please correct the detach() semantics defined by __torch_dispatch__() implementation.`
Use the latest version of `accelerate` with a command such as: `pip install -U accelerate` and the problem should be solved.
### `Parameter has no attribue .CB`
Same solution as above.
### `RuntimeError: CUDA error: an illegal memory access was encountered ... consider passing CUDA_LAUNCH_BLOCKING=1`
Run your script by pre-pending `CUDA_LAUNCH_BLOCKING=1` and you should observe an error as described in the next section.
### `CUDA illegal memory error: an illegal memory access at line...`:
Check the CUDA verisons with:
```bash
nvcc --version
```
and confirm it is the same version as the one detected by `bitsandbytes`. If not, run:
```bash
ls -l $CONDA_PREFIX/lib/libcudart.so
```
or
```bash
ls -l $LD_LIBRARY_PATH
```
Check if `libcudart.so` has a correct symlink that is set. Sometimes `nvcc` detects the correct CUDA version but `bitsandbytes` doesn't. You have to make sure that the symlink that is set for the file `libcudart.so` is redirected to the correct CUDA file.
Here is an example of a badly configured CUDA installation:
`nvcc --version` gives:
which means that the detected CUDA version is 11.3 but `bitsandbytes` outputs:
First check:
```bash
echo $LD_LIBRARY_PATH
```
If this contains multiple paths separated by `:`. Then you have to make sure that the correct CUDA version is set. By doing:
```bash
ls -l $path/libcudart.so
```
On each path (`$path`) separated by `:`.
If not, simply run
```bash
ls -l $LD_LIBRARY_PATH/libcudart.so
```
and you can see
If you see that the file is linked to the wrong CUDA version (here 10.2), find the correct location for `libcudart.so` (`find --name libcudart.so`) and replace the environment variable `LD_LIBRARY_PATH` with the one containing the correct `libcudart.so` file. | transformers/tests/quantization/bnb/README.md/0 | {
"file_path": "transformers/tests/quantization/bnb/README.md",
"repo_id": "transformers",
"token_count": 1405
} |
# coding=utf-8
# Copyright 2019 HuggingFace Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import collections
import copy
import gc
import inspect
import math
import os
import os.path
import random
import re
import tempfile
import warnings
from collections import defaultdict
from contextlib import contextmanager
from typing import Dict, List, Tuple
import numpy as np
from packaging import version
from parameterized import parameterized
from pytest import mark
import transformers
from transformers import (
AutoModel,
AutoModelForCausalLM,
AutoModelForSequenceClassification,
PretrainedConfig,
PreTrainedModel,
is_torch_available,
logging,
set_seed,
)
from transformers.integrations import HfDeepSpeedConfig
from transformers.integrations.deepspeed import (
is_deepspeed_available,
is_deepspeed_zero3_enabled,
unset_hf_deepspeed_config,
)
from transformers.models.auto import get_values
from transformers.models.auto.modeling_auto import (
MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES,
MODEL_FOR_BACKBONE_MAPPING_NAMES,
MODEL_FOR_CAUSAL_IMAGE_MODELING_MAPPING_NAMES,
MODEL_FOR_CAUSAL_LM_MAPPING_NAMES,
MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES,
MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_MASKED_IMAGE_MODELING_MAPPING_NAMES,
MODEL_FOR_MASKED_LM_MAPPING_NAMES,
MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES,
MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES,
MODEL_FOR_PRETRAINING_MAPPING_NAMES,
MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES,
MODEL_FOR_SEMANTIC_SEGMENTATION_MAPPING_NAMES,
MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES,
MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_VISION_2_SEQ_MAPPING_NAMES,
MODEL_MAPPING_NAMES,
)
from transformers.testing_utils import (
CaptureLogger,
is_flaky,
is_pt_flax_cross_test,
is_pt_tf_cross_test,
require_accelerate,
require_bitsandbytes,
require_deepspeed,
require_flash_attn,
require_non_xpu,
require_safetensors,
require_torch,
require_torch_accelerator,
require_torch_gpu,
require_torch_multi_accelerator,
require_torch_multi_gpu,
require_torch_sdpa,
set_config_for_less_flaky_test,
set_model_for_less_flaky_test,
set_model_tester_for_less_flaky_test,
slow,
torch_device,
)
from transformers.utils import (
CONFIG_NAME,
GENERATION_CONFIG_NAME,
SAFE_WEIGHTS_NAME,
is_accelerate_available,
is_flax_available,
is_tf_available,
is_torch_bf16_available_on_device,
is_torch_fp16_available_on_device,
is_torch_fx_available,
is_torch_sdpa_available,
)
from transformers.utils.generic import ContextManagers, ModelOutput
if is_accelerate_available():
from accelerate.utils import compute_module_sizes
if is_torch_available():
import torch
import torch.nn.functional as F
from safetensors.torch import load_file as safe_load_file
from safetensors.torch import save_file as safe_save_file
from torch import nn
from transformers import MODEL_MAPPING, AdaptiveEmbedding
from transformers.cache_utils import Cache, DynamicCache
from transformers.modeling_utils import load_state_dict, no_init_weights
from transformers.pytorch_utils import id_tensor_storage
if is_tf_available():
import tensorflow as tf
if is_flax_available():
import jax.numpy as jnp
from tests.utils.test_modeling_flax_utils import check_models_equal
from transformers.modeling_flax_pytorch_utils import (
convert_pytorch_state_dict_to_flax,
load_flax_weights_in_pytorch_model,
)
if is_torch_fx_available():
from transformers.utils.fx import _FX_SUPPORTED_MODELS_WITH_KV_CACHE, symbolic_trace
if is_deepspeed_available():
import deepspeed
def _config_zero_init(config):
configs_no_init = copy.deepcopy(config)
for key in configs_no_init.__dict__.keys():
if "_range" in key or "_std" in key or "initializer_factor" in key or "layer_scale" in key:
setattr(configs_no_init, key, 1e-10)
if isinstance(getattr(configs_no_init, key, None), PretrainedConfig):
no_init_subconfig = _config_zero_init(getattr(configs_no_init, key))
setattr(configs_no_init, key, no_init_subconfig)
return configs_no_init
def _mock_init_weights(self, module):
for name, param in module.named_parameters(recurse=False):
# Use the first letter of the name to get a value and go from a <> -13 to z <> 12
value = ord(name[0].lower()) - 110
param.data.fill_(value)
def _mock_all_init_weights(self):
# Prune heads if needed
if self.config.pruned_heads:
self.prune_heads(self.config.pruned_heads)
import transformers.modeling_utils
if transformers.modeling_utils._init_weights:
for module in self.modules():
module._is_hf_initialized = False
# Initialize weights
self.apply(self._initialize_weights)
# Tie weights should be skipped when not initializing all weights
# since from_pretrained(...) calls tie weights anyways
self.tie_weights()
@contextmanager
def _deepspeed_zero3(ds_config):
dschf = HfDeepSpeedConfig(ds_config)
try:
yield dschf
finally:
unset_hf_deepspeed_config()
def sdpa_kernel(enable_flash, enable_math, enable_mem_efficient):
if version.parse(torch.__version__).release < version.parse("2.3").release:
return torch.backends.cuda.sdp_kernel(
enable_flash=enable_flash, enable_math=enable_math, enable_mem_efficient=enable_mem_efficient
)
backends = []
if enable_flash:
backends += [torch.nn.attention.SDPBackend.FLASH_ATTENTION]
if enable_math:
backends += [torch.nn.attention.SDPBackend.MATH]
if enable_mem_efficient:
backends += [torch.nn.attention.SDPBackend.EFFICIENT_ATTENTION]
return torch.nn.attention.sdpa_kernel(backends)
@require_torch
class ModelTesterMixin:
model_tester = None
all_model_classes = ()
all_generative_model_classes = ()
fx_compatible = False
test_torchscript = True
test_pruning = True
test_resize_embeddings = True
test_resize_position_embeddings = False
test_head_masking = True
test_mismatched_shapes = True
test_missing_keys = True
test_model_parallel = False
# Used in `check_training_gradient_checkpointing` to NOT check all params having gradient (e.g. for some MOE models)
test_all_params_have_gradient = True
is_encoder_decoder = False
has_attentions = True
_is_composite = False
model_split_percents = [0.5, 0.7, 0.9]
def _prepare_for_class(self, inputs_dict, model_class, return_labels=False):
inputs_dict = copy.deepcopy(inputs_dict)
if model_class.__name__ in get_values(MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES):
inputs_dict = {
k: v.unsqueeze(1).expand(-1, self.model_tester.num_choices, -1).contiguous()
if isinstance(v, torch.Tensor) and v.ndim > 1
else v
for k, v in inputs_dict.items()
}
elif model_class.__name__ in get_values(MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES):
inputs_dict.pop("attention_mask")
elif model_class.__name__ == MODEL_FOR_PRETRAINING_MAPPING_NAMES["hiera"]:
config = self.model_tester.get_config()
mask_spatial_shape = [
i // s // ms for i, s, ms in zip(config.image_size, config.patch_stride, config.masked_unit_size)
]
num_windows = math.prod(mask_spatial_shape)
torch.manual_seed(0)
inputs_dict["noise"] = torch.rand(self.model_tester.batch_size, num_windows)
if return_labels:
if model_class.__name__ in get_values(MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES):
inputs_dict["labels"] = torch.ones(self.model_tester.batch_size, dtype=torch.long, device=torch_device)
elif model_class.__name__ in [
*get_values(MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES),
*get_values(MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES),
]:
inputs_dict["start_positions"] = torch.zeros(
self.model_tester.batch_size, dtype=torch.long, device=torch_device
)
inputs_dict["end_positions"] = torch.zeros(
self.model_tester.batch_size, dtype=torch.long, device=torch_device
)
elif model_class.__name__ in [
*get_values(MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES),
*get_values(MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES),
*get_values(MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES),
*get_values(MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES),
*get_values(MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES),
]:
inputs_dict["labels"] = torch.zeros(
self.model_tester.batch_size, dtype=torch.long, device=torch_device
)
elif model_class.__name__ in [
*get_values(MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES),
*get_values(MODEL_FOR_CAUSAL_LM_MAPPING_NAMES),
*get_values(MODEL_FOR_CAUSAL_IMAGE_MODELING_MAPPING_NAMES),
*get_values(MODEL_FOR_MASKED_LM_MAPPING_NAMES),
*get_values(MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES),
*get_values(MODEL_FOR_VISION_2_SEQ_MAPPING_NAMES),
]:
inputs_dict["labels"] = torch.zeros(
(self.model_tester.batch_size, self.model_tester.seq_length), dtype=torch.long, device=torch_device
)
elif model_class.__name__ in get_values(MODEL_FOR_MASKED_IMAGE_MODELING_MAPPING_NAMES):
num_patches = self.model_tester.image_size // self.model_tester.patch_size
inputs_dict["bool_masked_pos"] = torch.zeros(
(self.model_tester.batch_size, num_patches**2), dtype=torch.long, device=torch_device
)
elif model_class.__name__ in get_values(MODEL_FOR_SEMANTIC_SEGMENTATION_MAPPING_NAMES):
batch_size, num_channels, height, width = inputs_dict["pixel_values"].shape
inputs_dict["labels"] = torch.zeros(
[self.model_tester.batch_size, height, width], device=torch_device
).long()
return inputs_dict
def test_save_load(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
def check_save_load(out1, out2):
# make sure we don't have nans
out_2 = out2.cpu().numpy()
out_2[np.isnan(out_2)] = 0
out_2 = out_2[~np.isneginf(out_2)]
out_1 = out1.cpu().numpy()
out_1[np.isnan(out_1)] = 0
out_1 = out_1[~np.isneginf(out_1)]
max_diff = np.amax(np.abs(out_1 - out_2))
self.assertLessEqual(max_diff, 1e-5)
for model_class in self.all_model_classes:
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
first = model(**self._prepare_for_class(inputs_dict, model_class))[0]
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
# the config file (and the generation config file, if it can generate) should be saved
self.assertTrue(os.path.exists(os.path.join(tmpdirname, CONFIG_NAME)))
self.assertEqual(
model.can_generate(), os.path.exists(os.path.join(tmpdirname, GENERATION_CONFIG_NAME))
)
model = model_class.from_pretrained(tmpdirname)
model.to(torch_device)
with torch.no_grad():
second = model(**self._prepare_for_class(inputs_dict, model_class))[0]
# Save and load second time because `from_pretrained` adds a bunch of new config fields
# so we need to make sure those fields can be loaded back after saving
# Simply init as `model(config)` doesn't add those fields
model.save_pretrained(tmpdirname)
model = model_class.from_pretrained(tmpdirname)
if isinstance(first, tuple) and isinstance(second, tuple):
for tensor1, tensor2 in zip(first, second):
check_save_load(tensor1, tensor2)
else:
check_save_load(first, second)
def test_from_pretrained_no_checkpoint(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
state_dict = model.state_dict()
new_model = model_class.from_pretrained(
pretrained_model_name_or_path=None, config=config, state_dict=state_dict
)
for p1, p2 in zip(model.parameters(), new_model.parameters()):
self.assertTrue(torch.equal(p1, p2))
def test_keep_in_fp32_modules(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class._keep_in_fp32_modules is None:
self.skipTest(reason="Model class has no _keep_in_fp32_modules attribute defined")
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.float16)
for name, param in model.named_parameters():
if any(n in model_class._keep_in_fp32_modules for n in name.split(".")):
self.assertTrue(param.dtype == torch.float32)
else:
self.assertTrue(param.dtype == torch.float16, name)
def test_save_load_keys_to_ignore_on_save(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
_keys_to_ignore_on_save = getattr(model, "_keys_to_ignore_on_save", None)
if _keys_to_ignore_on_save is None:
continue
# check the keys are in the original state_dict
for k in _keys_to_ignore_on_save:
self.assertIn(k, model.state_dict().keys(), "\n".join(model.state_dict().keys()))
# check that certain keys didn't get saved with the model
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
output_model_file = os.path.join(tmpdirname, SAFE_WEIGHTS_NAME)
state_dict_saved = safe_load_file(output_model_file)
for k in _keys_to_ignore_on_save:
self.assertNotIn(k, state_dict_saved.keys(), "\n".join(state_dict_saved.keys()))
# Test we can load the state dict in the model, necessary for the checkpointing API in Trainer.
load_result = model.load_state_dict(state_dict_saved, strict=False)
keys_to_ignore = set(model._keys_to_ignore_on_save)
if hasattr(model, "_tied_weights_keys"):
keys_to_ignore.update(set(model._tied_weights_keys))
self.assertTrue(len(load_result.missing_keys) == 0 or set(load_result.missing_keys) == keys_to_ignore)
self.assertTrue(len(load_result.unexpected_keys) == 0)
def test_gradient_checkpointing_backward_compatibility(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if not model_class.supports_gradient_checkpointing:
continue
config.gradient_checkpointing = True
model = model_class(config)
self.assertTrue(model.is_gradient_checkpointing)
def test_gradient_checkpointing_enable_disable(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if not model_class.supports_gradient_checkpointing:
continue
# at init model should have gradient checkpointing disabled
model = model_class(config)
self.assertFalse(model.is_gradient_checkpointing)
# check enable works
model.gradient_checkpointing_enable()
self.assertTrue(model.is_gradient_checkpointing)
# Loop over all modules and check that relevant modules have gradient_checkpointing set to True
for n, m in model.named_modules():
if hasattr(m, "gradient_checkpointing"):
self.assertTrue(
m.gradient_checkpointing, f"Module {n} does not have gradient_checkpointing set to True"
)
# check disable works
model.gradient_checkpointing_disable()
self.assertFalse(model.is_gradient_checkpointing)
# Loop over all modules and check that relevant modules have gradient_checkpointing set to False
for n, m in model.named_modules():
if hasattr(m, "gradient_checkpointing"):
self.assertFalse(
m.gradient_checkpointing, f"Module {n} does not have gradient_checkpointing set to False"
)
def test_peft_gradient_checkpointing_enable_disable(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if not model_class.supports_gradient_checkpointing:
continue
# at init model should have gradient checkpointing disabled
model = model_class(config)
self.assertFalse(model.is_gradient_checkpointing)
# check enable works
model._hf_peft_config_loaded = True
try:
model.gradient_checkpointing_enable()
except NotImplementedError:
continue
self.assertTrue(model.is_gradient_checkpointing)
# Loop over all modules and check that relevant modules have gradient_checkpointing set to True
for n, m in model.named_modules():
if hasattr(m, "gradient_checkpointing"):
self.assertTrue(
m.gradient_checkpointing, f"Module {n} does not have gradient_checkpointing set to True"
)
# check disable works
model.gradient_checkpointing_disable()
self.assertFalse(model.is_gradient_checkpointing)
# Loop over all modules and check that relevant modules have gradient_checkpointing set to False
for n, m in model.named_modules():
if hasattr(m, "gradient_checkpointing"):
self.assertFalse(
m.gradient_checkpointing, f"Module {n} does not have gradient_checkpointing set to False"
)
@is_flaky(description="low likelihood of failure, reason not yet discovered")
def test_save_load_fast_init_from_base(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
if config.__class__ not in MODEL_MAPPING:
self.skipTest(reason=f"{config.__class__.__name__} not in MODEL_MAPPING")
base_class = MODEL_MAPPING[config.__class__]
if isinstance(base_class, tuple):
base_class = base_class[0]
for model_class in self.all_model_classes:
if model_class == base_class:
continue
# make a copy of model class to not break future tests
# from https://stackoverflow.com/questions/9541025/how-to-copy-a-python-class
class CopyClass(model_class):
pass
model_class_copy = CopyClass
# make sure that all keys are expected for test
model_class_copy._keys_to_ignore_on_load_missing = []
# make init deterministic, but make sure that
# non-initialized weights throw errors nevertheless
model_class_copy._init_weights = _mock_init_weights
model_class_copy.init_weights = _mock_all_init_weights
model = base_class(config)
state_dict = model.state_dict()
# this will often delete a single weight of a multi-weight module
# to test an edge case
random_key_to_del = random.choice(list(state_dict.keys()))
del state_dict[random_key_to_del]
# check that certain keys didn't get saved with the model
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
torch.save(state_dict, os.path.join(tmpdirname, "pytorch_model.bin"))
model_fast_init = model_class_copy.from_pretrained(tmpdirname)
model_slow_init = model_class_copy.from_pretrained(tmpdirname, _fast_init=False)
# Before we test anything
for key in model_fast_init.state_dict().keys():
if isinstance(model_slow_init.state_dict()[key], torch.BoolTensor):
max_diff = (model_slow_init.state_dict()[key] ^ model_fast_init.state_dict()[key]).sum().item()
else:
max_diff = (model_slow_init.state_dict()[key] - model_fast_init.state_dict()[key]).sum().item()
self.assertLessEqual(max_diff, 1e-3, msg=f"{key} not identical")
@slow
@require_accelerate
@mark.accelerate_tests
def test_save_load_low_cpu_mem_usage(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
with tempfile.TemporaryDirectory() as saved_model_path:
for model_class in self.all_model_classes:
model_to_save = model_class(config)
model_to_save.save_pretrained(saved_model_path)
self._check_save_load_low_cpu_mem_usage(model_class, saved_model_path)
@slow
@require_accelerate
@mark.accelerate_tests
def test_save_load_low_cpu_mem_usage_checkpoints(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
with tempfile.TemporaryDirectory() as saved_model_path:
for model_class in self.all_model_classes:
model_to_save = model_class(config)
model_to_save.config.save_pretrained(saved_model_path)
torch.save(model_to_save.state_dict(), os.path.join(saved_model_path, "pytorch_model.bin"))
self._check_save_load_low_cpu_mem_usage(model_class, saved_model_path)
@slow
@require_accelerate
@mark.accelerate_tests
def test_save_load_low_cpu_mem_usage_no_safetensors(self):
with tempfile.TemporaryDirectory() as saved_model_path:
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model_to_save = model_class(config)
model_to_save.save_pretrained(saved_model_path, safe_serialization=False)
self._check_save_load_low_cpu_mem_usage(model_class, saved_model_path)
def _check_save_load_low_cpu_mem_usage(self, model_class, saved_model_path):
from accelerate.utils.modeling import named_module_tensors
# Load the low usage and the normal models.
model_low_usage, loading_info = model_class.from_pretrained(
saved_model_path,
low_cpu_mem_usage=True,
output_loading_info=True,
)
model_non_low_usage = model_class.from_pretrained(saved_model_path)
# Check that there were no missing keys.
self.assertEqual(loading_info["missing_keys"], [])
# The low_cpu_mem_usage=True causes the model params to be initialized with device=meta, and then
# subsequently loaded with the correct values and onto the correct device. We check if there are any
# remaining params that were not properly loaded.
for name, tensor in named_module_tensors(model_low_usage, recurse=True):
self.assertNotEqual(
tensor.device,
torch.device("meta"),
"Tensor '" + name + "' has not been properly loaded and has device=meta.",
)
# Check that the parameters are equal.
for p1, p2 in zip(model_low_usage.parameters(), model_non_low_usage.parameters()):
self.assertEqual(p1.data.ne(p2.data).sum(), 0)
# Check that the state dict keys are equal.
self.assertEqual(set(model_low_usage.state_dict().keys()), set(model_non_low_usage.state_dict().keys()))
# Check that the shared tensors are equal.
tensor_ptrs1 = collections.defaultdict(list)
for name, tensor in model_low_usage.state_dict().items():
tensor_ptrs1[id_tensor_storage(tensor)].append(name)
tied_params1 = [names for _, names in tensor_ptrs1.items() if len(names) > 1]
tensor_ptrs2 = collections.defaultdict(list)
for name, tensor in model_non_low_usage.state_dict().items():
tensor_ptrs2[id_tensor_storage(tensor)].append(name)
tied_params2 = [names for _, names in tensor_ptrs2.items() if len(names) > 1]
self.assertEqual(tied_params1, tied_params2)
def test_save_load_fast_init_to_base(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
if config.__class__ not in MODEL_MAPPING:
self.skipTest(reason=f"{config.__class__.__name__} not in MODEL_MAPPING")
base_class = MODEL_MAPPING[config.__class__]
if isinstance(base_class, tuple):
base_class = base_class[0]
for model_class in self.all_model_classes:
if model_class == base_class:
continue
# make a copy of model class to not break future tests
# from https://stackoverflow.com/questions/9541025/how-to-copy-a-python-class
class CopyClass(base_class):
pass
base_class_copy = CopyClass
# make sure that all keys are expected for test
base_class_copy._keys_to_ignore_on_load_missing = []
# make init deterministic, but make sure that
# non-initialized weights throw errors nevertheless
base_class_copy._init_weights = _mock_init_weights
base_class_copy.init_weights = _mock_all_init_weights
model = model_class(config)
state_dict = model.state_dict()
# this will often delete a single weight of a multi-weight module
# to test an edge case
random_key_to_del = random.choice(list(state_dict.keys()))
del state_dict[random_key_to_del]
# check that certain keys didn't get saved with the model
with tempfile.TemporaryDirectory() as tmpdirname:
model.config.save_pretrained(tmpdirname)
torch.save(state_dict, os.path.join(tmpdirname, "pytorch_model.bin"))
model_fast_init = base_class_copy.from_pretrained(tmpdirname)
model_slow_init = base_class_copy.from_pretrained(tmpdirname, _fast_init=False)
for key in model_fast_init.state_dict().keys():
if isinstance(model_slow_init.state_dict()[key], torch.BoolTensor):
max_diff = torch.max(
model_slow_init.state_dict()[key] ^ model_fast_init.state_dict()[key]
).item()
else:
max_diff = torch.max(
torch.abs(model_slow_init.state_dict()[key] - model_fast_init.state_dict()[key])
).item()
self.assertLessEqual(max_diff, 1e-3, msg=f"{key} not identical")
def test_torch_save_load(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
if config.__class__ not in MODEL_MAPPING:
self.skipTest(reason=f"{config.__class__.__name__} not in MODEL_MAPPING")
base_class = MODEL_MAPPING[config.__class__]
if isinstance(base_class, tuple):
base_class = base_class[0]
for model_class in self.all_model_classes:
if model_class == base_class:
continue
# make a copy of model class to not break future tests
# from https://stackoverflow.com/questions/9541025/how-to-copy-a-python-class
class CopyClass(base_class):
pass
base_class_copy = CopyClass
# make sure that all keys are expected for test
base_class_copy._keys_to_ignore_on_load_missing = []
# make init deterministic, but make sure that
# non-initialized weights throw errors nevertheless
base_class_copy._init_weights = _mock_init_weights
base_class_copy.init_weights = _mock_all_init_weights
model = model_class(config)
state_dict = model.state_dict()
def check_equal(loaded):
for key in state_dict.keys():
max_diff = torch.max(
state_dict()[key] ^ loaded[key]
if isinstance(state_dict[key], torch.BoolTensor)
else torch.abs(state_dict[key] - loaded[key])
).item()
self.assertLessEqual(max_diff, 1e-6, msg=f"{key} not identical")
# check that certain keys didn't get saved with the model
with tempfile.TemporaryDirectory() as tmpdirname:
pt_checkpoint_path = os.path.join(tmpdirname, "pytorch_model.bin")
torch.save(state_dict, pt_checkpoint_path, _use_new_zipfile_serialization=True)
check_equal(load_state_dict(pt_checkpoint_path))
torch.save(state_dict, pt_checkpoint_path, _use_new_zipfile_serialization=False)
check_equal(load_state_dict(pt_checkpoint_path))
def test_initialization(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
configs_no_init = _config_zero_init(config)
for model_class in self.all_model_classes:
model = model_class(config=configs_no_init)
for name, param in model.named_parameters():
if param.requires_grad:
self.assertIn(
((param.data.mean() * 1e9).round() / 1e9).item(),
[0.0, 1.0],
msg=f"Parameter {name} of model {model_class} seems not properly initialized",
)
def test_determinism(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
def check_determinism(first, second):
out_1 = first.cpu().numpy()
out_2 = second.cpu().numpy()
out_1 = out_1[~np.isnan(out_1)]
out_2 = out_2[~np.isnan(out_2)]
out_1 = out_1[~np.isneginf(out_1)]
out_2 = out_2[~np.isneginf(out_2)]
max_diff = np.amax(np.abs(out_1 - out_2))
self.assertLessEqual(max_diff, 1e-5)
for model_class in self.all_model_classes:
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
first = model(**self._prepare_for_class(inputs_dict, model_class))[0]
second = model(**self._prepare_for_class(inputs_dict, model_class))[0]
if isinstance(first, tuple) and isinstance(second, tuple):
for tensor1, tensor2 in zip(first, second):
check_determinism(tensor1, tensor2)
else:
check_determinism(first, second)
def test_batching_equivalence(self, atol=1e-5, rtol=1e-5):
"""
Tests that the model supports batching and that the output is the nearly the same for the same input in
different batch sizes.
(Why "nearly the same" not "exactly the same"? Batching uses different matmul shapes, which often leads to
different results: https://github.com/huggingface/transformers/issues/25420#issuecomment-1775317535)
"""
def recursive_check(batched_object, single_row_object, model_name, key):
if isinstance(batched_object, (list, tuple)):
for batched_object_value, single_row_object_value in zip(batched_object, single_row_object):
recursive_check(batched_object_value, single_row_object_value, model_name, key)
elif isinstance(batched_object, dict):
for batched_object_value, single_row_object_value in zip(
batched_object.values(), single_row_object.values()
):
recursive_check(batched_object_value, single_row_object_value, model_name, key)
# do not compare returned loss (0-dim tensor) / codebook ids (int) / caching objects
elif batched_object is None or not isinstance(batched_object, torch.Tensor):
return
elif batched_object.dim() == 0:
return
# do not compare int or bool outputs as they are mostly computed with max/argmax/topk methods which are
# very sensitive to the inputs (e.g. tiny differences may give totally different results)
elif not torch.is_floating_point(batched_object):
return
else:
# indexing the first element does not always work
# e.g. models that output similarity scores of size (N, M) would need to index [0, 0]
slice_ids = [slice(0, index) for index in single_row_object.shape]
batched_row = batched_object[slice_ids]
self.assertFalse(
torch.isnan(batched_row).any(), f"Batched output has `nan` in {model_name} for key={key}"
)
self.assertFalse(
torch.isinf(batched_row).any(), f"Batched output has `inf` in {model_name} for key={key}"
)
self.assertFalse(
torch.isnan(single_row_object).any(), f"Single row output has `nan` in {model_name} for key={key}"
)
self.assertFalse(
torch.isinf(single_row_object).any(), f"Single row output has `inf` in {model_name} for key={key}"
)
try:
torch.testing.assert_close(batched_row, single_row_object, atol=atol, rtol=rtol)
except AssertionError as e:
msg = f"Batched and Single row outputs are not equal in {model_name} for key={key}.\n\n"
msg += str(e)
raise AssertionError(msg)
set_model_tester_for_less_flaky_test(self)
config, batched_input = self.model_tester.prepare_config_and_inputs_for_common()
set_config_for_less_flaky_test(config)
for model_class in self.all_model_classes:
config.output_hidden_states = True
model_name = model_class.__name__
if hasattr(self.model_tester, "prepare_config_and_inputs_for_model_class"):
config, batched_input = self.model_tester.prepare_config_and_inputs_for_model_class(model_class)
batched_input_prepared = self._prepare_for_class(batched_input, model_class)
model = model_class(config).to(torch_device).eval()
set_model_for_less_flaky_test(model)
batch_size = self.model_tester.batch_size
single_row_input = {}
for key, value in batched_input_prepared.items():
if isinstance(value, torch.Tensor) and value.shape[0] % batch_size == 0:
# e.g. musicgen has inputs of size (bs*codebooks). in most cases value.shape[0] == batch_size
single_batch_shape = value.shape[0] // batch_size
single_row_input[key] = value[:single_batch_shape]
else:
single_row_input[key] = value
with torch.no_grad():
model_batched_output = model(**batched_input_prepared)
model_row_output = model(**single_row_input)
if isinstance(model_batched_output, torch.Tensor):
model_batched_output = {"model_output": model_batched_output}
model_row_output = {"model_output": model_row_output}
for key in model_batched_output:
# DETR starts from zero-init queries to decoder, leading to cos_similarity = `nan`
if hasattr(self, "zero_init_hidden_state") and "decoder_hidden_states" in key:
model_batched_output[key] = model_batched_output[key][1:]
model_row_output[key] = model_row_output[key][1:]
recursive_check(model_batched_output[key], model_row_output[key], model_name, key)
def check_training_gradient_checkpointing(self, gradient_checkpointing_kwargs=None):
if not self.model_tester.is_training:
self.skipTest(reason="ModelTester is not configured to run training tests")
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__):
if (
model_class.__name__
in [
*get_values(MODEL_MAPPING_NAMES),
*get_values(MODEL_FOR_BACKBONE_MAPPING_NAMES),
]
or not model_class.supports_gradient_checkpointing
):
# TODO (ydshieh): use `skipTest` once pytest-dev/pytest-subtests/pull/169 is merged
# self.skipTest(reason=f"`supports_gradient_checkpointing` is False for {model_class.__name__}.")
continue
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.use_cache = False
config.return_dict = True
model = model_class(config)
model.to(torch_device)
model.gradient_checkpointing_enable(gradient_checkpointing_kwargs=gradient_checkpointing_kwargs)
model.train()
# unfreeze additional layers
for p in model.parameters():
p.requires_grad_(True)
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
loss = model(**inputs).loss
loss.backward()
optimizer.step()
if self.test_all_params_have_gradient:
for k, v in model.named_parameters():
if v.requires_grad:
self.assertTrue(v.grad is not None, f"{k} in {model_class.__name__} has no gradient!")
def test_training(self):
if not self.model_tester.is_training:
self.skipTest(reason="ModelTester is not configured to run training tests")
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.return_dict = True
if model_class.__name__ in [
*get_values(MODEL_MAPPING_NAMES),
*get_values(MODEL_FOR_BACKBONE_MAPPING_NAMES),
]:
continue
model = model_class(config)
model.to(torch_device)
model.train()
inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
loss = model(**inputs).loss
loss.backward()
def test_causal_lm_can_accept_kwargs(self):
if not getattr(self.model_tester, "is_training", False):
self.skipTest(reason="ModelTester is not configured to run training tests")
valid_model_class = False
incompatible_models = (
"MusicgenForCausalLM",
"MusicgenMelodyForCausalLM",
"MllamaForCausalLM",
"CpmAntForCausalLM",
"GotOcr2ForConditionalGeneration",
)
for model_class in self.all_model_classes:
if (
model_class.__name__ in get_values(MODEL_FOR_CAUSAL_LM_MAPPING_NAMES)
and model_class.__name__ not in incompatible_models
):
valid_model_class = True
if not valid_model_class:
self.skipTest(reason="No causal lm model classes found")
for model_class in self.all_model_classes:
model_name = model_class.__name__
if model_name in get_values(MODEL_FOR_CAUSAL_LM_MAPPING_NAMES) and model_name not in incompatible_models:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
with tempfile.TemporaryDirectory() as tmpdir:
with torch.device(torch_device):
model_eager = AutoModelForCausalLM.from_config(config, torch_dtype=torch.float32)
model_eager.save_pretrained(tmpdir)
with torch.device(torch_device):
model = AutoModelForCausalLM.from_pretrained(tmpdir, torch_dtype=torch.float32)
inputs_dict["num_items_in_batch"] = inputs_dict["input_ids"].shape[0]
inputs_dict["labels"] = inputs_dict["input_ids"]
_ = model(**inputs_dict, return_dict=False)
def test_training_gradient_checkpointing(self):
# Scenario - 1 default behaviour
self.check_training_gradient_checkpointing()
def test_training_gradient_checkpointing_use_reentrant(self):
# Scenario - 2 with `use_reentrant=True` - this is the default value that is used in pytorch's
# torch.utils.checkpoint.checkpoint
self.check_training_gradient_checkpointing(gradient_checkpointing_kwargs={"use_reentrant": True})
def test_training_gradient_checkpointing_use_reentrant_false(self):
# Scenario - 3 with `use_reentrant=False` pytorch suggests users to use this value for
# future releases: https://pytorch.org/docs/stable/checkpoint.html
self.check_training_gradient_checkpointing(gradient_checkpointing_kwargs={"use_reentrant": False})
def test_attention_outputs(self):
if not self.has_attentions:
self.skipTest(reason="Model does not output attentions")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.return_dict = True
seq_len = getattr(self.model_tester, "seq_length", None)
decoder_seq_length = getattr(self.model_tester, "decoder_seq_length", seq_len)
encoder_seq_length = getattr(self.model_tester, "encoder_seq_length", seq_len)
decoder_key_length = getattr(self.model_tester, "decoder_key_length", decoder_seq_length)
encoder_key_length = getattr(self.model_tester, "key_length", encoder_seq_length)
chunk_length = getattr(self.model_tester, "chunk_length", None)
if chunk_length is not None and hasattr(self.model_tester, "num_hashes"):
encoder_seq_length = encoder_seq_length * self.model_tester.num_hashes
for model_class in self.all_model_classes:
inputs_dict["output_attentions"] = True
inputs_dict["output_hidden_states"] = False
config.return_dict = True
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs.encoder_attentions if config.is_encoder_decoder else outputs.attentions
self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
config.output_attentions = True
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs.encoder_attentions if config.is_encoder_decoder else outputs.attentions
self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
if chunk_length is not None:
self.assertListEqual(
list(attentions[0].shape[-4:]),
[self.model_tester.num_attention_heads, encoder_seq_length, chunk_length, encoder_key_length],
)
else:
self.assertListEqual(
list(attentions[0].shape[-3:]),
[self.model_tester.num_attention_heads, encoder_seq_length, encoder_key_length],
)
out_len = len(outputs)
if self.is_encoder_decoder:
correct_outlen = 5
# loss is at first position
if "labels" in inputs_dict:
correct_outlen += 1 # loss is added to beginning
# Question Answering model returns start_logits and end_logits
if model_class.__name__ in [
*get_values(MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES),
*get_values(MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES),
]:
correct_outlen += 1 # start_logits and end_logits instead of only 1 output
if "past_key_values" in outputs:
correct_outlen += 1 # past_key_values have been returned
self.assertEqual(out_len, correct_outlen)
# decoder attentions
decoder_attentions = outputs.decoder_attentions
self.assertIsInstance(decoder_attentions, (list, tuple))
self.assertEqual(len(decoder_attentions), self.model_tester.num_hidden_layers)
self.assertListEqual(
list(decoder_attentions[0].shape[-3:]),
[self.model_tester.num_attention_heads, decoder_seq_length, decoder_key_length],
)
# cross attentions
cross_attentions = outputs.cross_attentions
self.assertIsInstance(cross_attentions, (list, tuple))
self.assertEqual(len(cross_attentions), self.model_tester.num_hidden_layers)
self.assertListEqual(
list(cross_attentions[0].shape[-3:]),
[
self.model_tester.num_attention_heads,
decoder_seq_length,
encoder_key_length,
],
)
# Check attention is always last and order is fine
inputs_dict["output_attentions"] = True
inputs_dict["output_hidden_states"] = True
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
if hasattr(self.model_tester, "num_hidden_states_types"):
added_hidden_states = self.model_tester.num_hidden_states_types
elif self.is_encoder_decoder:
added_hidden_states = 2
else:
added_hidden_states = 1
self.assertEqual(out_len + added_hidden_states, len(outputs))
self_attentions = outputs.encoder_attentions if config.is_encoder_decoder else outputs.attentions
self.assertEqual(len(self_attentions), self.model_tester.num_hidden_layers)
if chunk_length is not None:
self.assertListEqual(
list(self_attentions[0].shape[-4:]),
[self.model_tester.num_attention_heads, encoder_seq_length, chunk_length, encoder_key_length],
)
else:
self.assertListEqual(
list(self_attentions[0].shape[-3:]),
[self.model_tester.num_attention_heads, encoder_seq_length, encoder_key_length],
)
@slow
def test_torchscript_simple(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
self._create_and_check_torchscript(config, inputs_dict)
@slow
def test_torchscript_output_attentions(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.output_attentions = True
self._create_and_check_torchscript(config, inputs_dict)
@slow
def test_torchscript_output_hidden_state(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.output_hidden_states = True
self._create_and_check_torchscript(config, inputs_dict)
# This is copied from `torch/testing/_internal/jit_utils.py::clear_class_registry`
def clear_torch_jit_class_registry(self):
torch._C._jit_clear_class_registry()
torch.jit._recursive.concrete_type_store = torch.jit._recursive.ConcreteTypeStore()
# torch 1.8 has no `_clear_class_state` in `torch.jit._state`
if hasattr(torch.jit._state, "_clear_class_state"):
torch.jit._state._clear_class_state()
def _create_and_check_torchscript(self, config, inputs_dict):
if not self.test_torchscript:
self.skipTest(reason="test_torchscript is set to `False`")
configs_no_init = _config_zero_init(config) # To be sure we have no Nan
configs_no_init.torchscript = True
for model_class in self.all_model_classes:
for attn_implementation in ["eager", "sdpa"]:
if attn_implementation == "sdpa" and (not model_class._supports_sdpa or not is_torch_sdpa_available()):
continue
configs_no_init._attn_implementation = attn_implementation
model = model_class(config=configs_no_init)
model.to(torch_device)
model.eval()
inputs = self._prepare_for_class(inputs_dict, model_class)
main_input_name = model_class.main_input_name
try:
if model.config.is_encoder_decoder:
model.config.use_cache = False # FSTM still requires this hack -> FSTM should probably be refactored similar to BART afterward
main_input = inputs[main_input_name]
attention_mask = inputs["attention_mask"]
decoder_input_ids = inputs["decoder_input_ids"]
decoder_attention_mask = inputs["decoder_attention_mask"]
outputs = model(main_input, attention_mask, decoder_input_ids, decoder_attention_mask)
# `torchscript` doesn't work with outputs containing `Cache` object. However, #35235 makes
# several models to use `Cache` by default instead of the legacy cache (tuple), and
# their `torchscript` tests are failing. We won't support them anyway, but we still want to keep
# the tests for encoder models like `BERT`. So we skip the checks if the model's output contains
# a `Cache` object.
if any(isinstance(x, Cache) for x in outputs):
continue
traced_model = torch.jit.trace(
model, (main_input, attention_mask, decoder_input_ids, decoder_attention_mask)
)
elif "bbox" in inputs and "image" in inputs: # LayoutLMv2 requires additional inputs
input_ids = inputs["input_ids"]
bbox = inputs["bbox"]
image = inputs["image"].tensor
outputs = model(input_ids, bbox, image)
if any(isinstance(x, Cache) for x in outputs):
continue
traced_model = torch.jit.trace(
model, (input_ids, bbox, image), check_trace=False
) # when traced model is checked, an error is produced due to name mangling
elif "bbox" in inputs: # Bros requires additional inputs (bbox)
input_ids = inputs["input_ids"]
bbox = inputs["bbox"]
outputs = model(input_ids, bbox)
if any(isinstance(x, Cache) for x in outputs):
continue
traced_model = torch.jit.trace(
model, (input_ids, bbox), check_trace=False
) # when traced model is checked, an error is produced due to name mangling
elif (
"pixel_values" in inputs and "prompt_pixel_values" in inputs and "prompt_masks" in inputs
): # SegGpt requires additional inputs
pixel_values = inputs["pixel_values"]
prompt_pixel_values = inputs["prompt_pixel_values"]
prompt_masks = inputs["prompt_masks"]
outputs = model(pixel_values, prompt_pixel_values, prompt_masks)
if any(isinstance(x, Cache) for x in outputs):
continue
traced_model = torch.jit.trace(
model, (pixel_values, prompt_pixel_values, prompt_masks), check_trace=False
) # when traced model is checked, an error is produced due to name mangling
else:
main_input = inputs[main_input_name]
if model.config._attn_implementation == "sdpa":
trace_input = {main_input_name: main_input}
if "attention_mask" in inputs:
trace_input["attention_mask"] = inputs["attention_mask"]
else:
self.skipTest(reason="testing SDPA without attention_mask is not supported")
outputs = model(main_input, attention_mask=inputs["attention_mask"])
if any(isinstance(x, Cache) for x in outputs):
continue
# example_kwarg_inputs was introduced in torch==2.0, but it is fine here since SDPA has a requirement on torch>=2.1.
traced_model = torch.jit.trace(model, example_kwarg_inputs=trace_input)
else:
outputs = model(main_input)
if any(isinstance(x, Cache) for x in outputs):
continue
traced_model = torch.jit.trace(model, (main_input,))
except RuntimeError:
self.fail("Couldn't trace module.")
with tempfile.TemporaryDirectory() as tmp_dir_name:
pt_file_name = os.path.join(tmp_dir_name, "traced_model.pt")
try:
torch.jit.save(traced_model, pt_file_name)
except Exception:
self.fail("Couldn't save module.")
try:
loaded_model = torch.jit.load(pt_file_name)
except Exception:
self.fail("Couldn't load module.")
model.to(torch_device)
model.eval()
loaded_model.to(torch_device)
loaded_model.eval()
model_state_dict = model.state_dict()
loaded_model_state_dict = loaded_model.state_dict()
non_persistent_buffers = {}
for key in loaded_model_state_dict.keys():
if key not in model_state_dict.keys():
non_persistent_buffers[key] = loaded_model_state_dict[key]
loaded_model_state_dict = {
key: value for key, value in loaded_model_state_dict.items() if key not in non_persistent_buffers
}
self.assertEqual(set(model_state_dict.keys()), set(loaded_model_state_dict.keys()))
model_buffers = list(model.buffers())
for non_persistent_buffer in non_persistent_buffers.values():
found_buffer = False
for i, model_buffer in enumerate(model_buffers):
if torch.equal(non_persistent_buffer, model_buffer):
found_buffer = True
break
self.assertTrue(found_buffer)
model_buffers.pop(i)
models_equal = True
for layer_name, p1 in model_state_dict.items():
if layer_name in loaded_model_state_dict:
p2 = loaded_model_state_dict[layer_name]
if p1.data.ne(p2.data).sum() > 0:
models_equal = False
self.assertTrue(models_equal)
# Avoid memory leak. Without this, each call increase RAM usage by ~20MB.
# (Even with this call, there are still memory leak by ~0.04MB)
self.clear_torch_jit_class_registry()
def test_torch_fx(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
self._create_and_check_torch_fx_tracing(config, inputs_dict)
def test_torch_fx_output_loss(self):
if self.all_model_classes[0].__name__ == "BloomModel":
self.skipTest(reason="Bloom currently has issues, @michaelbenayoun")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
self._create_and_check_torch_fx_tracing(config, inputs_dict, output_loss=True)
def _create_and_check_torch_fx_tracing(self, config, inputs_dict, output_loss=False):
if not is_torch_fx_available() or not self.fx_compatible:
self.skipTest(
f"Either torch.fx is not available, or the model type {config.model_type} is not compatible with torch.fx"
)
configs_no_init = _config_zero_init(config) # To be sure we have no Nan
configs_no_init.return_dict = False
for model_class in self.all_model_classes:
model = model_class(config=configs_no_init)
model.to(torch_device)
model.eval()
inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=output_loss)
# We may want to test several inputs (various shapes, etc.).
inputs_to_test = [inputs]
if model.config.is_encoder_decoder:
model.config.use_cache = False # FSTM still requires this hack -> FSTM should probably be refactored similar to BART afterward
labels = inputs.get("labels", None)
input_names = [
"attention_mask",
"decoder_attention_mask",
"decoder_input_ids",
"input_features",
"input_ids",
"input_values",
]
if labels is not None:
input_names.append("labels")
else:
input_names = [
"attention_mask",
"bbox",
"input_features",
"input_ids",
"input_values",
"inputs_embeds",
"pixel_values",
"token_type_ids",
"visual_feats",
"visual_pos",
"noise",
]
labels = inputs.get("labels", None)
start_positions = inputs.get("start_positions", None)
end_positions = inputs.get("end_positions", None)
if labels is not None:
input_names.append("labels")
if start_positions is not None:
input_names.append("start_positions")
if end_positions is not None:
input_names.append("end_positions")
if model.config.model_type in _FX_SUPPORTED_MODELS_WITH_KV_CACHE:
input_names.append("past_key_values")
# Generally model_tester.prepare_config_and_inputs_for_common seem not to generate past key values inputs.
if "past_key_values" not in inputs:
batch_size = inputs[next(iter(inputs))].shape[0]
num_heads = model.config.num_attention_heads
head_dim = model.config.hidden_size // model.config.num_attention_heads
cache_shape = (batch_size, num_heads, 0, head_dim)
empty_pkv = tuple(
(
torch.rand(cache_shape, dtype=torch.float, device=torch_device),
torch.rand(cache_shape, dtype=torch.float, device=torch_device),
)
for i in range(model.config.num_hidden_layers)
)
empty_pkv = (
DynamicCache.from_legacy_cache(empty_pkv)
if model_class._supports_cache_class
else empty_pkv
)
cache_length = 9
cache_shape = (batch_size, num_heads, cache_length, head_dim)
non_empty_pkv = tuple(
(
torch.rand(cache_shape, dtype=torch.float, device=torch_device),
torch.rand(cache_shape, dtype=torch.float, device=torch_device),
)
for i in range(model.config.num_hidden_layers)
)
non_empty_pkv = (
DynamicCache.from_legacy_cache(non_empty_pkv)
if model_class._supports_cache_class
else non_empty_pkv
)
inps = copy.deepcopy(inputs_to_test[0])
inputs_to_test[0]["past_key_values"] = empty_pkv
inps["past_key_values"] = non_empty_pkv
inputs_to_test.append(inps)
past_mask = torch.ones(batch_size, cache_length, device=torch_device, dtype=torch.float)
inputs_to_test[1]["attention_mask"] = torch.cat(
(past_mask, inputs_to_test[1]["attention_mask"]), dim=1
)
forward_parameters = inspect.signature(model.forward).parameters
if "input_ids" in forward_parameters and "inputs_embeds" in forward_parameters:
inps = copy.deepcopy(inputs_to_test[0])
embedding_size = (
model.config.embedding_size
if getattr(model.config, "embedding_size", None) is not None
and model.config.model_type != "megatron-bert"
else model.config.hidden_size
)
if (
model.config.model_type in MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES
and model.__class__.__name__
== MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES[model.config.model_type]
):
batch_size, num_choices, sequence_length = inputs["input_ids"].shape
shape = (batch_size, num_choices, sequence_length, embedding_size)
elif inps["input_ids"].ndim == 2:
batch_size, sequence_length = inputs["input_ids"].shape
shape = (batch_size, sequence_length, embedding_size)
else:
self.skipTest("Unknown case")
del inps["input_ids"]
inps["inputs_embeds"] = torch.rand(shape, dtype=torch.float, device=torch_device)
inputs_to_test.append(inps)
for inps in inputs_to_test:
filtered_inputs = {k: v for (k, v) in inps.items() if k in input_names}
input_names_to_trace = list(filtered_inputs.keys())
if model.__class__.__name__ in set(MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES.values()) and (
not hasattr(model.config, "problem_type") or model.config.problem_type is None
):
model.config.problem_type = "single_label_classification"
model.config.use_cache = "past_key_values" in input_names_to_trace
traced_model = symbolic_trace(model, input_names_to_trace)
with torch.no_grad():
traced_output = traced_model(**filtered_inputs)
model_output = model(**filtered_inputs)
def flatten_output(output):
flatten = []
for x in output:
if isinstance(x, (tuple, list)):
flatten += flatten_output(x)
elif not isinstance(x, torch.Tensor):
continue
else:
flatten.append(x)
return flatten
model_output = flatten_output(model_output)
traced_output = flatten_output(traced_output)
num_outputs = len(model_output)
for i in range(num_outputs):
self.assertTrue(
torch.allclose(model_output[i], traced_output[i]),
f"traced {i}th output doesn't match model {i}th output for {model_class}",
)
# Avoid memory leak. Without this, each call increase RAM usage by ~20MB.
# (Even with this call, there are still memory leak by ~0.04MB)
self.clear_torch_jit_class_registry()
def test_headmasking(self):
if not self.test_head_masking:
self.skipTest(reason="Model does not support head masking")
global_rng.seed(42)
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
global_rng.seed()
inputs_dict["output_attentions"] = True
config.output_hidden_states = True
configs_no_init = _config_zero_init(config) # To be sure we have no Nan
for model_class in self.all_model_classes:
model = model_class(config=configs_no_init)
model.to(torch_device)
model.eval()
# Prepare head_mask
# Set require_grad after having prepared the tensor to avoid error (leaf variable has been moved into the graph interior)
head_mask = torch.ones(
self.model_tester.num_hidden_layers,
self.model_tester.num_attention_heads,
device=torch_device,
)
head_mask[0, 0] = 0
head_mask[-1, :-1] = 0
head_mask.requires_grad_(requires_grad=True)
inputs = self._prepare_for_class(inputs_dict, model_class).copy()
inputs["head_mask"] = head_mask
if model.config.is_encoder_decoder:
signature = inspect.signature(model.forward)
arg_names = [*signature.parameters.keys()]
if "decoder_head_mask" in arg_names: # necessary diferentiation because of T5 model
inputs["decoder_head_mask"] = head_mask
if "cross_attn_head_mask" in arg_names:
inputs["cross_attn_head_mask"] = head_mask
outputs = model(**inputs, return_dict=True)
# Test that we can get a gradient back for importance score computation
output = sum(t.sum() for t in outputs[0])
output = output.sum()
output.backward()
multihead_outputs = head_mask.grad
self.assertIsNotNone(multihead_outputs)
self.assertEqual(len(multihead_outputs), self.model_tester.num_hidden_layers)
def check_attentions_validity(attentions):
# Remove Nan
for t in attentions:
self.assertLess(
torch.sum(torch.isnan(t)), t.numel() / 4
) # Check we don't have more than 25% nans (arbitrary)
attentions = [
t.masked_fill(torch.isnan(t), 0.0) for t in attentions
] # remove them (the test is less complete)
self.assertAlmostEqual(attentions[0][..., 0, :, :].flatten().sum().item(), 0.0)
self.assertNotEqual(attentions[0][..., -1, :, :].flatten().sum().item(), 0.0)
if len(attentions) > 2: # encoder-decoder models have only 2 layers in each module
self.assertNotEqual(attentions[1][..., 0, :, :].flatten().sum().item(), 0.0)
self.assertAlmostEqual(attentions[-1][..., -2, :, :].flatten().sum().item(), 0.0)
self.assertNotEqual(attentions[-1][..., -1, :, :].flatten().sum().item(), 0.0)
if model.config.is_encoder_decoder:
check_attentions_validity(outputs.encoder_attentions)
check_attentions_validity(outputs.decoder_attentions)
check_attentions_validity(outputs.cross_attentions)
else:
check_attentions_validity(outputs.attentions)
def test_head_pruning(self):
if not self.test_pruning:
self.skipTest(reason="Pruning is not activated")
for model_class in self.all_model_classes:
(
config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
if "head_mask" in inputs_dict:
del inputs_dict["head_mask"]
inputs_dict["output_attentions"] = True
config.output_hidden_states = False
model = model_class(config=config)
model.to(torch_device)
model.eval()
heads_to_prune = {
0: list(range(1, self.model_tester.num_attention_heads)),
-1: [0],
}
model.prune_heads(heads_to_prune)
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], 1)
# TODO: To have this check, we will need at least 3 layers. Do we really need it?
# self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads)
self.assertEqual(attentions[-1].shape[-3], self.model_tester.num_attention_heads - 1)
def test_head_pruning_save_load_from_pretrained(self):
if not self.test_pruning:
self.skipTest(reason="Pruning is not activated")
for model_class in self.all_model_classes:
(
config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
if "head_mask" in inputs_dict:
del inputs_dict["head_mask"]
inputs_dict["output_attentions"] = True
config.output_hidden_states = False
model = model_class(config=config)
model.to(torch_device)
model.eval()
heads_to_prune = {
0: list(range(1, self.model_tester.num_attention_heads)),
-1: [0],
}
model.prune_heads(heads_to_prune)
with tempfile.TemporaryDirectory() as temp_dir_name:
model.save_pretrained(temp_dir_name)
model = model_class.from_pretrained(temp_dir_name)
model.to(torch_device)
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], 1)
# TODO: To have this check, we will need at least 3 layers. Do we really need it?
# self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads)
self.assertEqual(attentions[-1].shape[-3], self.model_tester.num_attention_heads - 1)
def test_head_pruning_save_load_from_config_init(self):
if not self.test_pruning:
self.skipTest(reason="Pruning is not activated")
for model_class in self.all_model_classes:
(
config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
if "head_mask" in inputs_dict:
del inputs_dict["head_mask"]
inputs_dict["output_attentions"] = True
config.output_hidden_states = False
heads_to_prune = {
0: list(range(1, self.model_tester.num_attention_heads)),
-1: [0],
}
config.pruned_heads = heads_to_prune
model = model_class(config=config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], 1)
# TODO: To have this check, we will need at least 3 layers. Do we really need it?
# self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads)
self.assertEqual(attentions[-1].shape[-3], self.model_tester.num_attention_heads - 1)
def test_head_pruning_integration(self):
if not self.test_pruning:
self.skipTest(reason="Pruning is not activated")
for model_class in self.all_model_classes:
(
config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
if "head_mask" in inputs_dict:
del inputs_dict["head_mask"]
inputs_dict["output_attentions"] = True
config.output_hidden_states = False
heads_to_prune = {1: [1, 2]}
config.pruned_heads = heads_to_prune
model = model_class(config=config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], self.model_tester.num_attention_heads - 0)
self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads - 2)
with tempfile.TemporaryDirectory() as temp_dir_name:
model.save_pretrained(temp_dir_name)
model = model_class.from_pretrained(temp_dir_name)
model.to(torch_device)
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], self.model_tester.num_attention_heads - 0)
self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads - 2)
heads_to_prune = {0: [0], 1: [1, 2]}
model.prune_heads(heads_to_prune)
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
attentions = outputs[-1]
self.assertEqual(attentions[0].shape[-3], self.model_tester.num_attention_heads - 1)
self.assertEqual(attentions[1].shape[-3], self.model_tester.num_attention_heads - 2)
self.assertDictEqual(model.config.pruned_heads, {0: [0], 1: [1, 2]})
def test_hidden_states_output(self):
def check_hidden_states_output(inputs_dict, config, model_class):
model = model_class(config)
model.to(torch_device)
model.eval()
with torch.no_grad():
outputs = model(**self._prepare_for_class(inputs_dict, model_class))
hidden_states = outputs.encoder_hidden_states if config.is_encoder_decoder else outputs.hidden_states
expected_num_layers = getattr(
self.model_tester, "expected_num_hidden_layers", self.model_tester.num_hidden_layers + 1
)
self.assertEqual(len(hidden_states), expected_num_layers)
if hasattr(self.model_tester, "encoder_seq_length"):
seq_length = self.model_tester.encoder_seq_length
if hasattr(self.model_tester, "chunk_length") and self.model_tester.chunk_length > 1:
seq_length = seq_length * self.model_tester.chunk_length
else:
seq_length = self.model_tester.seq_length
self.assertListEqual(
list(hidden_states[0].shape[-2:]),
[seq_length, self.model_tester.hidden_size],
)
if config.is_encoder_decoder:
hidden_states = outputs.decoder_hidden_states
self.assertIsInstance(hidden_states, (list, tuple))
self.assertEqual(len(hidden_states), expected_num_layers)
seq_len = getattr(self.model_tester, "seq_length", None)
decoder_seq_length = getattr(self.model_tester, "decoder_seq_length", seq_len)
self.assertListEqual(
list(hidden_states[0].shape[-2:]),
[decoder_seq_length, self.model_tester.hidden_size],
)
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
inputs_dict["output_hidden_states"] = True
check_hidden_states_output(inputs_dict, config, model_class)
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
config.output_hidden_states = True
check_hidden_states_output(inputs_dict, config, model_class)
def test_retain_grad_hidden_states_attentions(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config.output_hidden_states = True
config.output_attentions = self.has_attentions
# no need to test all models as different heads yield the same functionality
model_class = self.all_model_classes[0]
model = model_class(config)
model.to(torch_device)
inputs = self._prepare_for_class(inputs_dict, model_class)
outputs = model(**inputs)
output = outputs[0]
if config.is_encoder_decoder:
# Seq2Seq models
encoder_hidden_states = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
decoder_hidden_states = outputs.decoder_hidden_states[0]
decoder_hidden_states.retain_grad()
if self.has_attentions:
encoder_attentions = outputs.encoder_attentions[0]
encoder_attentions.retain_grad()
decoder_attentions = outputs.decoder_attentions[0]
decoder_attentions.retain_grad()
cross_attentions = outputs.cross_attentions[0]
cross_attentions.retain_grad()
output.flatten()[0].backward(retain_graph=True)
self.assertIsNotNone(encoder_hidden_states.grad)
self.assertIsNotNone(decoder_hidden_states.grad)
if self.has_attentions:
self.assertIsNotNone(encoder_attentions.grad)
self.assertIsNotNone(decoder_attentions.grad)
self.assertIsNotNone(cross_attentions.grad)
else:
# Encoder-/Decoder-only models
hidden_states = outputs.hidden_states[0]
hidden_states.retain_grad()
if self.has_attentions:
attentions = outputs.attentions[0]
attentions.retain_grad()
output.flatten()[0].backward(retain_graph=True)
self.assertIsNotNone(hidden_states.grad)
if self.has_attentions:
self.assertIsNotNone(attentions.grad)
def test_feed_forward_chunking(self):
(
original_config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
torch.manual_seed(0)
config = copy.deepcopy(original_config)
model = model_class(config)
model.to(torch_device)
model.eval()
hidden_states_no_chunk = model(**self._prepare_for_class(inputs_dict, model_class))[0]
torch.manual_seed(0)
config.chunk_size_feed_forward = 1
model = model_class(config)
model.to(torch_device)
model.eval()
hidden_states_with_chunk = model(**self._prepare_for_class(inputs_dict, model_class))[0]
torch.testing.assert_close(hidden_states_no_chunk, hidden_states_with_chunk, rtol=1e-3, atol=1e-3)
def test_resize_position_vector_embeddings(self):
if not self.test_resize_position_embeddings:
self.skipTest(reason="Model does not have position embeddings")
(
original_config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
config = copy.deepcopy(original_config)
model = model_class(config)
model.to(torch_device)
if self.model_tester.is_training is False:
model.eval()
max_position_embeddings = config.max_position_embeddings
# Retrieve the embeddings and clone theme
if model.config.is_encoder_decoder:
encoder_model_embed, decoder_model_embed = model.get_position_embeddings()
encoder_cloned_embeddings = encoder_model_embed.weight.clone()
decoder_cloned_embeddings = decoder_model_embed.weight.clone()
else:
model_embed = model.get_position_embeddings()
cloned_embeddings = model_embed.weight.clone()
# Check that resizing the position embeddings with a larger max_position_embeddings increases
# the model's postion embeddings size
model.resize_position_embeddings(max_position_embeddings + 10)
self.assertEqual(model.config.max_position_embeddings, max_position_embeddings + 10)
# Check that it actually resizes the embeddings matrix
if model.config.is_encoder_decoder:
encoder_model_embed, decoder_model_embed = model.get_position_embeddings()
self.assertEqual(encoder_model_embed.weight.shape[0], encoder_cloned_embeddings.shape[0] + 10)
self.assertEqual(decoder_model_embed.weight.shape[0], decoder_cloned_embeddings.shape[0] + 10)
else:
model_embed = model.get_position_embeddings()
self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] + 10)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
model(**self._prepare_for_class(inputs_dict, model_class))
# Check that resizing the position embeddings with a smaller max_position_embeddings decreases
# the model's max_position_embeddings
model.resize_position_embeddings(max_position_embeddings - 5)
self.assertEqual(model.config.max_position_embeddings, max_position_embeddings - 5)
# Check that it actually resizes the embeddings matrix
if model.config.is_encoder_decoder:
encoder_model_embed, decoder_model_embed = model.get_position_embeddings()
self.assertEqual(encoder_model_embed.weight.shape[0], encoder_cloned_embeddings.shape[0] - 5)
self.assertEqual(decoder_model_embed.weight.shape[0], decoder_cloned_embeddings.shape[0] - 5)
else:
model_embed = model.get_position_embeddings()
self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] - 5)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
model(**self._prepare_for_class(inputs_dict, model_class))
# Check that adding and removing tokens has not modified the first part of the embedding matrix.
models_equal = True
if model.config.is_encoder_decoder:
for p1, p2 in zip(encoder_cloned_embeddings, encoder_model_embed.weight):
if p1.data.ne(p2.data).sum() > 0:
models_equal = False
for p1, p2 in zip(decoder_cloned_embeddings, decoder_model_embed.weight):
if p1.data.ne(p2.data).sum() > 0:
models_equal = False
else:
for p1, p2 in zip(cloned_embeddings, model_embed.weight):
if p1.data.ne(p2.data).sum() > 0:
models_equal = False
self.assertTrue(models_equal)
def test_resize_tokens_embeddings(self):
if not self.test_resize_embeddings:
self.skipTest(reason="test_resize_embeddings is set to `False`")
(
original_config,
inputs_dict,
) = self.model_tester.prepare_config_and_inputs_for_common()
inputs_dict.pop("labels", None)
for model_class in self.all_model_classes:
config = copy.deepcopy(original_config)
if is_deepspeed_zero3_enabled():
with deepspeed.zero.Init():
model = model_class(config)
else:
model = model_class(config)
model.to(torch_device)
model_embed_pre_resize = model.get_input_embeddings()
type_model_embed_pre_resize = type(model_embed_pre_resize)
if self.model_tester.is_training is False:
model.eval()
model_vocab_size = config.get_text_config().vocab_size
# Retrieve the embeddings and clone theme
model_embed = model.resize_token_embeddings(model_vocab_size)
cloned_embeddings = model_embed.weight.clone()
# Check that resizing the token embeddings with a larger vocab size increases the model's vocab size
model_embed = model.resize_token_embeddings(model_vocab_size + 10)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertEqual(new_model_vocab_size, model_vocab_size + 10)
# Check that it actually resizes the embeddings matrix
self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] + 10)
# Check to make sure the type of embeddings returned post resizing is same as type of input
type_model_embed_post_resize = type(model_embed)
self.assertEqual(type_model_embed_pre_resize, type_model_embed_post_resize)
# Check that added embeddings mean is close to the old embeddings mean
if is_deepspeed_zero3_enabled():
with deepspeed.zero.GatheredParameters(model_embed.weight, modifier_rank=None):
old_embeddings_mean = torch.mean(model_embed.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(model_embed.weight.data[-10:, :], axis=0)
else:
old_embeddings_mean = torch.mean(model_embed.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(model_embed.weight.data[-10:, :], axis=0)
torch.testing.assert_close(old_embeddings_mean, new_embeddings_mean, rtol=1e-3, atol=1e-3)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
if not is_deepspeed_zero3_enabled():
# A distriputed launcher is needed for the forward pass when deepspeed is enabled
model_inputs = self._prepare_for_class(inputs_dict, model_class)
model(**model_inputs)
# Check that resizing the token embeddings with a smaller vocab size decreases the model's vocab size
model_embed = model.resize_token_embeddings(model_vocab_size - 15)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertEqual(new_model_vocab_size, model_vocab_size - 15)
# Check that it actually resizes the embeddings matrix
self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] - 15)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
# Input ids should be clamped to the maximum size of the vocabulary
inputs_dict["input_ids"].clamp_(max=model_vocab_size - 15 - 1)
# make sure that decoder_input_ids are resized as well
if not is_deepspeed_zero3_enabled():
# A distriputed launcher is needed for the forward pass when deepspeed is enabled
if "decoder_input_ids" in inputs_dict:
inputs_dict["decoder_input_ids"].clamp_(max=model_vocab_size - 15 - 1)
model_inputs = self._prepare_for_class(inputs_dict, model_class)
model(**model_inputs)
# Check that adding and removing tokens has not modified the first part of the embedding matrix.
models_equal = True
for p1, p2 in zip(cloned_embeddings, model_embed.weight):
if p1.data.ne(p2.data).sum() > 0:
models_equal = False
self.assertTrue(models_equal)
del model
del config
# Copy again. config changed with embedding resizing (`vocab_size` changed)
config = copy.deepcopy(original_config)
if is_deepspeed_zero3_enabled():
with deepspeed.zero.Init():
model = model_class(config)
else:
model = model_class(config)
model.to(torch_device)
model_vocab_size = config.get_text_config().vocab_size
model.resize_token_embeddings(model_vocab_size + 10, pad_to_multiple_of=1)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertTrue(new_model_vocab_size + 10, model_vocab_size)
model_embed = model.resize_token_embeddings(model_vocab_size, pad_to_multiple_of=64)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertTrue(model_embed.weight.shape[0] // 64, 0)
self.assertTrue(model_embed.weight.shape[0], new_model_vocab_size)
self.assertTrue(new_model_vocab_size, model.vocab_size)
model_embed = model.resize_token_embeddings(model_vocab_size + 13, pad_to_multiple_of=64)
self.assertTrue(model_embed.weight.shape[0] // 64, 0)
# Check that resizing a model to a multiple of pad_to_multiple leads to a model of exactly that size
target_dimension = 128
model_embed = model.resize_token_embeddings(target_dimension, pad_to_multiple_of=64)
self.assertTrue(model_embed.weight.shape[0], target_dimension)
with self.assertRaisesRegex(
ValueError,
"Asking to pad the embedding matrix to a multiple of `1.3`, which is not and integer. Please make sure to pass an integer",
):
model.resize_token_embeddings(model_vocab_size, pad_to_multiple_of=1.3)
# Test when `vocab_size` is smaller than `hidden_size`.
del model
del config
# Copy again. config changed with embedding resizing (`vocab_size` changed)
config = copy.deepcopy(original_config)
config.vocab_size = 4
config.pad_token_id = 3
if is_deepspeed_zero3_enabled():
with deepspeed.zero.Init():
model = model_class(config)
else:
model = model_class(config)
model.to(torch_device)
model_vocab_size = config.get_text_config().vocab_size
# Retrieve the embeddings and clone theme
model_embed = model.resize_token_embeddings(model_vocab_size)
cloned_embeddings = model_embed.weight.clone()
# Check that resizing the token embeddings with a larger vocab size increases the model's vocab size
model_embed = model.resize_token_embeddings(model_vocab_size + 10)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertEqual(new_model_vocab_size, model_vocab_size + 10)
# Check that it actually resizes the embeddings matrix
self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] + 10)
# Check to make sure the type of embeddings returned post resizing is same as type of input
type_model_embed_post_resize = type(model_embed)
self.assertEqual(type_model_embed_pre_resize, type_model_embed_post_resize)
# Check that added embeddings mean is close to the old embeddings mean
if is_deepspeed_zero3_enabled():
with deepspeed.zero.GatheredParameters(model_embed.weight, modifier_rank=None):
old_embeddings_mean = torch.mean(model_embed.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(model_embed.weight.data[-10:, :], axis=0)
else:
old_embeddings_mean = torch.mean(model_embed.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(model_embed.weight.data[-10:, :], axis=0)
torch.testing.assert_close(old_embeddings_mean, new_embeddings_mean, rtol=1e-3, atol=1e-3)
@require_deepspeed
@require_torch_accelerator
def test_resize_tokens_embeddings_with_deepspeed(self):
ds_config = {
"zero_optimization": {
"stage": 3,
"offload_param": {"device": "cpu", "pin_memory": True},
},
}
with _deepspeed_zero3(ds_config):
self.test_resize_tokens_embeddings()
@require_deepspeed
@require_torch_multi_gpu
def test_resize_tokens_embeddings_with_deepspeed_multi_gpu(self):
ds_config = {
"zero_optimization": {
"stage": 3,
},
}
with _deepspeed_zero3(ds_config):
self.test_resize_tokens_embeddings()
def test_resize_embeddings_untied(self):
if not self.test_resize_embeddings:
self.skipTest(reason="test_resize_embeddings is set to `False`")
original_config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
original_config.tie_word_embeddings = False
inputs_dict.pop("labels", None)
# if model cannot untied embeddings -> leave test
if original_config.tie_word_embeddings:
self.skipTest(reason="Model cannot untied embeddings")
for model_class in self.all_model_classes:
config = copy.deepcopy(original_config)
if is_deepspeed_zero3_enabled():
with deepspeed.zero.Init():
model = model_class(config)
else:
model = model_class(config).to(torch_device)
# if no output embeddings -> leave test
if model.get_output_embeddings() is None:
continue
# Check that resizing the token embeddings with a larger vocab size increases the model's vocab size
model_vocab_size = config.get_text_config().vocab_size
model.resize_token_embeddings(model_vocab_size + 10)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertEqual(new_model_vocab_size, model_vocab_size + 10)
output_embeds = model.get_output_embeddings()
self.assertEqual(output_embeds.weight.shape[0], model_vocab_size + 10)
# Check bias if present
if output_embeds.bias is not None:
self.assertEqual(output_embeds.bias.shape[0], model_vocab_size + 10)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
if not is_deepspeed_zero3_enabled():
# A distriputed launcher is needed for the forward pass when deepspeed is enabled
model(**self._prepare_for_class(inputs_dict, model_class))
# Test multivariate resizing.
model.resize_token_embeddings(model_vocab_size + 10)
output_embeds = model.get_output_embeddings()
# Check that added embeddings mean is close to the old embeddings mean
if is_deepspeed_zero3_enabled():
with deepspeed.zero.GatheredParameters(output_embeds.weight, modifier_rank=None):
old_embeddings_mean = torch.mean(output_embeds.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(output_embeds.weight.data[-10:, :], axis=0)
else:
old_embeddings_mean = torch.mean(output_embeds.weight.data[:-10, :], axis=0)
new_embeddings_mean = torch.mean(output_embeds.weight.data[-10:, :], axis=0)
torch.testing.assert_close(old_embeddings_mean, new_embeddings_mean, rtol=1e-3, atol=1e-3)
# check if the old bias mean close to added bias mean.
if output_embeds.bias is not None:
if is_deepspeed_zero3_enabled():
with deepspeed.zero.GatheredParameters(output_embeds.bias, modifier_rank=None):
old_bias_mean = torch.mean(output_embeds.bias.data[:-10], axis=0)
new_bias_mean = torch.mean(output_embeds.bias.data[-10:], axis=0)
else:
old_bias_mean = torch.mean(output_embeds.bias.data[:-10], axis=0)
new_bias_mean = torch.mean(output_embeds.bias.data[-10:], axis=0)
torch.testing.assert_close(old_bias_mean, new_bias_mean, rtol=1e-5, atol=1e-5)
# Check that resizing the token embeddings with a smaller vocab size decreases the model's vocab size
model.resize_token_embeddings(model_vocab_size - 15)
new_model_vocab_size = model.config.get_text_config().vocab_size
self.assertEqual(new_model_vocab_size, model_vocab_size - 15)
# Check that it actually resizes the embeddings matrix
output_embeds = model.get_output_embeddings()
self.assertEqual(output_embeds.weight.shape[0], model_vocab_size - 15)
# Check bias if present
if output_embeds.bias is not None:
self.assertEqual(output_embeds.bias.shape[0], model_vocab_size - 15)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
# Input ids should be clamped to the maximum size of the vocabulary
inputs_dict["input_ids"].clamp_(max=model_vocab_size - 15 - 1)
if "decoder_input_ids" in inputs_dict:
inputs_dict["decoder_input_ids"].clamp_(max=model_vocab_size - 15 - 1)
# Check that the model can still do a forward pass successfully (every parameter should be resized)
if not is_deepspeed_zero3_enabled():
# A distriputed launcher is needed for the forward pass when deepspeed is enabled
model(**self._prepare_for_class(inputs_dict, model_class))
@require_deepspeed
@require_torch_accelerator
def test_resize_embeddings_untied_with_deepspeed(self):
ds_config = {
"zero_optimization": {
"stage": 3,
"offload_param": {"device": "cpu", "pin_memory": True},
},
}
with _deepspeed_zero3(ds_config):
self.test_resize_embeddings_untied()
@require_deepspeed
@require_torch_multi_gpu
def test_resize_embeddings_untied_with_deepspeed_multi_gpu(self):
ds_config = {
"zero_optimization": {
"stage": 3,
},
}
with _deepspeed_zero3(ds_config):
self.test_resize_embeddings_untied()
def test_model_get_set_embeddings(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
self.assertIsInstance(model.get_input_embeddings(), (nn.Embedding, AdaptiveEmbedding))
new_input_embedding_layer = nn.Embedding(10, 10)
model.set_input_embeddings(new_input_embedding_layer)
self.assertEqual(model.get_input_embeddings(), new_input_embedding_layer)
x = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(x, nn.Linear))
def test_model_main_input_name(self):
for model_class in self.all_model_classes:
model_signature = inspect.signature(getattr(model_class, "forward"))
# The main input is the name of the argument after `self`
observed_main_input_name = list(model_signature.parameters.keys())[1]
self.assertEqual(model_class.main_input_name, observed_main_input_name)
def test_correct_missing_keys(self):
if not self.test_missing_keys:
self.skipTest(reason="test_missing_keys is set to `False`")
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
base_model_prefix = model.base_model_prefix
if hasattr(model, base_model_prefix):
extra_params = {k: v for k, v in model.named_parameters() if not k.startswith(base_model_prefix)}
extra_params.update({k: v for k, v in model.named_buffers() if not k.startswith(base_model_prefix)})
# Some models define this as None
if model._keys_to_ignore_on_load_missing:
for key in model._keys_to_ignore_on_load_missing:
extra_params.pop(key, None)
if not extra_params:
# In that case, we *are* on a head model, but every
# single key is not actual parameters and this is
# tested in `test_tied_model_weights_key_ignore` test.
continue
with tempfile.TemporaryDirectory() as temp_dir_name:
model.base_model.save_pretrained(temp_dir_name)
model, loading_info = model_class.from_pretrained(temp_dir_name, output_loading_info=True)
self.assertGreater(len(loading_info["missing_keys"]), 0, model.__class__.__name__)
def test_tie_model_weights(self):
if not self.test_torchscript:
self.skipTest(reason="test_torchscript is set to `False`")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
def check_same_values(layer_1, layer_2):
equal = True
for p1, p2 in zip(layer_1.weight, layer_2.weight):
if p1.data.ne(p2.data).sum() > 0:
equal = False
return equal
for model_class in self.all_model_classes:
config.torchscript = True
model_not_tied = model_class(config)
if model_not_tied.get_output_embeddings() is None:
continue
config_tied = copy.deepcopy(config)
config_tied.torchscript = False
model_tied = model_class(config_tied)
params_tied = list(model_tied.parameters())
# Check that the embedding layer and decoding layer are the same in size and in value
# self.assertTrue(check_same_values(embeddings, decoding))
# Check that after resize they remain tied.
vocab_size = config.get_text_config().vocab_size
model_tied.resize_token_embeddings(vocab_size + 10)
params_tied_2 = list(model_tied.parameters())
self.assertEqual(len(params_tied_2), len(params_tied))
@require_safetensors
def test_can_use_safetensors(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model_tied = model_class(config)
with tempfile.TemporaryDirectory() as d:
try:
model_tied.save_pretrained(d, safe_serialization=True)
except Exception as e:
raise Exception(f"Class {model_class.__name__} cannot be saved using safetensors: {e}")
model_reloaded, infos = model_class.from_pretrained(d, output_loading_info=True)
# Checking the state dicts are correct
reloaded_state = model_reloaded.state_dict()
for k, v in model_tied.state_dict().items():
self.assertIn(k, reloaded_state, f"Key {k} is missing from reloaded")
torch.testing.assert_close(
v, reloaded_state[k], msg=lambda x: f"{model_class.__name__}: Tensor {k}: {x}"
)
# Checking there was no complain of missing weights
self.assertEqual(infos["missing_keys"], [])
# Checking the tensor sharing are correct
ptrs = defaultdict(list)
for k, v in model_tied.state_dict().items():
ptrs[v.data_ptr()].append(k)
shared_ptrs = {k: v for k, v in ptrs.items() if len(v) > 1}
for _, shared_names in shared_ptrs.items():
reloaded_ptrs = {reloaded_state[k].data_ptr() for k in shared_names}
self.assertEqual(
len(reloaded_ptrs),
1,
f"The shared pointers are incorrect, found different pointers for keys {shared_names}",
)
def test_load_save_without_tied_weights(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
config.tie_word_embeddings = False
for model_class in self.all_model_classes:
model = model_class(config)
with tempfile.TemporaryDirectory() as d:
model.save_pretrained(d)
model_reloaded, infos = model_class.from_pretrained(d, output_loading_info=True)
# Checking the state dicts are correct
reloaded_state = model_reloaded.state_dict()
for k, v in model.state_dict().items():
self.assertIn(k, reloaded_state, f"Key {k} is missing from reloaded")
torch.testing.assert_close(
v, reloaded_state[k], msg=lambda x: f"{model_class.__name__}: Tensor {k}: {x}"
)
# Checking there was no complain of missing weights
self.assertEqual(infos["missing_keys"], [])
def test_tied_weights_keys(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
config.get_text_config().tie_word_embeddings = True
for model_class in self.all_model_classes:
model_tied = model_class(config)
ptrs = collections.defaultdict(list)
for name, tensor in model_tied.state_dict().items():
ptrs[id_tensor_storage(tensor)].append(name)
# These are all the pointers of shared tensors.
tied_params = [names for _, names in ptrs.items() if len(names) > 1]
tied_weight_keys = model_tied._tied_weights_keys if model_tied._tied_weights_keys is not None else []
# Detect we get a hit for each key
for key in tied_weight_keys:
is_tied_key = any(re.search(key, p) for group in tied_params for p in group)
self.assertTrue(is_tied_key, f"{key} is not a tied weight key for {model_class}.")
# Removed tied weights found from tied params -> there should only be one left after
for key in tied_weight_keys:
for i in range(len(tied_params)):
tied_params[i] = [p for p in tied_params[i] if re.search(key, p) is None]
tied_params = [group for group in tied_params if len(group) > 1]
self.assertListEqual(
tied_params,
[],
f"Missing `_tied_weights_keys` for {model_class}: add all of {tied_params} except one.",
)
def test_model_weights_reload_no_missing_tied_weights(self):
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(tmp_dir)
# We are nuking ALL weights on file, so every parameter should
# yell on load. We're going to detect if we yell too much, or too little.
placeholder_dict = {"tensor": torch.tensor([1, 2])}
safe_save_file(placeholder_dict, os.path.join(tmp_dir, "model.safetensors"), metadata={"format": "pt"})
model_reloaded, infos = model_class.from_pretrained(tmp_dir, output_loading_info=True)
prefix = f"{model_reloaded.base_model_prefix}."
params = dict(model_reloaded.named_parameters())
params.update(dict(model_reloaded.named_buffers()))
param_names = {k[len(prefix) :] if k.startswith(prefix) else k for k in params.keys()}
missing_keys = set(infos["missing_keys"])
extra_missing = missing_keys - param_names
# Remove tied weights from extra missing: they are normally not warned as missing if their tied
# counterpart is present but here there are no weights at all so we do get the warning.
ptrs = collections.defaultdict(list)
for name, tensor in model_reloaded.state_dict().items():
ptrs[id_tensor_storage(tensor)].append(name)
tied_params = [names for _, names in ptrs.items() if len(names) > 1]
for group in tied_params:
group = {k[len(prefix) :] if k.startswith(prefix) else k for k in group}
# We remove the group from extra_missing if not all weights from group are in it
if len(group - extra_missing) > 0:
extra_missing = extra_missing - set(group)
self.assertEqual(
extra_missing,
set(),
f"This model {model_class.__name__} might be missing some `keys_to_ignore`: {extra_missing}. "
f"For debugging, tied parameters are {tied_params}",
)
missed_missing = param_names - missing_keys
# Remove nonpersistent buffers from missed_missing
buffers = [n for n, _ in model_reloaded.named_buffers()]
nonpersistent_buffers = {n for n in buffers if n not in model_reloaded.state_dict()}
nonpersistent_buffers = {
k[len(prefix) :] if k.startswith(prefix) else k for k in nonpersistent_buffers
}
missed_missing = missed_missing - nonpersistent_buffers
if model_reloaded._keys_to_ignore_on_load_missing is None:
expected_missing = set()
else:
expected_missing = set(model_reloaded._keys_to_ignore_on_load_missing)
self.assertEqual(
missed_missing,
expected_missing,
f"This model {model_class.__name__} ignores keys {missed_missing} but they look like real"
" parameters. If they are non persistent buffers make sure to instantiate them with"
" `persistent=False`",
)
def test_model_outputs_equivalence(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(t):
t[t != t] = 0
return t
def check_equivalence(model, tuple_inputs, dict_inputs, additional_kwargs={}):
with torch.no_grad():
tuple_output = model(**tuple_inputs, return_dict=False, **additional_kwargs)
dict_output = model(**dict_inputs, return_dict=True, **additional_kwargs).to_tuple()
def recursive_check(tuple_object, dict_object):
if isinstance(tuple_object, (List, Tuple)):
for tuple_iterable_value, dict_iterable_value in zip(tuple_object, dict_object):
recursive_check(tuple_iterable_value, dict_iterable_value)
elif isinstance(tuple_object, Dict):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values(), dict_object.values()
):
recursive_check(tuple_iterable_value, dict_iterable_value)
elif tuple_object is None:
return
# model might return non-tensors objects (e.g. Cache class)
elif isinstance(tuple_object, torch.Tensor):
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(tuple_object), set_nan_tensor_to_zero(dict_object), atol=1e-5
),
msg=(
"Tuple and dict output are not equal. Difference:"
f" {torch.max(torch.abs(tuple_object - dict_object))}. Tuple has `nan`:"
f" {torch.isnan(tuple_object).any()} and `inf`: {torch.isinf(tuple_object)}. Dict has"
f" `nan`: {torch.isnan(dict_object).any()} and `inf`: {torch.isinf(dict_object)}."
),
)
recursive_check(tuple_output, dict_output)
for model_class in self.all_model_classes:
model = model_class(config)
model.to(torch_device)
model.eval()
tuple_inputs = self._prepare_for_class(inputs_dict, model_class)
dict_inputs = self._prepare_for_class(inputs_dict, model_class)
check_equivalence(model, tuple_inputs, dict_inputs)
tuple_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
dict_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
check_equivalence(model, tuple_inputs, dict_inputs)
tuple_inputs = self._prepare_for_class(inputs_dict, model_class)
dict_inputs = self._prepare_for_class(inputs_dict, model_class)
check_equivalence(model, tuple_inputs, dict_inputs, {"output_hidden_states": True})
tuple_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
dict_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
check_equivalence(model, tuple_inputs, dict_inputs, {"output_hidden_states": True})
if self.has_attentions:
tuple_inputs = self._prepare_for_class(inputs_dict, model_class)
dict_inputs = self._prepare_for_class(inputs_dict, model_class)
check_equivalence(model, tuple_inputs, dict_inputs, {"output_attentions": True})
tuple_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
dict_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
check_equivalence(model, tuple_inputs, dict_inputs, {"output_attentions": True})
tuple_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
dict_inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
check_equivalence(
model, tuple_inputs, dict_inputs, {"output_hidden_states": True, "output_attentions": True}
)
# Don't copy this method to model specific test file!
# TODO: remove this method once the issues are all fixed!
def _make_attention_mask_non_null(self, inputs_dict):
"""Make sure no sequence has all zeros as attention mask"""
for k in ["attention_mask", "encoder_attention_mask", "decoder_attention_mask"]:
if k in inputs_dict:
attention_mask = inputs_dict[k]
# Make sure no all 0s attention masks - to avoid failure at this moment.
# Put `1` at the beginning of sequences to make it still work when combining causal attention masks.
# TODO: remove this line once a fix regarding large negative values for attention mask is done.
attention_mask = torch.cat(
[torch.ones_like(attention_mask[:, :1], dtype=attention_mask.dtype), attention_mask[:, 1:]], dim=-1
)
# Here we make the first sequence with all 0s as attention mask.
# Currently, this will fail for `TFWav2Vec2Model`. This is caused by the different large negative
# values, like `1e-4`, `1e-9`, `1e-30` and `-inf` for attention mask across models/frameworks.
# TODO: enable this block once the large negative values thing is cleaned up.
# (see https://github.com/huggingface/transformers/issues/14859)
# attention_mask = torch.cat(
# [torch.zeros_like(attention_mask[:1], dtype=attention_mask.dtype), attention_mask[1:]],
# dim=0
# )
inputs_dict[k] = attention_mask
# Don't copy this method to model specific test file!
# TODO: remove this method once the issues are all fixed!
def _postprocessing_to_ignore_test_cases(self, tf_outputs, pt_outputs, model_class):
"""For temporarily ignoring some failed test cases (issues to be fixed)"""
tf_keys = {k for k, v in tf_outputs.items() if v is not None}
pt_keys = {k for k, v in pt_outputs.items() if v is not None}
key_differences = tf_keys.symmetric_difference(pt_keys)
if model_class.__name__ in [
"FlaubertWithLMHeadModel",
"FunnelForPreTraining",
"ElectraForPreTraining",
"XLMWithLMHeadModel",
]:
for k in key_differences:
if k in ["loss", "losses"]:
tf_keys.discard(k)
pt_keys.discard(k)
elif model_class.__name__.startswith("GPT2"):
# `TFGPT2` has `past_key_values` as a tensor while `GPT2` has it as a tuple.
tf_keys.discard("past_key_values")
pt_keys.discard("past_key_values")
# create new outputs from the remaining fields
new_tf_outputs = type(tf_outputs)(**{k: tf_outputs[k] for k in tf_keys})
new_pt_outputs = type(pt_outputs)(**{k: pt_outputs[k] for k in pt_keys})
return new_tf_outputs, new_pt_outputs
# Copied from tests.test_modeling_tf_common.TFModelTesterMixin.check_pt_tf_outputs
def check_pt_tf_outputs(self, tf_outputs, pt_outputs, model_class, tol=1e-4, name="outputs", attributes=None):
"""Check the outputs from PyTorch and TensorFlow models are close enough. Checks are done in a recursive way.
Args:
model_class: The class of the model that is currently testing. For example, `TFBertModel`,
TFBertForMaskedLM`, `TFBertForSequenceClassification`, etc. Mainly used for providing more informative
error messages.
name (`str`): The name of the output. For example, `output.hidden_states`, `output.attentions`, etc.
attributes (`Tuple[str]`): The names of the output's element if the output is a tuple/list with each element
being a named field in the output.
"""
self.assertEqual(type(name), str)
if attributes is not None:
self.assertEqual(type(attributes), tuple, f"{name}: The argument `attributes` should be a `tuple`")
# Allow `ModelOutput` (e.g. `CLIPOutput` has `text_model_output` and `vision_model_output`).
if isinstance(tf_outputs, ModelOutput):
self.assertTrue(
isinstance(pt_outputs, ModelOutput),
f"{name}: `pt_outputs` should an instance of `ModelOutput` when `tf_outputs` is",
)
# Don't copy this block to model specific test file!
# TODO: remove this method and this line after issues are fixed
tf_outputs, pt_outputs = self._postprocessing_to_ignore_test_cases(tf_outputs, pt_outputs, model_class)
tf_keys = [k for k, v in tf_outputs.items() if v is not None]
pt_keys = [k for k, v in pt_outputs.items() if v is not None]
self.assertEqual(tf_keys, pt_keys, f"{name}: Output keys differ between TF and PyTorch")
# convert to the case of `tuple`
# appending each key to the current (string) `name`
attributes = tuple([f"{name}.{k}" for k in tf_keys])
self.check_pt_tf_outputs(
tf_outputs.to_tuple(), pt_outputs.to_tuple(), model_class, tol=tol, name=name, attributes=attributes
)
# Allow `list` (e.g. `TransfoXLModelOutput.mems` is a list of tensors.)
elif type(tf_outputs) in [tuple, list]:
self.assertEqual(type(tf_outputs), type(pt_outputs), f"{name}: Output types differ between TF and PyTorch")
self.assertEqual(len(tf_outputs), len(pt_outputs), f"{name}: Output lengths differ between TF and PyTorch")
if attributes is not None:
# case 1: each output has assigned name (e.g. a tuple form of a `ModelOutput`)
self.assertEqual(
len(attributes),
len(tf_outputs),
f"{name}: The tuple `attributes` should have the same length as `tf_outputs`",
)
else:
# case 2: each output has no assigned name (e.g. hidden states of each layer) -> add an index to `name`
attributes = tuple([f"{name}_{idx}" for idx in range(len(tf_outputs))])
for tf_output, pt_output, attr in zip(tf_outputs, pt_outputs, attributes):
if isinstance(pt_output, DynamicCache):
pt_output = pt_output.to_legacy_cache()
self.check_pt_tf_outputs(tf_output, pt_output, model_class, tol=tol, name=attr)
elif isinstance(tf_outputs, tf.Tensor):
self.assertTrue(
isinstance(pt_outputs, torch.Tensor), f"{name}: `pt_outputs` should a tensor when `tf_outputs` is"
)
tf_outputs = tf_outputs.numpy()
pt_outputs = pt_outputs.detach().to("cpu").numpy()
self.assertEqual(
tf_outputs.shape, pt_outputs.shape, f"{name}: Output shapes differ between TF and PyTorch"
)
# deal with NumPy's scalars to make replacing nan values by 0 work.
if np.isscalar(tf_outputs):
tf_outputs = np.array([tf_outputs])
pt_outputs = np.array([pt_outputs])
tf_nans = np.isnan(tf_outputs)
pt_nans = np.isnan(pt_outputs)
pt_outputs[tf_nans] = 0
tf_outputs[tf_nans] = 0
pt_outputs[pt_nans] = 0
tf_outputs[pt_nans] = 0
max_diff = np.amax(np.abs(tf_outputs - pt_outputs))
self.assertLessEqual(
max_diff,
tol,
f"{name}: Difference between PyTorch and TF is {max_diff} (>= {tol}) for {model_class.__name__}",
)
else:
raise ValueError(
"`tf_outputs` should be an instance of `ModelOutput`, a `tuple`, or an instance of `tf.Tensor`. Got"
f" {type(tf_outputs)} instead."
)
def prepare_tf_inputs_from_pt_inputs(self, pt_inputs_dict):
tf_inputs_dict = {}
for key, tensor in pt_inputs_dict.items():
# skip key that does not exist in tf
if isinstance(tensor, bool):
tf_inputs_dict[key] = tensor
elif key == "input_values":
tf_inputs_dict[key] = tf.convert_to_tensor(tensor.cpu().numpy(), dtype=tf.float32)
elif key == "pixel_values":
tf_inputs_dict[key] = tf.convert_to_tensor(tensor.cpu().numpy(), dtype=tf.float32)
elif key == "input_features":
tf_inputs_dict[key] = tf.convert_to_tensor(tensor.cpu().numpy(), dtype=tf.float32)
# other general float inputs
elif tensor.is_floating_point():
tf_inputs_dict[key] = tf.convert_to_tensor(tensor.cpu().numpy(), dtype=tf.float32)
else:
tf_inputs_dict[key] = tf.convert_to_tensor(tensor.cpu().numpy(), dtype=tf.int32)
return tf_inputs_dict
def check_pt_tf_models(self, tf_model, pt_model, pt_inputs_dict):
tf_inputs_dict = self.prepare_tf_inputs_from_pt_inputs(pt_inputs_dict)
# send pytorch inputs to the correct device
pt_inputs_dict = {
k: v.to(device=torch_device) if isinstance(v, torch.Tensor) else v for k, v in pt_inputs_dict.items()
}
# send pytorch model to the correct device
pt_model.to(torch_device)
# Check predictions on first output (logits/hidden-states) are close enough given low-level computational differences
pt_model.eval()
with torch.no_grad():
pt_outputs = pt_model(**pt_inputs_dict)
tf_outputs = tf_model(tf_inputs_dict)
# tf models returned loss is usually a tensor rather than a scalar.
# (see `hf_compute_loss`: it uses `tf.keras.losses.Reduction.NONE`)
# Change it here to a scalar to match PyTorch models' loss
tf_loss = getattr(tf_outputs, "loss", None)
if tf_loss is not None:
tf_outputs.loss = tf.math.reduce_mean(tf_loss)
self.check_pt_tf_outputs(tf_outputs, pt_outputs, type(pt_model))
@is_pt_tf_cross_test
def test_pt_tf_model_equivalence(self, allow_missing_keys=False):
import transformers
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
tf_model_class_name = "TF" + model_class.__name__ # Add the "TF" at the beginning
if not hasattr(transformers, tf_model_class_name):
self.skipTest(reason="transformers does not have TF version of this model yet")
# Output all for aggressive testing
config.output_hidden_states = True
config.output_attentions = self.has_attentions
# Make sure no sequence has all zeros as attention mask, otherwise some tests fail due to the inconsistency
# of the usage `1e-4`, `1e-9`, `1e-30`, `-inf`.
# TODO: Use a uniform value for all models, make sure all tests pass without this processing, and remove it.
self._make_attention_mask_non_null(inputs_dict)
tf_model_class = getattr(transformers, tf_model_class_name)
pt_model = model_class(config).eval()
tf_model = tf_model_class(config)
pt_inputs_dict = self._prepare_for_class(inputs_dict, model_class)
pt_inputs_dict_with_labels = self._prepare_for_class(
inputs_dict,
model_class,
# Not all models accept "labels" in the forward pass (yet :) )
return_labels=True if "labels" in inspect.signature(model_class.forward).parameters.keys() else False,
)
# make sure only tf inputs are forward that actually exist in function args
tf_input_keys = set(inspect.signature(tf_model.call).parameters.keys())
# remove all head masks
tf_input_keys.discard("head_mask")
tf_input_keys.discard("cross_attn_head_mask")
tf_input_keys.discard("decoder_head_mask")
pt_inputs_dict = {k: v for k, v in pt_inputs_dict.items() if k in tf_input_keys}
pt_inputs_dict_with_labels = {k: v for k, v in pt_inputs_dict_with_labels.items() if k in tf_input_keys}
# For some models (e.g. base models), there is no label returned.
# Set the input dict to `None` to avoid check outputs twice for the same input dicts.
if not set(pt_inputs_dict_with_labels.keys()).symmetric_difference(pt_inputs_dict.keys()):
pt_inputs_dict_with_labels = None
# Check we can load pt model in tf and vice-versa with model => model functions
# Here requires `tf_inputs_dict` to build `tf_model`
tf_inputs_dict = self.prepare_tf_inputs_from_pt_inputs(pt_inputs_dict)
tf_model = transformers.load_pytorch_model_in_tf2_model(
tf_model, pt_model, tf_inputs=tf_inputs_dict, allow_missing_keys=allow_missing_keys
)
pt_model = transformers.load_tf2_model_in_pytorch_model(
pt_model, tf_model, allow_missing_keys=allow_missing_keys
)
# Original test: check without `labels`
self.check_pt_tf_models(tf_model, pt_model, pt_inputs_dict)
# check with `labels`
if pt_inputs_dict_with_labels:
self.check_pt_tf_models(tf_model, pt_model, pt_inputs_dict_with_labels)
# Check we can load pt model in tf and vice-versa with checkpoint => model functions
with tempfile.TemporaryDirectory() as tmpdirname:
pt_checkpoint_path = os.path.join(tmpdirname, "pt_model.bin")
torch.save(pt_model.state_dict(), pt_checkpoint_path)
tf_model = transformers.load_pytorch_checkpoint_in_tf2_model(
tf_model, pt_checkpoint_path, allow_missing_keys=allow_missing_keys
)
tf_checkpoint_path = os.path.join(tmpdirname, "tf_model.h5")
tf_model.save_weights(tf_checkpoint_path)
pt_model = transformers.load_tf2_checkpoint_in_pytorch_model(
pt_model, tf_checkpoint_path, allow_missing_keys=allow_missing_keys
)
# Original test: check without `labels`
self.check_pt_tf_models(tf_model, pt_model, pt_inputs_dict)
# check with `labels`
if pt_inputs_dict_with_labels:
self.check_pt_tf_models(tf_model, pt_model, pt_inputs_dict_with_labels)
def assert_almost_equals(self, a: np.ndarray, b: np.ndarray, tol: float):
diff = np.abs((a - b)).max()
self.assertLessEqual(diff, tol, f"Difference between torch and flax is {diff} (>= {tol}).")
def check_pt_flax_outputs(self, fx_outputs, pt_outputs, model_class, tol=1e-4, name="outputs", attributes=None):
"""
Args:
model_class: The class of the model that is currently testing. For example, ..., etc.
Currently unused, but it could make debugging easier and faster.
names: A string, or a list of strings. These specify what fx_outputs/pt_outputs represent in the model outputs.
Currently unused, but in the future, we could use this information to make the error message clearer
by giving the name(s) of the output tensor(s) with large difference(s) between PT and Flax.
"""
self.assertEqual(type(name), str)
if attributes is not None:
self.assertEqual(type(attributes), tuple, f"{name}: The argument `attributes` should be a `tuple`")
# Allow `ModelOutput` (e.g. `CLIPOutput` has `text_model_output` and `vision_model_output`).
if isinstance(fx_outputs, ModelOutput):
self.assertTrue(
isinstance(pt_outputs, ModelOutput),
f"{name}: `pt_outputs` should an instance of `ModelOutput` when `fx_outputs` is",
)
fx_keys = tuple([k for k, v in fx_outputs.items() if v is not None])
pt_keys = tuple([k for k, v in pt_outputs.items() if v is not None])
self.assertEqual(fx_keys, pt_keys, f"{name}: Output keys differ between Flax and PyTorch")
# convert to the case of `tuple`
# appending each key to the current (string) `name`
attributes = tuple([f"{name}.{k}" for k in fx_keys])
self.check_pt_flax_outputs(
fx_outputs.to_tuple(), pt_outputs.to_tuple(), model_class, tol=tol, name=name, attributes=attributes
)
# Allow `list` (e.g. `TransfoXLModelOutput.mems` is a list of tensors.)
elif type(fx_outputs) in [tuple, list]:
self.assertEqual(
type(fx_outputs), type(pt_outputs), f"{name}: Output types differ between Flax and PyTorch"
)
self.assertEqual(
len(fx_outputs), len(pt_outputs), f"{name}: Output lengths differ between Flax and PyTorch"
)
if attributes is not None:
# case 1: each output has assigned name (e.g. a tuple form of a `ModelOutput`)
self.assertEqual(
len(attributes),
len(fx_outputs),
f"{name}: The tuple `attributes` should have the same length as `fx_outputs`",
)
else:
# case 2: each output has no assigned name (e.g. hidden states of each layer) -> add an index to `name`
attributes = tuple([f"{name}_{idx}" for idx in range(len(fx_outputs))])
for fx_output, pt_output, attr in zip(fx_outputs, pt_outputs, attributes):
self.check_pt_flax_outputs(fx_output, pt_output, model_class, tol=tol, name=attr)
elif isinstance(fx_outputs, jnp.ndarray):
self.assertTrue(
isinstance(pt_outputs, torch.Tensor), f"{name}: `pt_outputs` should a tensor when `fx_outputs` is"
)
# Using `np.asarray` gives `ValueError: assignment destination is read-only` at the line `fx_outputs[fx_nans] = 0`.
fx_outputs = np.array(fx_outputs)
pt_outputs = pt_outputs.detach().to("cpu").numpy()
self.assertEqual(
fx_outputs.shape, pt_outputs.shape, f"{name}: Output shapes differ between Flax and PyTorch"
)
# deal with NumPy's scalars to make replacing nan values by 0 work.
if np.isscalar(fx_outputs):
fx_outputs = np.array([fx_outputs])
pt_outputs = np.array([pt_outputs])
fx_nans = np.isnan(fx_outputs)
pt_nans = np.isnan(pt_outputs)
pt_outputs[fx_nans] = 0
fx_outputs[fx_nans] = 0
pt_outputs[pt_nans] = 0
fx_outputs[pt_nans] = 0
max_diff = np.amax(np.abs(fx_outputs - pt_outputs))
self.assertLessEqual(
max_diff, tol, f"{name}: Difference between PyTorch and Flax is {max_diff} (>= {tol})."
)
else:
raise ValueError(
"`fx_outputs` should be an instance of `ModelOutput`, a `tuple`, or an instance of `jnp.ndarray`. Got"
f" {type(fx_outputs)} instead."
)
@is_pt_flax_cross_test
def test_equivalence_pt_to_flax(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__):
fx_model_class_name = "Flax" + model_class.__name__
if not hasattr(transformers, fx_model_class_name):
self.skipTest(reason="No Flax model exists for this class")
# Output all for aggressive testing
config.output_hidden_states = True
config.output_attentions = self.has_attentions
fx_model_class = getattr(transformers, fx_model_class_name)
# load PyTorch class
pt_model = model_class(config).eval()
# Flax models don't use the `use_cache` option and cache is not returned as a default.
# So we disable `use_cache` here for PyTorch model.
pt_model.config.use_cache = False
# load Flax class
fx_model = fx_model_class(config, dtype=jnp.float32)
# make sure only flax inputs are forward that actually exist in function args
fx_input_keys = inspect.signature(fx_model.__call__).parameters.keys()
# prepare inputs
pt_inputs = self._prepare_for_class(inputs_dict, model_class)
# remove function args that don't exist in Flax
pt_inputs = {k: v for k, v in pt_inputs.items() if k in fx_input_keys}
# send pytorch inputs to the correct device
pt_inputs = {
k: v.to(device=torch_device) if isinstance(v, torch.Tensor) else v for k, v in pt_inputs.items()
}
# convert inputs to Flax
fx_inputs = {k: np.array(v.to("cpu")) for k, v in pt_inputs.items() if torch.is_tensor(v)}
fx_state = convert_pytorch_state_dict_to_flax(pt_model.state_dict(), fx_model)
fx_model.params = fx_state
# send pytorch model to the correct device
pt_model.to(torch_device)
with torch.no_grad():
pt_outputs = pt_model(**pt_inputs)
fx_outputs = fx_model(**fx_inputs)
fx_keys = tuple([k for k, v in fx_outputs.items() if v is not None])
pt_keys = tuple([k for k, v in pt_outputs.items() if v is not None])
self.assertEqual(fx_keys, pt_keys)
self.check_pt_flax_outputs(fx_outputs, pt_outputs, model_class)
with tempfile.TemporaryDirectory() as tmpdirname:
pt_model.save_pretrained(tmpdirname)
fx_model_loaded = fx_model_class.from_pretrained(tmpdirname, from_pt=True)
fx_outputs_loaded = fx_model_loaded(**fx_inputs)
fx_keys = tuple([k for k, v in fx_outputs_loaded.items() if v is not None])
pt_keys = tuple([k for k, v in pt_outputs.items() if v is not None])
self.assertEqual(fx_keys, pt_keys)
self.check_pt_flax_outputs(fx_outputs_loaded, pt_outputs, model_class)
@is_pt_flax_cross_test
def test_equivalence_flax_to_pt(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__):
fx_model_class_name = "Flax" + model_class.__name__
if not hasattr(transformers, fx_model_class_name):
self.skipTest(reason="No Flax model exists for this class")
# Output all for aggressive testing
config.output_hidden_states = True
config.output_attentions = self.has_attentions
fx_model_class = getattr(transformers, fx_model_class_name)
# load PyTorch class
pt_model = model_class(config).eval()
# Flax models don't use the `use_cache` option and cache is not returned as a default.
# So we disable `use_cache` here for PyTorch model.
pt_model.config.use_cache = False
# load Flax class
fx_model = fx_model_class(config, dtype=jnp.float32)
# make sure only flax inputs are forward that actually exist in function args
fx_input_keys = inspect.signature(fx_model.__call__).parameters.keys()
# prepare inputs
pt_inputs = self._prepare_for_class(inputs_dict, model_class)
# remove function args that don't exist in Flax
pt_inputs = {k: v for k, v in pt_inputs.items() if k in fx_input_keys}
# send pytorch inputs to the correct device
pt_inputs = {
k: v.to(device=torch_device) if isinstance(v, torch.Tensor) else v for k, v in pt_inputs.items()
}
# convert inputs to Flax
fx_inputs = {k: np.array(v.to("cpu")) for k, v in pt_inputs.items() if torch.is_tensor(v)}
pt_model = load_flax_weights_in_pytorch_model(pt_model, fx_model.params)
# make sure weights are tied in PyTorch
pt_model.tie_weights()
# send pytorch model to the correct device
pt_model.to(torch_device)
with torch.no_grad():
pt_outputs = pt_model(**pt_inputs)
fx_outputs = fx_model(**fx_inputs)
fx_keys = tuple([k for k, v in fx_outputs.items() if v is not None])
pt_keys = tuple([k for k, v in pt_outputs.items() if v is not None])
self.assertEqual(fx_keys, pt_keys)
self.check_pt_flax_outputs(fx_outputs, pt_outputs, model_class)
with tempfile.TemporaryDirectory() as tmpdirname:
fx_model.save_pretrained(tmpdirname)
pt_model_loaded = model_class.from_pretrained(
tmpdirname, from_flax=True, attn_implementation=fx_model.config._attn_implementation
)
# send pytorch model to the correct device
pt_model_loaded.to(torch_device)
pt_model_loaded.eval()
with torch.no_grad():
pt_outputs_loaded = pt_model_loaded(**pt_inputs)
fx_keys = tuple([k for k, v in fx_outputs.items() if v is not None])
pt_keys = tuple([k for k, v in pt_outputs_loaded.items() if v is not None])
self.assertEqual(fx_keys, pt_keys)
self.check_pt_flax_outputs(fx_outputs, pt_outputs_loaded, model_class)
def test_inputs_embeds(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
model.to(torch_device)
model.eval()
model_forward_args = inspect.signature(model.forward).parameters
if "inputs_embeds" not in model_forward_args:
self.skipTest(reason="This model doesn't use `inputs_embeds`")
inputs = copy.deepcopy(self._prepare_for_class(inputs_dict, model_class))
if not self.is_encoder_decoder:
input_ids = inputs["input_ids"]
del inputs["input_ids"]
else:
encoder_input_ids = inputs["input_ids"]
decoder_input_ids = inputs.get("decoder_input_ids", encoder_input_ids)
del inputs["input_ids"]
inputs.pop("decoder_input_ids", None)
wte = model.get_input_embeddings()
if not self.is_encoder_decoder:
inputs["inputs_embeds"] = wte(input_ids)
else:
inputs["inputs_embeds"] = wte(encoder_input_ids)
inputs["decoder_inputs_embeds"] = wte(decoder_input_ids)
with torch.no_grad():
model(**inputs)[0]
def test_inputs_embeds_matches_input_ids(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class.__name__ not in get_values(MODEL_MAPPING_NAMES):
continue
model = model_class(config)
model.to(torch_device)
model.eval()
model_forward_args = inspect.signature(model.forward).parameters
if "inputs_embeds" not in model_forward_args:
self.skipTest(reason="This model doesn't use `inputs_embeds`")
inputs = copy.deepcopy(self._prepare_for_class(inputs_dict, model_class))
pad_token_id = config.pad_token_id if config.pad_token_id is not None else 1
wte = model.get_input_embeddings()
if not self.is_encoder_decoder:
input_ids = inputs["input_ids"]
# some models infer position ids/attn mask differently when input ids
# by check if pad_token let's make sure no padding is in input ids
not_pad_token_id = pad_token_id + 1 if max(0, pad_token_id - 1) == 0 else pad_token_id - 1
input_ids[input_ids == pad_token_id] = not_pad_token_id
del inputs["input_ids"]
inputs_embeds = wte(input_ids)
with torch.no_grad():
out_ids = model(input_ids=input_ids, **inputs)[0]
out_embeds = model(inputs_embeds=inputs_embeds, **inputs)[0]
else:
encoder_input_ids = inputs["input_ids"]
decoder_input_ids = inputs.get("decoder_input_ids", encoder_input_ids)
encoder_input_ids[encoder_input_ids == pad_token_id] = max(0, pad_token_id + 1)
decoder_input_ids[decoder_input_ids == pad_token_id] = max(0, pad_token_id + 1)
del inputs["input_ids"]
inputs.pop("decoder_input_ids", None)
inputs_embeds = wte(encoder_input_ids)
decoder_inputs_embeds = wte(decoder_input_ids)
with torch.no_grad():
out_ids = model(input_ids=encoder_input_ids, decoder_input_ids=decoder_input_ids, **inputs)[0]
out_embeds = model(
inputs_embeds=inputs_embeds, decoder_inputs_embeds=decoder_inputs_embeds, **inputs
)[0]
torch.testing.assert_close(out_embeds, out_ids)
@require_non_xpu
@require_torch_multi_gpu
def test_multi_gpu_data_parallel_forward(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
# some params shouldn't be scattered by nn.DataParallel
# so just remove them if they are present.
blacklist_non_batched_params = ["head_mask", "decoder_head_mask", "cross_attn_head_mask"]
for k in blacklist_non_batched_params:
inputs_dict.pop(k, None)
# move input tensors to cuda:O
for k, v in inputs_dict.items():
if torch.is_tensor(v):
inputs_dict[k] = v.to(0)
for model_class in self.all_model_classes:
model = model_class(config=config)
model.to(0)
model.eval()
# Wrap model in nn.DataParallel
model = nn.DataParallel(model)
with torch.no_grad():
_ = model(**self._prepare_for_class(inputs_dict, model_class))
@require_torch_gpu
@require_torch_multi_gpu
def test_model_parallelization(self):
if not self.test_model_parallel:
self.skipTest(reason="test_model_parallel is set to False")
# a candidate for testing_utils
def get_current_gpu_memory_use():
"""returns a list of cuda memory allocations per GPU in MBs"""
per_device_memory = []
for id in range(torch.cuda.device_count()):
with torch.cuda.device(id):
per_device_memory.append(torch.cuda.memory_allocated() >> 20)
return per_device_memory
# Needs a large model to see the difference.
config = self.model_tester.get_large_model_config()
for model_class in self.all_parallelizable_model_classes:
torch.cuda.empty_cache()
# 1. single gpu memory load + unload + memory measurements
# Retrieve initial memory usage (can easily be ~0.6-1.5GB if cuda-kernels have been preloaded by previous tests)
memory_at_start = get_current_gpu_memory_use()
# Put model on device 0 and take a memory snapshot
model = model_class(config)
model.to("cuda:0")
memory_after_model_load = get_current_gpu_memory_use()
# The memory use on device 0 should be higher than it was initially.
self.assertGreater(memory_after_model_load[0], memory_at_start[0])
del model
gc.collect()
torch.cuda.empty_cache()
# 2. MP test
# it's essential to re-calibrate the usage before the next stage
memory_at_start = get_current_gpu_memory_use()
# Spread model layers over multiple devices
model = model_class(config)
model.parallelize()
memory_after_parallelization = get_current_gpu_memory_use()
# Assert that the memory use on all devices is higher than it was when loaded only on CPU
for n in range(len(model.device_map.keys())):
self.assertGreater(memory_after_parallelization[n], memory_at_start[n])
# Assert that the memory use of device 0 is lower than it was when the entire model was loaded on it
self.assertLess(memory_after_parallelization[0], memory_after_model_load[0])
# Assert that the memory use of device 1 is higher than it was when the entire model was loaded
# on device 0 and device 1 wasn't used at all
self.assertGreater(memory_after_parallelization[1], memory_after_model_load[1])
del model
gc.collect()
torch.cuda.empty_cache()
@require_torch_gpu
@require_torch_multi_gpu
def test_model_parallel_equal_results(self):
if not self.test_model_parallel:
self.skipTest(reason="test_model_parallel is set to False")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_parallelizable_model_classes:
inputs_dict = self._prepare_for_class(inputs_dict, model_class)
def cast_to_device(dictionary, device):
output = {}
for k, v in dictionary.items():
if isinstance(v, torch.Tensor):
output[k] = v.to(device)
else:
output[k] = v
return output
model = model_class(config)
output = model(**cast_to_device(inputs_dict, "cpu"))
model.parallelize()
parallel_output = model(**cast_to_device(inputs_dict, "cuda:0"))
for value, parallel_value in zip(output, parallel_output):
if isinstance(value, torch.Tensor):
torch.testing.assert_close(value, parallel_value.to("cpu"), rtol=1e-7, atol=1e-7)
elif isinstance(value, (Tuple, List)):
for value_, parallel_value_ in zip(value, parallel_value):
torch.testing.assert_close(value_, parallel_value_.to("cpu"), rtol=1e-7, atol=1e-7)
def check_device_map_is_respected(self, model, device_map):
for param_name, param in model.named_parameters():
# Find device in device_map
while len(param_name) > 0 and param_name not in device_map:
param_name = ".".join(param_name.split(".")[:-1])
if param_name not in device_map:
raise ValueError("device map is incomplete, it does not contain any device for `param_name`.")
param_device = device_map[param_name]
if param_device in ["cpu", "disk"]:
self.assertEqual(param.device, torch.device("meta"))
elif param_device in ["mps"]:
self.assertEqual(param.device, torch.device("mps"))
else:
# when loaded with device_map, `param_device` are integer values for cuda/xpu/npu/mlu
self.assertEqual(param.device, torch.device(f"{torch_device}:{param_device}"))
@require_accelerate
@mark.accelerate_tests
@require_torch_accelerator
def test_disk_offload_bin(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class._no_split_modules is None:
continue
inputs_dict_class = self._prepare_for_class(inputs_dict, model_class)
model = model_class(config).eval()
model = model.to(torch_device)
torch.manual_seed(0)
base_output = model(**inputs_dict_class)
model_size = compute_module_sizes(model)[""]
with tempfile.TemporaryDirectory() as tmp_dir:
model.cpu().save_pretrained(tmp_dir, safe_serialization=False)
with self.assertRaises(ValueError):
max_size = int(self.model_split_percents[0] * model_size)
max_memory = {0: max_size, "cpu": max_size}
# This errors out cause it's missing an offload folder
new_model = model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
max_size = int(self.model_split_percents[1] * model_size)
max_memory = {0: max_size, "cpu": max_size}
new_model = model_class.from_pretrained(
tmp_dir, device_map="auto", max_memory=max_memory, offload_folder=tmp_dir
)
self.check_device_map_is_respected(new_model, new_model.hf_device_map)
torch.manual_seed(0)
new_output = new_model(**inputs_dict_class)
if isinstance(base_output[0], tuple) and isinstance(new_output[0], tuple):
[
torch.testing.assert_close(a, b, rtol=1e-5, atol=1e-5)
for a, b in zip(base_output[0], new_output[0])
]
else:
torch.testing.assert_close(base_output[0], new_output[0], rtol=1e-5, atol=1e-5)
@require_accelerate
@mark.accelerate_tests
@require_torch_accelerator
def test_disk_offload_safetensors(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class._no_split_modules is None:
continue
inputs_dict_class = self._prepare_for_class(inputs_dict, model_class)
model = model_class(config).eval()
model = model.to(torch_device)
torch.manual_seed(0)
base_output = model(**inputs_dict_class)
model_size = compute_module_sizes(model)[""]
with tempfile.TemporaryDirectory() as tmp_dir:
model.cpu().save_pretrained(tmp_dir)
max_size = int(self.model_split_percents[1] * model_size)
max_memory = {0: max_size, "cpu": max_size}
# This doesn't error out as it's in safetensors and doesn't need an offload folder
new_model = model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
self.check_device_map_is_respected(new_model, new_model.hf_device_map)
torch.manual_seed(0)
new_output = new_model(**inputs_dict_class)
if isinstance(base_output[0], tuple) and isinstance(new_output[0], tuple):
[
torch.testing.assert_close(a, b, rtol=1e-5, atol=1e-5)
for a, b in zip(base_output[0], new_output[0])
]
else:
torch.testing.assert_close(base_output[0], new_output[0], rtol=1e-5, atol=1e-5)
@require_accelerate
@mark.accelerate_tests
@require_torch_accelerator
def test_cpu_offload(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class._no_split_modules is None:
continue
inputs_dict_class = self._prepare_for_class(inputs_dict, model_class)
model = model_class(config).eval()
model = model.to(torch_device)
torch.manual_seed(0)
base_output = model(**inputs_dict_class)
model_size = compute_module_sizes(model)[""]
# We test several splits of sizes to make sure it works.
max_gpu_sizes = [int(p * model_size) for p in self.model_split_percents[1:]]
with tempfile.TemporaryDirectory() as tmp_dir:
model.cpu().save_pretrained(tmp_dir)
for max_size in max_gpu_sizes:
max_memory = {0: max_size, "cpu": model_size * 2}
new_model = model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
# Making sure part of the model will actually end up offloaded
self.assertSetEqual(set(new_model.hf_device_map.values()), {0, "cpu"})
self.check_device_map_is_respected(new_model, new_model.hf_device_map)
torch.manual_seed(0)
new_output = new_model(**inputs_dict_class)
if isinstance(base_output[0], tuple) and isinstance(new_output[0], tuple):
[
torch.testing.assert_close(a, b, rtol=1e-5, atol=1e-5)
for a, b in zip(base_output[0], new_output[0])
]
else:
torch.testing.assert_close(base_output[0], new_output[0], rtol=1e-5, atol=1e-5)
@require_accelerate
@mark.accelerate_tests
@require_torch_multi_accelerator
def test_model_parallelism(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class._no_split_modules is None:
continue
inputs_dict_class = self._prepare_for_class(inputs_dict, model_class)
model = model_class(config).eval()
model = model.to(torch_device)
torch.manual_seed(0)
base_output = model(**inputs_dict_class)
model_size = compute_module_sizes(model)[""]
# We test several splits of sizes to make sure it works.
max_gpu_sizes = [int(p * model_size) for p in self.model_split_percents[1:]]
with tempfile.TemporaryDirectory() as tmp_dir:
model.cpu().save_pretrained(tmp_dir)
for max_size in max_gpu_sizes:
max_memory = {0: max_size, 1: model_size * 2, "cpu": model_size * 2}
new_model = model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
# Making sure part of the model will actually end up offloaded
self.assertSetEqual(set(new_model.hf_device_map.values()), {0, 1})
self.check_device_map_is_respected(new_model, new_model.hf_device_map)
torch.manual_seed(0)
new_output = new_model(**inputs_dict_class)
if isinstance(base_output[0], tuple) and isinstance(new_output[0], tuple):
[
torch.testing.assert_close(a, b, rtol=1e-5, atol=1e-5)
for a, b in zip(base_output[0], new_output[0])
]
else:
torch.testing.assert_close(base_output[0], new_output[0], rtol=1e-5, atol=1e-5)
def test_problem_types(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
problem_types = [
{"title": "multi_label_classification", "num_labels": 2, "dtype": torch.float},
{"title": "single_label_classification", "num_labels": 1, "dtype": torch.long},
{"title": "regression", "num_labels": 1, "dtype": torch.float},
]
for model_class in self.all_model_classes:
if model_class.__name__ not in [
*get_values(MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES),
*get_values(MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES),
]:
continue
for problem_type in problem_types:
with self.subTest(msg=f"Testing {model_class} with {problem_type['title']}"):
config.problem_type = problem_type["title"]
config.num_labels = problem_type["num_labels"]
model = model_class(config)
model.to(torch_device)
model.train()
inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
if problem_type["num_labels"] > 1:
inputs["labels"] = inputs["labels"].unsqueeze(1).repeat(1, problem_type["num_labels"])
inputs["labels"] = inputs["labels"].to(problem_type["dtype"])
# This tests that we do not trigger the warning form PyTorch "Using a target size that is different
# to the input size. This will likely lead to incorrect results due to broadcasting. Please ensure
# they have the same size." which is a symptom something in wrong for the regression problem.
# See https://github.com/huggingface/transformers/issues/11780
with warnings.catch_warnings(record=True) as warning_list:
loss = model(**inputs).loss
for w in warning_list:
if "Using a target size that is different to the input size" in str(w.message):
raise ValueError(
f"Something is going wrong in the regression problem: intercepted {w.message}"
)
loss.backward()
def test_load_with_mismatched_shapes(self):
if not self.test_mismatched_shapes:
self.skipTest(reason="test_missmatched_shapes is set to False")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
if model_class.__name__ not in get_values(MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES):
continue
with self.subTest(msg=f"Testing {model_class}"):
with tempfile.TemporaryDirectory() as tmp_dir:
model = model_class(config)
model.save_pretrained(tmp_dir)
# Fails when we don't set ignore_mismatched_sizes=True
with self.assertRaises(RuntimeError):
new_model = AutoModelForSequenceClassification.from_pretrained(tmp_dir, num_labels=42)
with self.assertRaises(RuntimeError):
new_model_without_prefix = AutoModel.from_pretrained(tmp_dir, vocab_size=10)
logger = logging.get_logger("transformers.modeling_utils")
with CaptureLogger(logger) as cl:
new_model = AutoModelForSequenceClassification.from_pretrained(
tmp_dir, num_labels=42, ignore_mismatched_sizes=True
)
self.assertIn("the shapes did not match", cl.out)
new_model.to(torch_device)
inputs = self._prepare_for_class(inputs_dict, model_class)
logits = new_model(**inputs).logits
self.assertEqual(logits.shape[1], 42)
with CaptureLogger(logger) as cl:
new_model_without_prefix = AutoModel.from_pretrained(
tmp_dir, vocab_size=10, ignore_mismatched_sizes=True
)
self.assertIn("the shapes did not match", cl.out)
input_ids = ids_tensor((2, 8), 10)
new_model_without_prefix.to(torch_device)
if self.is_encoder_decoder:
new_model_without_prefix(input_ids, decoder_input_ids=input_ids)
else:
new_model_without_prefix(input_ids)
def test_mismatched_shapes_have_properly_initialized_weights(self):
if not self.test_mismatched_shapes:
self.skipTest(reason="test_missmatched_shapes is set to False")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
configs_no_init = _config_zero_init(config)
for model_class in self.all_model_classes:
mappings = [
MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES,
MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES,
]
is_classication_model = any(model_class.__name__ in get_values(mapping) for mapping in mappings)
if not is_classication_model:
continue
# TODO: ydshieh
is_special_classes = model_class.__name__ in [
"wav2vec2.masked_spec_embed",
"Wav2Vec2ForSequenceClassification",
"CLIPForImageClassification",
"RegNetForImageClassification",
"ResNetForImageClassification",
"UniSpeechSatForSequenceClassification",
"Wav2Vec2BertForSequenceClassification",
"PvtV2ForImageClassification",
"Wav2Vec2ConformerForSequenceClassification",
"WavLMForSequenceClassification",
"SwiftFormerForImageClassification",
"SEWForSequenceClassification",
"BitForImageClassification",
"SEWDForSequenceClassification",
"SiglipForImageClassification",
"HubertForSequenceClassification",
"Swinv2ForImageClassification",
"Data2VecAudioForSequenceClassification",
"UniSpeechForSequenceClassification",
"PvtForImageClassification",
"ModernBertForSequenceClassification",
"ModernBertForTokenClassification",
"TimmWrapperForImageClassification",
]
special_param_names = [
r"^bit\.",
r"^classifier\.weight",
r"^classifier\.bias",
r"^classifier\..+\.weight",
r"^classifier\..+\.bias",
r"^data2vec_audio\.",
r"^dist_head\.",
r"^head\.",
r"^hubert\.",
r"^pvt\.",
r"^pvt_v2\.",
r"^regnet\.",
r"^resnet\.",
r"^sew\.",
r"^sew_d\.",
r"^swiftformer\.",
r"^swinv2\.",
r"^transformers\.models\.swiftformer\.",
r"^timm_model\.",
r"^unispeech\.",
r"^unispeech_sat\.",
r"^vision_model\.",
r"^wav2vec2\.",
r"^wav2vec2_bert\.",
r"^wav2vec2_conformer\.",
r"^wavlm\.",
]
with self.subTest(msg=f"Testing {model_class}"):
with tempfile.TemporaryDirectory() as tmp_dir:
model = model_class(configs_no_init)
model.save_pretrained(tmp_dir)
# Fails when we don't set ignore_mismatched_sizes=True
with self.assertRaises(RuntimeError):
new_model = model_class.from_pretrained(tmp_dir, num_labels=42)
logger = logging.get_logger("transformers.modeling_utils")
with CaptureLogger(logger) as cl:
new_model = model_class.from_pretrained(tmp_dir, num_labels=42, ignore_mismatched_sizes=True)
self.assertIn("the shapes did not match", cl.out)
for name, param in new_model.named_parameters():
if param.requires_grad:
param_mean = ((param.data.mean() * 1e9).round() / 1e9).item()
if not (
is_special_classes
and any(len(re.findall(target, name)) > 0 for target in special_param_names)
):
self.assertIn(
param_mean,
[0.0, 1.0],
msg=f"Parameter {name} of model {model_class} seems not properly initialized",
)
else:
# Here we allow the parameters' mean to be in the range [-5.0, 5.0] instead of being
# either `0.0` or `1.0`, because their initializations are not using
# `config.initializer_factor` (or something similar). The purpose of this test is simply
# to make sure they are properly initialized (to avoid very large value or even `nan`).
self.assertGreaterEqual(
param_mean,
-5.0,
msg=f"Parameter {name} of model {model_class} seems not properly initialized",
)
self.assertLessEqual(
param_mean,
5.0,
msg=f"Parameter {name} of model {model_class} seems not properly initialized",
)
def test_matched_shapes_have_loaded_weights_when_some_mismatched_shapes_exist(self):
# 1. Create a dummy class. Should have buffers as well? To make sure we test __init__
class MyClass(PreTrainedModel):
config_class = PretrainedConfig
def __init__(self, config=None):
super().__init__(config if config is not None else PretrainedConfig())
self.linear = nn.Linear(10, config.num_labels, bias=True)
self.embedding = nn.Embedding(10, 10)
self.std = 1
def _init_weights(self, module):
if isinstance(module, nn.Linear):
module.weight.data = nn.init.kaiming_uniform_(module.weight.data, np.sqrt(5))
if module.bias is not None:
module.bias.data = module.bias.data.normal_(mean=0.0, std=self.std)
# Used to make sure the weights with matched shape are loaded correctly
config = PretrainedConfig()
config.num_labels = 3
model = MyClass(config=config)
# Used to make sure the weights with mismatched shape are properly initialized
set_seed(0)
config = PretrainedConfig()
config.num_labels = 4
# not to init. the weights during the creation: to match the logic in `from_pretrained`, so we can keep the
# same sequence of random ops in the execution path to allow us to compare `target_model` and `new_model` below
# for `linear` part.
with ContextManagers([no_init_weights(True)]):
target_model = MyClass(config=config)
target_model.apply(target_model._initialize_weights)
with tempfile.TemporaryDirectory() as tmpdirname:
state_dict = model.state_dict()
del state_dict["linear.weight"]
model.config.save_pretrained(tmpdirname)
torch.save(state_dict, os.path.join(tmpdirname, "pytorch_model.bin"))
set_seed(0)
new_model = MyClass.from_pretrained(tmpdirname, num_labels=4, ignore_mismatched_sizes=True)
for key in new_model.state_dict().keys():
# check weight values for weights with matched shapes are identical
# (i.e. correctly loaded from the checkpoint)
if key not in ["linear.weight", "linear.bias"]:
max_diff = torch.max(torch.abs(model.state_dict()[key] - new_model.state_dict()[key]))
self.assertLessEqual(
max_diff.item(),
1e-6,
msg=f"the weight values for `{key}` in `new_model` and `model` are not identical",
)
else:
# check we have some mismatched shapes
self.assertNotEqual(
model.state_dict()[key].shape,
new_model.state_dict()[key].shape,
msg=f"the weight shapes for {key} in `model` and `new_model` should differ",
)
# check the weights with mismatched shape are properly initialized
max_diff = torch.max(torch.abs(new_model.state_dict()[key] - target_model.state_dict()[key]))
self.assertLessEqual(
max_diff.item(),
1e-6,
msg=f"the weight values for `{key}` in `new_model` and `target_model` are not identical",
)
def test_model_is_small(self):
# Just a consistency check to make sure we are not running tests on 80M parameter models.
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
num_params = model.num_parameters()
assert (
num_params < 1000000
), f"{model_class} is too big for the common tests ({num_params})! It should have 1M max."
@require_flash_attn
@require_torch_gpu
@mark.flash_attn_test
@slow
@is_flaky()
def test_flash_attn_2_inference_equivalence(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
for model_class in self.all_model_classes:
if not model_class._supports_flash_attn_2:
self.skipTest(f"{model_class.__name__} does not support Flash Attention 2")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model_fa = model_class.from_pretrained(
tmpdirname, torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2"
)
model_fa.to(torch_device)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.bfloat16)
model.to(torch_device)
dummy_input = inputs_dict[model.main_input_name][:1]
if dummy_input.dtype in [torch.float32, torch.float16]:
dummy_input = dummy_input.to(torch.bfloat16)
dummy_attention_mask = inputs_dict.get("attention_mask", None)
if dummy_attention_mask is not None:
dummy_attention_mask = dummy_attention_mask[:1]
dummy_attention_mask[:, 1:] = 1
dummy_attention_mask[:, :1] = 0
if model.config.is_encoder_decoder:
decoder_input_ids = inputs_dict.get("decoder_input_ids", dummy_input)[:1]
outputs = model(dummy_input, decoder_input_ids=decoder_input_ids, output_hidden_states=True)
outputs_fa = model_fa(dummy_input, decoder_input_ids=decoder_input_ids, output_hidden_states=True)
else:
outputs = model(dummy_input, output_hidden_states=True)
outputs_fa = model_fa(dummy_input, output_hidden_states=True)
logits = (
outputs.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs.decoder_hidden_states[-1]
)
logits_fa = (
outputs_fa.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs_fa.decoder_hidden_states[-1]
)
assert torch.allclose(logits_fa, logits, atol=4e-2, rtol=4e-2)
if model.config.is_encoder_decoder:
other_inputs = {
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": dummy_attention_mask,
"output_hidden_states": True,
}
if dummy_attention_mask is not None:
other_inputs["attention_mask"] = dummy_attention_mask
outputs = model(dummy_input, **other_inputs)
outputs_fa = model_fa(dummy_input, **other_inputs)
else:
other_inputs = {
"output_hidden_states": True,
}
if dummy_attention_mask is not None:
other_inputs["attention_mask"] = dummy_attention_mask
outputs = model(dummy_input, **other_inputs)
outputs_fa = model_fa(dummy_input, **other_inputs)
logits = (
outputs.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs.decoder_hidden_states[-1]
)
logits_fa = (
outputs_fa.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs_fa.decoder_hidden_states[-1]
)
assert torch.allclose(logits_fa[1:], logits[1:], atol=4e-2, rtol=4e-2)
# check with inference + dropout
model.train()
_ = model_fa(dummy_input, **other_inputs)
@require_flash_attn
@require_torch_gpu
@mark.flash_attn_test
@slow
@is_flaky()
def test_flash_attn_2_inference_equivalence_right_padding(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
for model_class in self.all_model_classes:
if not model_class._supports_flash_attn_2:
self.skipTest(f"{model_class.__name__} does not support Flash Attention 2")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model_fa = model_class.from_pretrained(
tmpdirname, torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2"
)
model_fa.to(torch_device)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.bfloat16)
model.to(torch_device)
dummy_input = inputs_dict[model.main_input_name][:1]
if dummy_input.dtype in [torch.float32, torch.float16]:
dummy_input = dummy_input.to(torch.bfloat16)
dummy_attention_mask = inputs_dict.get("attention_mask", None)
if dummy_attention_mask is not None:
dummy_attention_mask = dummy_attention_mask[:1]
dummy_attention_mask[:, :-1] = 1
dummy_attention_mask[:, -1:] = 0
if model.config.is_encoder_decoder:
decoder_input_ids = inputs_dict.get("decoder_input_ids", dummy_input)[:1]
outputs = model(dummy_input, decoder_input_ids=decoder_input_ids, output_hidden_states=True)
outputs_fa = model_fa(dummy_input, decoder_input_ids=decoder_input_ids, output_hidden_states=True)
else:
outputs = model(dummy_input, output_hidden_states=True)
outputs_fa = model_fa(dummy_input, output_hidden_states=True)
logits = (
outputs.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs.decoder_hidden_states[-1]
)
logits_fa = (
outputs_fa.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs_fa.decoder_hidden_states[-1]
)
assert torch.allclose(logits_fa, logits, atol=4e-2, rtol=4e-2)
if model.config.is_encoder_decoder:
other_inputs = {
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": dummy_attention_mask,
"output_hidden_states": True,
}
if dummy_attention_mask is not None:
other_inputs["attention_mask"] = dummy_attention_mask
outputs = model(dummy_input, **other_inputs)
outputs_fa = model_fa(dummy_input, **other_inputs)
else:
other_inputs = {
"output_hidden_states": True,
}
if dummy_attention_mask is not None:
other_inputs["attention_mask"] = dummy_attention_mask
outputs = model(dummy_input, **other_inputs)
outputs_fa = model_fa(dummy_input, **other_inputs)
logits = (
outputs.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs.decoder_hidden_states[-1]
)
logits_fa = (
outputs_fa.hidden_states[-1]
if not model.config.is_encoder_decoder
else outputs_fa.decoder_hidden_states[-1]
)
assert torch.allclose(logits_fa[:-1], logits[:-1], atol=4e-2, rtol=4e-2)
def test_attn_implementation_composite_models(self):
"""
Tests if composite models can receive a dict object as attn_implementation, where each key should be
one of the sub-configs from the model's config.
"""
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
for model_class in self.all_model_classes:
if not self._is_composite:
self.skipTest("Model is not a composite model.")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
# set eager as it will be the one supported in all models
# we just need to test if passing 'attn_implementation' as a dict fails or not
attn_implementation_per_subconfig = {}
for key in config.sub_configs.keys():
attn_implementation_per_subconfig[key] = "eager"
config._attn_implementation = attn_implementation_per_subconfig
model = model_class(config)
for key in config.sub_configs.keys():
sub_config = getattr(model.config, key)
self.assertTrue(sub_config._attn_implementation == "eager")
for name, submodule in model.named_modules():
class_name = submodule.__class__.__name__
if (
class_name.endswith("Attention")
and getattr(submodule, "config", None)
and submodule.config._attn_implementation != "eager"
):
raise ValueError(
f"The eager model should not have SDPA/FA2 attention layers but got `{class_name}.config._attn_implementation={submodule.config._attn_implementation}`"
)
@require_torch_sdpa
def test_sdpa_can_dispatch_non_composite_models(self):
"""
Tests if non-composite models dispatch correctly on SDPA/eager when requested so when loading the model.
This tests only by looking at layer names, as usually SDPA layers are calles "SDPAAttention".
"""
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
if not self.all_model_classes[0]._supports_sdpa or self._is_composite:
self.skipTest(f"{self.all_model_classes[0].__name__} does not support SDPA")
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model_sdpa = model_class.from_pretrained(tmpdirname)
model_sdpa = model_sdpa.eval().to(torch_device)
self.assertTrue(model_sdpa.config._attn_implementation == "sdpa")
model_eager = model_class.from_pretrained(tmpdirname, attn_implementation="eager")
model_eager = model_eager.eval().to(torch_device)
self.assertTrue(model_eager.config._attn_implementation == "eager")
for name, submodule in model_eager.named_modules():
class_name = submodule.__class__.__name__
if (
class_name.endswith("Attention")
and getattr(submodule, "config", None)
and submodule.config._attn_implementation == "sdpa"
):
raise ValueError(
f"The eager model should not have SDPA attention layers but got `{class_name}.config._attn_implementation={submodule.config._attn_implementation}`"
)
@require_torch_sdpa
def test_sdpa_can_dispatch_composite_models(self):
"""
Tests if composite models dispatch correctly on SDPA/eager when requested so when loading the model.
This tests only by looking at layer names, as usually SDPA layers are calles "SDPAAttention".
In contrast to the above test, this one checks if the "config._attn_implamentation" is a dict after the model
is loaded, because we manually replicate requested attn implementation on each sub-config when loading.
See https://github.com/huggingface/transformers/pull/32238 for more info
The test tries to cover most general cases of composite models, VLMs with vision and text configs. Any model
that has a different set of sub-configs has to overwrite this test.
"""
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
if not self._is_composite:
self.skipTest(f"{self.all_model_classes[0].__name__} does not support SDPA")
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model_sdpa = model_class.from_pretrained(tmpdirname)
model_sdpa = model_sdpa.eval().to(torch_device)
vision_model_names = {"visual", "image_tower", "vision_tower", "vision_model"}
language_model_names = {"language_model", "model", "text_model"}
vision_model_name = [name for name in vision_model_names if hasattr(model_sdpa, name)][0]
language_model_name = [name for name in language_model_names if hasattr(model_sdpa, name)][0]
vision_model_sdpa = getattr(model, vision_model_name)
language_model_sdpa = getattr(model, language_model_name)
text_attn = "sdpa" if language_model_sdpa._supports_sdpa else "eager"
vision_attn = "sdpa" if vision_model_sdpa._supports_sdpa else "eager"
# `None` as it is the requested one which will be assigned to each sub-config
# Sub-model will dispatch to SDPA if it can (checked below that `SDPA` layers are present)
self.assertTrue(language_model_sdpa.config._attn_implementation == text_attn)
self.assertTrue(vision_model_sdpa.config._attn_implementation == vision_attn)
model_eager = model_class.from_pretrained(tmpdirname, attn_implementation="eager")
model_eager = model_eager.eval().to(torch_device)
self.assertTrue(getattr(model_eager, language_model_name).config._attn_implementation == "eager")
self.assertTrue(getattr(model_eager, vision_model_name).config._attn_implementation == "eager")
for name, submodule in model_eager.named_modules():
class_name = submodule.__class__.__name__
if (
class_name.endswith("Attention")
and getattr(submodule, "config", None)
and submodule.config._attn_implementation == "sdpa"
):
raise ValueError("The eager model should not have SDPA attention layers")
@parameterized.expand([("float16",), ("bfloat16",), ("float32",)])
@require_torch_sdpa
def test_eager_matches_sdpa_inference(self, torch_dtype: str):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
if not self.all_model_classes[0]._supports_sdpa:
self.skipTest(f"{self.all_model_classes[0].__name__} does not support SDPA")
if torch_dtype == "float16" and not is_torch_fp16_available_on_device(torch_device):
self.skipTest(f"float16 not supported on {torch_device} (on the specific device currently used)")
if torch_dtype == "bfloat16" and not is_torch_bf16_available_on_device(torch_device):
self.skipTest(
f"bfloat16 not supported on {torch_device} (on the specific device currently used, e.g. Nvidia T4 GPU)"
)
# Not sure whether it's fine to put torch.XXX in a decorator if torch is not available so hacking it here instead.
if torch_dtype == "float16":
torch_dtype = torch.float16
elif torch_dtype == "bfloat16":
torch_dtype = torch.bfloat16
elif torch_dtype == "float32":
torch_dtype = torch.float32
atols = {
("cpu", False, torch.float32): 1e-6,
("cpu", False, torch.float16): 5e-3,
("cpu", False, torch.bfloat16): 1e-2,
("cpu", True, torch.float32): 1e-6,
("cpu", True, torch.float16): 5e-3,
("cpu", True, torch.bfloat16): 1e-2,
("cuda", False, torch.float32): 1e-6,
("cuda", False, torch.bfloat16): 1e-2,
("cuda", False, torch.float16): 5e-3,
("cuda", True, torch.float32): 1e-6,
("cuda", True, torch.bfloat16): 1e-2,
("cuda", True, torch.float16): 5e-3,
}
rtols = {
("cpu", False, torch.float32): 1e-4,
("cpu", False, torch.float16): 5e-3,
("cpu", False, torch.bfloat16): 1e-2,
("cpu", True, torch.float32): 1e-4,
("cpu", True, torch.float16): 5e-3,
("cpu", True, torch.bfloat16): 1e-2,
("cuda", False, torch.float32): 1e-4,
("cuda", False, torch.bfloat16): 1e-2,
("cuda", False, torch.float16): 5e-3,
("cuda", True, torch.float32): 1e-4,
("cuda", True, torch.bfloat16): 3e-2,
("cuda", True, torch.float16): 5e-3,
}
def get_mean_reldiff(failcase, x, ref, atol, rtol):
return f"{failcase}: mean relative difference: {((x - ref).abs() / (ref.abs() + 1e-12)).mean():.3e}, torch atol = {atol}, torch rtol = {rtol}"
set_model_tester_for_less_flaky_test(self)
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
set_config_for_less_flaky_test(config)
model = model_class(config)
# FIXME: we deactivate boolean mask for models using "use_mask_token" in their constructors.
# These models support masking only in the case `use_mask_token=True`. Otherwise they cannot consume an input mask.
# This means that the class needs to be instantiated much later, after `use_mask` is set, which means a significant refactor of the code.
# However masking there is not done at any layers that matters (i.e self-attention), therefore we can safely deactivate it.
deactivate_mask = "use_mask_token" in inspect.signature(model_class).parameters
is_encoder_decoder = model.config.is_encoder_decoder
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
try:
model_sdpa = model_class.from_pretrained(
tmpdirname, torch_dtype=torch_dtype, attn_implementation="sdpa"
)
except ValueError:
model_sdpa = model_class.from_pretrained(tmpdirname, torch_dtype=torch_dtype)
model_sdpa = model_sdpa.eval().to(torch_device, dtype=torch_dtype)
model_eager = model_class.from_pretrained(
tmpdirname,
torch_dtype=torch_dtype,
attn_implementation="eager",
)
model_eager = model_eager.eval().to(torch_device, dtype=torch_dtype)
set_model_for_less_flaky_test(model_eager)
set_model_for_less_flaky_test(model_sdpa)
# We use these for loops instead of parameterized.expand just for the interest of avoiding loading/saving 16 times the model,
# but it would be nicer to have an efficient way to use parameterized.expand
fail_cases = []
for padding_side in ["left", "right"]:
for use_mask in [False, True]:
for output_attentions in [True, False]:
can_output_attn = "output_attentions" in inspect.signature(model_sdpa.forward).parameters
if not (self.has_attentions and can_output_attn) and output_attentions:
continue
# TODO: if we can also check with `batch_size=1` without being flaky?
for batch_size in [7]:
dummy_input = inputs_dict[model.main_input_name]
if dummy_input.dtype in [torch.float32, torch.bfloat16, torch.float16]:
dummy_input = dummy_input.to(torch_dtype)
dummy_input = dummy_input[:batch_size]
if dummy_input.shape[0] != batch_size:
if dummy_input.dtype in [torch.float32, torch.bfloat16, torch.float16]:
extension = torch.rand(
batch_size - dummy_input.shape[0],
*dummy_input.shape[1:],
dtype=torch_dtype,
device=torch_device,
)
dummy_input = torch.cat((dummy_input, extension), dim=0).to(torch_device)
else:
extension = torch.randint(
high=5,
size=(batch_size - dummy_input.shape[0], *dummy_input.shape[1:]),
dtype=dummy_input.dtype,
device=torch_device,
)
dummy_input = torch.cat((dummy_input, extension), dim=0).to(torch_device)
if not use_mask:
dummy_attention_mask = None
else:
dummy_attention_mask = inputs_dict.get("attention_mask", None)
if dummy_attention_mask is None:
if is_encoder_decoder:
seqlen = inputs_dict.get("decoder_input_ids", dummy_input).shape[-1]
else:
seqlen = dummy_input.shape[-1]
dummy_attention_mask = (
torch.ones(batch_size, seqlen).to(torch.int64).to(torch_device)
)
dummy_attention_mask = dummy_attention_mask[:batch_size]
if dummy_attention_mask.shape[0] != batch_size:
extension = torch.ones(
batch_size - dummy_attention_mask.shape[0],
*dummy_attention_mask.shape[1:],
dtype=dummy_attention_mask.dtype,
device=torch_device,
)
dummy_attention_mask = torch.cat((dummy_attention_mask, extension), dim=0)
dummy_attention_mask = dummy_attention_mask.to(torch_device)
dummy_attention_mask[:] = 1
if padding_side == "left":
dummy_attention_mask[-1, :2] = 0
dummy_attention_mask[-1, 2:] = 1
elif padding_side == "right":
dummy_attention_mask[-1, -2:] = 0
dummy_attention_mask[-1, :-2] = 1
for enable_kernels in [False, True]:
failcase = f"padding_side={padding_side}, use_mask={use_mask}, enable_kernels={enable_kernels}"
if is_encoder_decoder:
decoder_input_ids = inputs_dict.get("decoder_input_ids", dummy_input)[
:batch_size
]
if decoder_input_ids.shape[0] != batch_size:
extension = torch.ones(
batch_size - decoder_input_ids.shape[0],
*decoder_input_ids.shape[1:],
dtype=decoder_input_ids.dtype,
device=torch_device,
)
decoder_input_ids = torch.cat((decoder_input_ids, extension), dim=0)
decoder_input_ids = decoder_input_ids.to(torch_device)
# TODO: never an `attention_mask` arg here?
processed_inputs = {
model.main_input_name: dummy_input,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": dummy_attention_mask,
"output_hidden_states": True,
}
else:
processed_inputs = {
model.main_input_name: dummy_input,
"output_hidden_states": True,
}
# Otherwise fails for e.g. WhisperEncoderModel
if "attention_mask" in inspect.signature(model_eager.forward).parameters:
processed_inputs["attention_mask"] = dummy_attention_mask
if (
self.has_attentions
and "output_attentions" in inspect.signature(model_sdpa.forward).parameters
):
processed_inputs["output_attentions"] = output_attentions
if not deactivate_mask and (
"bool_masked_pos" in inspect.signature(model_eager.forward).parameters
):
dummy_mask = torch.ones((self.model_tester.num_masks,))
# In case of additional token (like class) we define a custom `mask_length`
if hasattr(self.model_tester, "mask_length"):
mask_length = self.model_tester.mask_length - dummy_mask.size(0)
else:
mask_length = self.model_tester.seq_length - dummy_mask.size(0)
dummy_mask = torch.cat([dummy_mask, torch.zeros(mask_length)])
dummy_bool_masked_pos = dummy_mask.expand(batch_size, -1).bool()
processed_inputs["bool_masked_pos"] = dummy_bool_masked_pos.to(torch_device)
if "noise" in inspect.signature(model_eager.forward).parameters:
np.random.seed(2)
num_patches = int(
(self.model_tester.image_size // self.model_tester.patch_size) ** 2
)
noise = np.random.uniform(size=(batch_size, num_patches))
processed_inputs["noise"] = torch.from_numpy(noise)
# TODO: test gradients as well (& for FA2 as well!)
with torch.no_grad():
with sdpa_kernel(
enable_flash=enable_kernels,
enable_math=True,
enable_mem_efficient=enable_kernels,
):
prepared_inputs = self._prepare_for_class(processed_inputs, model_class)
outputs_eager = model_eager(**prepared_inputs)
outputs_sdpa = model_sdpa(**prepared_inputs)
if hasattr(outputs_eager, "vision_hidden_states"):
logits_eager = outputs_eager.vision_hidden_states[-1]
logits_sdpa = outputs_sdpa.vision_hidden_states[-1]
else:
logits_eager = (
outputs_eager.hidden_states[-1]
if not is_encoder_decoder
else outputs_eager.decoder_hidden_states[-1]
)
logits_sdpa = (
outputs_sdpa.hidden_states[-1]
if not is_encoder_decoder
else outputs_sdpa.decoder_hidden_states[-1]
)
if torch_device in ["cpu", "cuda"]:
atol = atols[torch_device, enable_kernels, torch_dtype]
rtol = rtols[torch_device, enable_kernels, torch_dtype]
elif torch_device == "xpu":
# As of PyTorch 2.5 XPU backend supports only torch.nn.attention.SDPBackend.MATH
# which is implemented on PyTorch level using aten operators and is
# device agnostic with respect to implementation of each aten operator.
atol = atols["cuda", False, torch_dtype]
rtol = rtols["cuda", False, torch_dtype]
else:
atol = 1e-7
rtol = 1e-4
# Masked tokens output slightly deviates - we don't mind that.
if use_mask:
_logits_sdpa = torch.zeros_like(input=logits_sdpa)
_logits_eager = torch.zeros_like(input=logits_eager)
_logits_sdpa[:-1] = logits_sdpa[:-1]
_logits_eager[:-1] = logits_eager[:-1]
if padding_side == "left":
_logits_sdpa[-1:, 2:] = logits_sdpa[-1:, 2:]
_logits_eager[-1:, 2:] = logits_eager[-1:, 2:]
elif padding_side == "right":
_logits_sdpa[-1:, 2:] = logits_sdpa[-1:, :-2]
_logits_eager[-1:, 2:] = logits_eager[-1:, :-2]
logits_sdpa = _logits_sdpa
logits_eager = _logits_eager
results = [
torch.allclose(_logits_sdpa, _logits_eager, atol=atol, rtol=rtol)
for (_logits_sdpa, _logits_eager) in zip(logits_sdpa, logits_eager)
]
# If 80% batch elements have matched results, it's fine
if np.mean(results) < 0.8:
fail_cases.append(
get_mean_reldiff(failcase, logits_sdpa, logits_eager, atol, rtol)
)
self.assertTrue(len(fail_cases) == 0, "\n".join(fail_cases))
@require_torch_sdpa
@require_torch_gpu
@slow
def test_sdpa_can_dispatch_on_flash(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
torch.compiler.reset()
compute_capability = torch.cuda.get_device_capability()
major, _ = compute_capability
if not torch.version.cuda or major < 8:
self.skipTest(reason="This test requires an NVIDIA GPU with compute capability >= 8.0")
for model_class in self.all_model_classes:
if not model_class._supports_sdpa:
self.skipTest(f"{model_class.__name__} does not support SDPA")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
inputs_dict = self._prepare_for_class(inputs_dict, model_class)
if config.model_type in ["llava", "llava_next", "vipllava", "video_llava"]:
self.skipTest(
reason="Llava-like models currently (transformers==4.39.1) requires an attention_mask input"
)
if config.model_type in ["paligemma"]:
self.skipTest(
"PaliGemma-like models currently (transformers==4.41.0) requires an attention_mask input"
)
if config.model_type in ["idefics", "idefics2", "idefics3"]:
self.skipTest(reason="Idefics currently (transformers==4.39.1) requires an image_attention_mask input")
if config.model_type in ["sam"]:
self.skipTest(reason="SAM requires an attention_mask input for relative positional embeddings")
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.float16, attn_implementation="sdpa")
model.to(torch_device)
inputs_dict.pop("attention_mask", None)
inputs_dict.pop("decoder_attention_mask", None)
for name, inp in inputs_dict.items():
if isinstance(inp, torch.Tensor) and inp.dtype in [torch.float32, torch.float16]:
inputs_dict[name] = inp.to(torch.float16)
with sdpa_kernel(enable_flash=True, enable_math=False, enable_mem_efficient=False):
_ = model(**inputs_dict)
@require_non_xpu
@require_torch_sdpa
@require_torch_accelerator
@slow
def test_sdpa_can_compile_dynamic(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
torch.compiler.reset()
if "cuda" in torch_device:
compute_capability = torch.cuda.get_device_capability()
major, _ = compute_capability
if not torch.version.cuda or major < 8:
self.skipTest(reason="This test requires an NVIDIA GPU with compute capability >= 8.0")
for model_class in self.all_model_classes:
if not model_class._supports_sdpa:
self.skipTest(f"{model_class.__name__} does not support SDPA")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
inputs_dict = self._prepare_for_class(inputs_dict, model_class)
if config.model_type in ["dbrx"]:
self.skipTest(
"DBRX (transformers==4.40) requires a modification to support dynamic shapes with compile."
)
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch.float16, attn_implementation="sdpa")
model.to(torch_device)
# For PyTorch 2.1 - 2.3.0 set `dynamic=True`. In the future setting `dynamic=None` and using `torch._dynamo.mark_dynamic()`
# on input tensors will be required. `mark_dynamic` currently raises inconsistent shape errors.
model = torch.compile(model, dynamic=True)
inputs_dict.pop("attention_mask", None)
inputs_dict.pop("decoder_attention_mask", None)
for name, inp in inputs_dict.items():
if isinstance(inp, torch.Tensor) and inp.dtype in [torch.float32, torch.float16]:
inputs_dict[name] = inp.to(torch.float16)
# use no_grad to save some memory
with torch.no_grad():
_ = model(**inputs_dict)
@require_torch_sdpa
def test_sdpa_matches_eager_sliding_window(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
WINDOW_ATTENTION_MODELS = ["mistral", "mixtral", "qwen2", "qwen_moe", "starcoder2"]
if len(self.all_generative_model_classes) == 0:
self.skipTest(f"No generative model classes for {self.__class__.__name__}")
for model_class in self.all_generative_model_classes:
if model_class._supports_sdpa:
self.skipTest(reason="Model architecture does not support attentions")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
if config.model_type not in WINDOW_ATTENTION_MODELS:
self.skipTest(f"{config.model_type} does not use window attention")
config.sliding_window = 2
dummy_input = inputs_dict[model_class.main_input_name]
attention_mask = inputs_dict["attention_mask"]
self.assertTrue(dummy_input.ndim == 2)
self.assertTrue(dummy_input.shape[1] > 6)
with tempfile.TemporaryDirectory() as tmpdir:
with torch.device(torch_device):
model_eager = AutoModelForCausalLM.from_config(
config, attn_implementation="eager", torch_dtype=torch.float32
)
model_eager.save_pretrained(tmpdir)
with torch.device(torch_device):
model_sdpa = AutoModelForCausalLM.from_pretrained(
tmpdir, attn_implementation="sdpa", torch_dtype=torch.float32
)
model_eager = model_eager.eval()
model_sdpa = model_sdpa.eval()
with torch.no_grad():
with sdpa_kernel(enable_flash=False, enable_math=True, enable_mem_efficient=False):
res_eager = model_eager(**inputs_dict, return_dict=False)[0]
res_sdpa = model_sdpa(**inputs_dict, return_dict=False)[0]
# Only non-padding tokens are expected to match.
self.assertTrue(
torch.allclose(res_eager[attention_mask == 1], res_sdpa[attention_mask == 1], rtol=1e-4, atol=1e-4)
)
@require_flash_attn
@require_torch_gpu
@mark.flash_attn_test
def test_flash_attn_2_can_dispatch_composite_models(self):
"""
Tests if composite models can dispatch on FA2 if the sub-models support FA2.
The tests is needed as we handle differently composite models and we cannot check them
with above tests. If any of the sub-models does not support FA2, we'll raise an error when dispatching
that particular sub-model. Otherwise we dispatch safely in all sub-models, where "sub-models" are specific
backbone models (LM/vision/audio/etc)
"""
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
if not is_torch_fp16_available_on_device(torch_device):
self.skipTest(f"float16 not supported on {torch_device} (on the specific device currently used)")
torch_dtype = torch.float16
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
if not self._is_composite:
self.skipTest("This model is not a composte model!")
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
model = model_class.from_pretrained(tmpdirname, torch_dtype=torch_dtype)
sub_models_supporting_fa2 = [
module._supports_flash_attn_2
for name, module in model.named_modules()
if isinstance(module, PreTrainedModel) and name != ""
]
supports_fa2_all_modules = (
all(sub_models_supporting_fa2)
if len(sub_models_supporting_fa2) > 0
else model._supports_flash_attn_2
)
if not supports_fa2_all_modules:
with self.assertRaises(ValueError):
model_fa2 = model_class.from_pretrained(
tmpdirname, torch_dtype=torch_dtype, attn_implementation="flash_attention_2"
)
else:
model_fa2 = model_class.from_pretrained(
tmpdirname, torch_dtype=torch_dtype, attn_implementation="flash_attention_2"
)
for key in model_fa2.config:
if isinstance(getattr(model_fa2.config, key), PretrainedConfig):
sub_config = getattr(model_fa2.config, key)
self.assertTrue(sub_config._attn_implementation == "flash_attention_2")
has_fa2 = False
for name, submodule in model_fa2.named_modules():
class_name = submodule.__class__.__name__
if (
"Attention" in class_name
and getattr(submodule, "config", None)
and submodule.config._attn_implementation == "flash_attention_2"
):
has_fa2 = True
break
if not has_fa2:
raise ValueError("The FA2 model should have FA2 layers")
@require_flash_attn
@require_torch_gpu
@require_bitsandbytes
@mark.flash_attn_test
@slow
def test_flash_attn_2_fp32_ln(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
for model_class in self.all_generative_model_classes:
if not model_class._supports_flash_attn_2:
self.skipTest(f"{model_class.__name__} does not support Flash Attention 2")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
dummy_input = inputs_dict[model.main_input_name]
dummy_attention_mask = inputs_dict.get("attention_mask", torch.ones_like(dummy_input))
batch_size = dummy_attention_mask.shape[0]
is_padding_right = dummy_attention_mask[:, -1].sum().item() != batch_size
# To avoid errors with padding_side=="right"
if is_padding_right:
dummy_attention_mask = torch.ones_like(dummy_input)
model = model_class.from_pretrained(
tmpdirname,
torch_dtype=torch.float16,
attn_implementation="flash_attention_2",
low_cpu_mem_usage=True,
load_in_4bit=True,
)
for _, param in model.named_parameters():
# upcast only layer norms
if (param.dtype == torch.float16) or (param.dtype == torch.bfloat16):
param.data = param.data.to(torch.float32)
if model.config.is_encoder_decoder:
dummy_decoder_input_ids = inputs_dict["decoder_input_ids"]
dummy_decoder_attention_mask = inputs_dict["decoder_attention_mask"]
_ = model(dummy_input, decoder_input_ids=dummy_decoder_input_ids)
# with attention mask
_ = model(
dummy_input,
attention_mask=dummy_attention_mask,
decoder_input_ids=dummy_decoder_input_ids,
decoder_attention_mask=dummy_decoder_attention_mask,
)
else:
_ = model(dummy_input)
# with attention mask
_ = model(dummy_input, attention_mask=dummy_attention_mask)
@require_flash_attn
@require_torch_gpu
@mark.flash_attn_test
@slow
def test_flash_attention_2_padding_matches_padding_free_with_position_ids(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
max_new_tokens = 30
for model_class in self.all_generative_model_classes:
if not model_class._supports_flash_attn_2:
self.skipTest(f"{model_class.__name__} does not support Flash Attention 2")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
if 0 not in inputs_dict.get("attention_mask", []) or "attention_mask" not in inputs_dict:
self.skipTest("Model dummy inputs should contain padding in their attention mask")
dummy_input = inputs_dict[model_class.main_input_name]
if dummy_input.dtype in [torch.float32, torch.bfloat16]:
dummy_input = dummy_input.to(torch.float16)
# make sure that all models have enough positions for generation
if hasattr(config, "max_position_embeddings"):
config.max_position_embeddings = max_new_tokens + dummy_input.shape[1] + 1
model = model_class(config)
if "position_ids" not in inspect.signature(model.forward).parameters:
self.skipTest("Model does not support position_ids")
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(tmpdirname)
# ensure left padding, to adapt for some models
if 0 in inputs_dict["attention_mask"][:, -1]:
inputs_dict["attention_mask"] = inputs_dict["attention_mask"].flip(1)
dummy_attention_mask = inputs_dict["attention_mask"]
inputs_dict["input_ids"][~dummy_attention_mask.bool()] = config.get_text_config().pad_token_id
model = (
model_class.from_pretrained(
tmpdirname,
torch_dtype=torch.float16,
attn_implementation="flash_attention_2",
low_cpu_mem_usage=True,
)
.to(torch_device)
.eval()
)
# flatten
padfree_inputs_dict = {
k: v[dummy_attention_mask.bool()].unsqueeze(0)
for k, v in inputs_dict.items()
if not k == "attention_mask"
}
# add position_ids
padfree_inputs_dict["position_ids"] = (
torch.cat([torch.arange(length) for length in dummy_attention_mask.sum(1).tolist()])
.long()
.unsqueeze(0)
.to(torch_device)
)
res_padded = model(**inputs_dict)
res_padfree = model(**padfree_inputs_dict)
logits_padded = res_padded.logits[inputs_dict["attention_mask"].bool()]
logits_padfree = res_padfree.logits[0]
torch.testing.assert_close(logits_padded.argmax(-1), logits_padfree.argmax(-1), rtol=0, atol=0)
# acceptable numerical instability
tol = torch.finfo(torch.float16).eps
torch.testing.assert_close(logits_padded, logits_padfree, rtol=tol, atol=tol)
@is_pt_tf_cross_test
def test_tf_from_pt_safetensors(self):
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
tf_model_class_name = "TF" + model_class.__name__ # Add the "TF" at the beginning
if not hasattr(transformers, tf_model_class_name):
self.skipTest(reason="transformers does not have this model in TF version yet")
tf_model_class = getattr(transformers, tf_model_class_name)
pt_model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
pt_model.save_pretrained(tmpdirname, safe_serialization=True)
tf_model_1 = tf_model_class.from_pretrained(tmpdirname, from_pt=True)
pt_model.save_pretrained(tmpdirname, safe_serialization=False)
tf_model_2 = tf_model_class.from_pretrained(tmpdirname, from_pt=True)
# Check models are equal
for p1, p2 in zip(tf_model_1.weights, tf_model_2.weights):
self.assertTrue(np.allclose(p1.numpy(), p2.numpy()))
@is_pt_flax_cross_test
def test_flax_from_pt_safetensors(self):
for model_class in self.all_model_classes:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
flax_model_class_name = "Flax" + model_class.__name__ # Add the "Flax at the beginning
if not hasattr(transformers, flax_model_class_name):
self.skipTest(reason="transformers does not have this model in Flax version yet")
flax_model_class = getattr(transformers, flax_model_class_name)
pt_model = model_class(config)
with tempfile.TemporaryDirectory() as tmpdirname:
pt_model.save_pretrained(tmpdirname, safe_serialization=True)
flax_model_1 = flax_model_class.from_pretrained(tmpdirname, from_pt=True)
pt_model.save_pretrained(tmpdirname, safe_serialization=False)
flax_model_2 = flax_model_class.from_pretrained(tmpdirname, from_pt=True)
# Check models are equal
self.assertTrue(check_models_equal(flax_model_1, flax_model_2))
@require_flash_attn
@require_torch_gpu
@mark.flash_attn_test
@slow
def test_flash_attn_2_from_config(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
for model_class in self.all_generative_model_classes:
if not model_class._supports_flash_attn_2:
self.skipTest(f"{model_class.__name__} does not support Flash Attention 2")
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
# TODO: to change it in the future with other relevant auto classes
fa2_model = model_class._from_config(
config, attn_implementation="flash_attention_2", torch_dtype=torch.bfloat16
).to(torch_device)
dummy_input = torch.LongTensor([[0, 2, 3, 4], [0, 2, 3, 4]]).to(torch_device)
dummy_attention_mask = torch.LongTensor([[1, 1, 1, 1], [0, 1, 1, 1]]).to(torch_device)
_ = fa2_model(input_ids=dummy_input, attention_mask=dummy_attention_mask)
with tempfile.TemporaryDirectory() as tmpdirname:
fa2_model.save_pretrained(tmpdirname)
model_from_pretrained = model_class.from_pretrained(tmpdirname)
self.assertTrue(model_from_pretrained.config._attn_implementation != "flash_attention_2")
def _get_custom_4d_mask_test_data(self):
# Sequence in which all but the last token is the same
input_ids = torch.tensor(
[[10, 11, 12, 13], [10, 11, 12, 14], [10, 11, 12, 15]], device=torch_device, dtype=torch.int64
)
position_ids = torch.tensor([[0, 1, 2, 3]] * 3, device=torch_device, dtype=torch.int64)
# Combining common prefix with the unique ending tokens:
input_ids_shared_prefix = torch.cat([input_ids[0][:-1], input_ids[:, -1]]).unsqueeze(0)
# Creating a 4D mask where each of the last 3 tokens do not attend to each other.
mask_shared_prefix = torch.tensor(
[
[
[
[1, 0, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0],
[1, 1, 1, 0, 0, 0],
[1, 1, 1, 1, 0, 0],
[1, 1, 1, 0, 1, 0],
[1, 1, 1, 0, 0, 1],
]
]
],
)
# inverting the attention mask
mask_dtype = torch.float32
min_dtype = torch.finfo(mask_dtype).min
mask_shared_prefix = (mask_shared_prefix.eq(0.0)).to(dtype=mask_dtype, device=torch_device) * min_dtype
# Creating a position_ids tensor. note the repeating figures in the end.
position_ids_shared_prefix = torch.tensor([[0, 1, 2, 3, 3, 3]], device=torch_device, dtype=torch.int64)
return input_ids, position_ids, input_ids_shared_prefix, mask_shared_prefix, position_ids_shared_prefix
def test_custom_4d_attention_mask(self):
if not self.has_attentions:
self.skipTest(reason="Model architecture does not support attentions")
if len(self.all_generative_model_classes) == 0:
self.skipTest(
reason="Model architecture has no generative classes, and thus not necessarily supporting 4D masks"
)
set_model_tester_for_less_flaky_test(self)
for model_class in self.all_generative_model_classes:
if not model_class._supports_static_cache:
self.skipTest(f"{model_class.__name__} is not guaranteed to work with custom 4D attention masks")
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
set_config_for_less_flaky_test(config)
if getattr(config, "sliding_window", 0) is not None and getattr(config, "sliding_window", 0) > 0:
self.skipTest(f"{model_class.__name__} with sliding window attention is not supported by this test")
model = model_class(config).to(device=torch_device, dtype=torch.float32)
set_model_for_less_flaky_test(model)
(
input_ids,
position_ids,
input_ids_shared_prefix,
mask_shared_prefix,
position_ids_shared_prefix,
) = self._get_custom_4d_mask_test_data()
logits = model.forward(input_ids, position_ids=position_ids).logits
# logits.shape == torch.Size([3, 4, ...])
logits_shared_prefix = model(
input_ids_shared_prefix,
attention_mask=mask_shared_prefix,
position_ids=position_ids_shared_prefix,
)[0]
# logits_shared_prefix.shape == torch.Size([1, 6, ...])
out_last_tokens = logits[:, -1, :] # last tokens in each batch line
out_shared_prefix_last_tokens = logits_shared_prefix[0, -3:, :] # last three tokens
# comparing softmax-normalized logits:
normalized_0 = F.softmax(out_last_tokens)
normalized_1 = F.softmax(out_shared_prefix_last_tokens)
torch.testing.assert_close(normalized_0, normalized_1, rtol=1e-3, atol=1e-4)
@slow
@require_torch_accelerator
def test_torch_compile_for_training(self):
if version.parse(torch.__version__) < version.parse("2.3"):
self.skipTest(reason="This test requires torch >= 2.3 to run.")
if not hasattr(self, "_torch_compile_train_cls"):
self.skipTest(f"{self.__class__.__name__} doesn't have the attribute `_torch_compile_train_cls`.")
config, _ = self.model_tester.prepare_config_and_inputs_for_common()
cls = self._torch_compile_train_cls
model = cls(config).to(torch_device)
inputs = {
"input_ids": torch.randint(low=1, high=model.config.vocab_size, size=(2, 10), device=torch_device),
"attention_mask": torch.tensor(
[[1, 1, 1, 1, 1, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]],
dtype=torch.int64,
device=torch_device,
),
"position_ids": torch.arange(0, 10, device=torch_device).unsqueeze(0),
"labels": torch.randint(low=1, high=model.config.vocab_size, size=(2, 10), device=torch_device),
}
# eager backward
set_seed(42)
loss = model(**inputs).loss
loss.backward()
params = {name: param.grad.clone().detach().cpu() for name, param in model.named_parameters()}
model.zero_grad()
del loss
model = torch.compile(model, fullgraph=True, mode="reduce-overhead")
# forward compilation
set_seed(42)
loss = model(**inputs).loss
# backward compilation
loss.backward()
# check grad matches
for name, param in model._orig_mod.named_parameters():
torch.testing.assert_close(param.grad.detach().cpu(), params[name], rtol=1e-4, atol=1e-4)
def test_forward_with_logits_to_keep(self):
for model_class in self.all_generative_model_classes:
if "logits_to_keep" not in set(inspect.signature(model_class.forward).parameters.keys()):
self.skipTest(reason="This model does not support `logits_to_keep` argument.")
config, inputs = self.model_tester.prepare_config_and_inputs_for_common()
batch_size, sequence_length = inputs["input_ids"].shape
vocab_size = config.get_text_config().vocab_size
model = model_class(config).to(device=torch_device).eval()
# some models have labels but `logits_to_keep` should not be used in train mode
_ = inputs.pop("labels", None)
# logits_to_keep=0 is a special case meaning "keep all logits"
all_logits = model(**inputs, logits_to_keep=0).logits
last_token_logits = model(**inputs, logits_to_keep=1).logits
# Assert all shapes are correct
self.assertEqual(tuple(all_logits.shape), (batch_size, sequence_length, vocab_size))
self.assertEqual(tuple(last_token_logits.shape), (batch_size, 1, vocab_size))
# Assert the last tokens are actually the same (except for the natural fluctuation due to order of FP ops)
torch.testing.assert_close(all_logits[:, -1:, :], last_token_logits, rtol=1e-5, atol=1e-5)
@require_torch_gpu
def test_flex_attention_with_grads(self):
for model_class in self.all_model_classes:
if not model_class._supports_flex_attn:
self.skipTest(reason="This model does not support flex attention")
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
config._attn_implementation = "flex_attention"
model = model_class(config).to(device=torch_device, dtype=torch.float16)
self.assertTrue(model.config._attn_implementation == "flex_attention")
# If this does not raise an error, the test passes (see https://github.com/huggingface/transformers/pull/35605)
_ = model(inputs_dict["input_ids"].to(torch_device))
global_rng = random.Random()
def ids_tensor(shape, vocab_size, rng=None, name=None):
# Creates a random int32 tensor of the shape within the vocab size
if rng is None:
rng = global_rng
total_dims = 1
for dim in shape:
total_dims *= dim
values = []
for _ in range(total_dims):
values.append(rng.randint(0, vocab_size - 1))
return torch.tensor(data=values, dtype=torch.long, device=torch_device).view(shape).contiguous()
def random_attention_mask(shape, rng=None, name=None):
attn_mask = ids_tensor(shape, vocab_size=2, rng=None, name=None)
# make sure that at least one token is attended to for each batch
# we choose the 1st token so this property of `at least one being non-zero` still holds after applying causal mask
attn_mask[:, 0] = 1
return attn_mask
def floats_tensor(shape, scale=1.0, rng=None, name=None):
"""Creates a random float32 tensor"""
if rng is None:
rng = global_rng
total_dims = 1
for dim in shape:
total_dims *= dim
values = []
for _ in range(total_dims):
values.append(rng.random() * scale)
return torch.tensor(data=values, dtype=torch.float, device=torch_device).view(shape).contiguous()
| transformers/tests/test_modeling_common.py/0 | {
"file_path": "transformers/tests/test_modeling_common.py",
"repo_id": "transformers",
"token_count": 120034
} |
# Copyright 2024 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import Dict
from transformers import is_torch_available
from transformers.testing_utils import (
TestCasePlus,
execute_subprocess_async,
get_torch_dist_unique_port,
require_accelerate,
require_fp8,
require_fsdp,
require_torch_multi_gpu,
)
if is_torch_available():
import torch
import torch.distributed
import torch.utils.data
from transformers import (
AutoModelForCausalLM,
AutoTokenizer,
DataCollatorForSeq2Seq,
EvalPrediction,
GenerationConfig,
HfArgumentParser,
PreTrainedTokenizerBase,
Seq2SeqTrainer,
Seq2SeqTrainingArguments,
)
class DummyTextDataset(torch.utils.data.Dataset[str]):
def __init__(self, tokenizer: PreTrainedTokenizerBase) -> None:
data = 4 * [
"Hello world!",
"The quick brown fox jumps over the lazy dog.",
]
self.data = [
{k: v.squeeze(0) for k, v in tokenizer(item, return_tensors="pt", return_attention_mask=True).items()}
for item in data
]
for item in self.data:
item["labels"] = item["input_ids"]
def __len__(self) -> int:
return len(self.data)
def __getitem__(self, i: int) -> str:
return self.data[i]
class TestFSDPTrainer(TestCasePlus):
@require_accelerate
@require_torch_multi_gpu
@require_fsdp
def test_trainer(self):
output_dir = self.get_auto_remove_tmp_dir()
cmd = [
"accelerate",
"launch",
"--use_fsdp",
"--main_process_port",
f"{get_torch_dist_unique_port()}",
"--num_processes",
f"{torch.cuda.device_count()}",
"--fsdp_transformer_layer_cls_to_wrap",
"GPT2Block",
f"{self.test_file_dir}/test_trainer_fsdp.py",
"--output_dir",
f"{output_dir}",
"--report_to",
"none",
]
execute_subprocess_async(cmd, env=self.get_env())
# successful return here == success - any errors would have caused an error in the sub-call
class TestFSDPTrainerFP8(TestCasePlus):
@require_accelerate
@require_torch_multi_gpu
@require_fsdp
@require_fp8
def test_trainer(self):
output_dir = self.get_auto_remove_tmp_dir()
cmd = [
"accelerate",
"launch",
"--use_fsdp",
"--main_process_port",
f"{get_torch_dist_unique_port()}",
"--num_processes",
f"{torch.cuda.device_count()}",
"--mixed_precision",
"fp8",
"--fsdp_transformer_layer_cls_to_wrap",
"GPT2Block",
f"{self.test_file_dir}/test_trainer_fsdp.py",
"--output_dir",
f"{output_dir}",
"--report_to",
"none",
]
execute_subprocess_async(cmd, env=self.get_env())
# successful return here == success - any errors would have caused an error in the sub-call
class TestFSDPTrainerWrap(TestCasePlus):
@require_accelerate
@require_torch_multi_gpu
@require_fsdp
def test_trainer(self):
output_dir = self.get_auto_remove_tmp_dir()
cmd = [
"accelerate",
"launch",
"--use_fsdp",
"--main_process_port",
f"{get_torch_dist_unique_port()}",
"--num_processes",
f"{torch.cuda.device_count()}",
"--fsdp_transformer_layer_cls_to_wrap",
"GPT2Block",
f"{self.test_file_dir}/test_trainer_fsdp.py",
"--output_dir",
f"{output_dir}",
"--report_to",
"none",
"--auto_find_batch_size",
"True",
]
execute_subprocess_async(cmd, env=self.get_env())
# successful return here == success - any errors would have caused an error in the sub-call
if __name__ == "__main__":
parser = HfArgumentParser((Seq2SeqTrainingArguments,))
training_args = parser.parse_args_into_dataclasses()[0]
training_args.per_device_eval_batch_size = 1
training_args.use_legacy_prediction_loop = False
training_args.predict_with_generate = True
training_args.generation_config = GenerationConfig(max_length=30)
pretrained_model_name = "hf-internal-testing/tiny-random-gpt2"
tokenizer = AutoTokenizer.from_pretrained(pretrained_model_name)
tokenizer.pad_token = tokenizer.eos_token
device = torch.device(torch.distributed.get_rank())
model = AutoModelForCausalLM.from_pretrained(pretrained_model_name).to(device)
def compute_metrics(p: EvalPrediction) -> Dict[str, bool]:
return {"accuracy": (p.predictions == p.label_ids).mean()}
trainer = Seq2SeqTrainer(
model=model,
args=training_args,
data_collator=DataCollatorForSeq2Seq(tokenizer, model),
eval_dataset=DummyTextDataset(tokenizer),
compute_metrics=compute_metrics,
)
metrics = trainer.evaluate()
| transformers/tests/trainer/test_trainer_fsdp.py/0 | {
"file_path": "transformers/tests/trainer/test_trainer_fsdp.py",
"repo_id": "transformers",
"token_count": 2775
} |
# coding=utf-8
# Copyright 2019-present, the HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import shutil
import unittest
from unittest.mock import patch
from transformers.testing_utils import CaptureStd, require_torch
class CLITest(unittest.TestCase):
@patch("sys.argv", ["fakeprogrampath", "env"])
def test_cli_env(self):
# test transformers-cli env
import transformers.commands.transformers_cli
with CaptureStd() as cs:
transformers.commands.transformers_cli.main()
self.assertIn("Python version", cs.out)
self.assertIn("Platform", cs.out)
self.assertIn("Using distributed or parallel set-up in script?", cs.out)
@require_torch
@patch("sys.argv", ["fakeprogrampath", "download", "hf-internal-testing/tiny-random-gptj", "--cache-dir", "/tmp"])
def test_cli_download(self):
import transformers.commands.transformers_cli
# # remove any previously downloaded model to start clean
shutil.rmtree("/tmp/models--hf-internal-testing--tiny-random-gptj", ignore_errors=True)
# run the command
transformers.commands.transformers_cli.main()
# check if the model files are downloaded correctly on /tmp/models--hf-internal-testing--tiny-random-gptj
self.assertTrue(os.path.exists("/tmp/models--hf-internal-testing--tiny-random-gptj/blobs"))
self.assertTrue(os.path.exists("/tmp/models--hf-internal-testing--tiny-random-gptj/refs"))
self.assertTrue(os.path.exists("/tmp/models--hf-internal-testing--tiny-random-gptj/snapshots"))
@require_torch
@patch(
"sys.argv",
[
"fakeprogrampath",
"download",
"hf-internal-testing/test_dynamic_model_with_tokenizer",
"--trust-remote-code",
"--cache-dir",
"/tmp",
],
)
def test_cli_download_trust_remote(self):
import transformers.commands.transformers_cli
# # remove any previously downloaded model to start clean
shutil.rmtree("/tmp/models--hf-internal-testing--test_dynamic_model_with_tokenizer", ignore_errors=True)
# run the command
transformers.commands.transformers_cli.main()
# check if the model files are downloaded correctly on /tmp/models--hf-internal-testing--test_dynamic_model_with_tokenizer
self.assertTrue(os.path.exists("/tmp/models--hf-internal-testing--test_dynamic_model_with_tokenizer/blobs"))
self.assertTrue(os.path.exists("/tmp/models--hf-internal-testing--test_dynamic_model_with_tokenizer/refs"))
self.assertTrue(
os.path.exists("/tmp/models--hf-internal-testing--test_dynamic_model_with_tokenizer/snapshots")
)
| transformers/tests/utils/test_cli.py/0 | {
"file_path": "transformers/tests/utils/test_cli.py",
"repo_id": "transformers",
"token_count": 1250
} |
# coding=utf-8
# Copyright 2020 The Hugging Face Team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import io
import unittest
from dataclasses import dataclass
from typing import Optional
from transformers import AlbertForMaskedLM
from transformers.testing_utils import require_torch
from transformers.utils import ModelOutput, is_torch_available
if is_torch_available():
import torch
from transformers.pytorch_utils import is_torch_greater_or_equal_than_2_2
@dataclass
class ModelOutputTest(ModelOutput):
a: float
b: Optional[float] = None
c: Optional[float] = None
class ModelOutputTester(unittest.TestCase):
def test_get_attributes(self):
x = ModelOutputTest(a=30)
self.assertEqual(x.a, 30)
self.assertIsNone(x.b)
self.assertIsNone(x.c)
with self.assertRaises(AttributeError):
_ = x.d
def test_index_with_ints_and_slices(self):
x = ModelOutputTest(a=30, b=10)
self.assertEqual(x[0], 30)
self.assertEqual(x[1], 10)
self.assertEqual(x[:2], (30, 10))
self.assertEqual(x[:], (30, 10))
x = ModelOutputTest(a=30, c=10)
self.assertEqual(x[0], 30)
self.assertEqual(x[1], 10)
self.assertEqual(x[:2], (30, 10))
self.assertEqual(x[:], (30, 10))
def test_index_with_strings(self):
x = ModelOutputTest(a=30, b=10)
self.assertEqual(x["a"], 30)
self.assertEqual(x["b"], 10)
with self.assertRaises(KeyError):
_ = x["c"]
x = ModelOutputTest(a=30, c=10)
self.assertEqual(x["a"], 30)
self.assertEqual(x["c"], 10)
with self.assertRaises(KeyError):
_ = x["b"]
def test_dict_like_properties(self):
x = ModelOutputTest(a=30)
self.assertEqual(list(x.keys()), ["a"])
self.assertEqual(list(x.values()), [30])
self.assertEqual(list(x.items()), [("a", 30)])
self.assertEqual(list(x), ["a"])
x = ModelOutputTest(a=30, b=10)
self.assertEqual(list(x.keys()), ["a", "b"])
self.assertEqual(list(x.values()), [30, 10])
self.assertEqual(list(x.items()), [("a", 30), ("b", 10)])
self.assertEqual(list(x), ["a", "b"])
x = ModelOutputTest(a=30, c=10)
self.assertEqual(list(x.keys()), ["a", "c"])
self.assertEqual(list(x.values()), [30, 10])
self.assertEqual(list(x.items()), [("a", 30), ("c", 10)])
self.assertEqual(list(x), ["a", "c"])
with self.assertRaises(Exception):
x = x.update({"d": 20})
with self.assertRaises(Exception):
del x["a"]
with self.assertRaises(Exception):
_ = x.pop("a")
with self.assertRaises(Exception):
_ = x.setdefault("d", 32)
def test_set_attributes(self):
x = ModelOutputTest(a=30)
x.a = 10
self.assertEqual(x.a, 10)
self.assertEqual(x["a"], 10)
def test_set_keys(self):
x = ModelOutputTest(a=30)
x["a"] = 10
self.assertEqual(x.a, 10)
self.assertEqual(x["a"], 10)
def test_instantiate_from_dict(self):
x = ModelOutputTest({"a": 30, "b": 10})
self.assertEqual(list(x.keys()), ["a", "b"])
self.assertEqual(x.a, 30)
self.assertEqual(x.b, 10)
def test_instantiate_from_iterator(self):
x = ModelOutputTest([("a", 30), ("b", 10)])
self.assertEqual(list(x.keys()), ["a", "b"])
self.assertEqual(x.a, 30)
self.assertEqual(x.b, 10)
with self.assertRaises(ValueError):
_ = ModelOutputTest([("a", 30), (10, 10)])
x = ModelOutputTest(a=(30, 30))
self.assertEqual(list(x.keys()), ["a"])
self.assertEqual(x.a, (30, 30))
@require_torch
def test_torch_pytree(self):
# ensure torch.utils._pytree treats ModelOutput subclasses as nodes (and not leaves)
# this is important for DistributedDataParallel gradient synchronization with static_graph=True
import torch.utils._pytree as pytree
x = ModelOutput({"a": 1.0, "c": 2.0})
self.assertFalse(pytree._is_leaf(x))
x = ModelOutputTest(a=1.0, c=2.0)
self.assertFalse(pytree._is_leaf(x))
expected_flat_outs = [1.0, 2.0]
expected_tree_spec = pytree.TreeSpec(ModelOutputTest, ["a", "c"], [pytree.LeafSpec(), pytree.LeafSpec()])
actual_flat_outs, actual_tree_spec = pytree.tree_flatten(x)
self.assertEqual(expected_flat_outs, actual_flat_outs)
self.assertEqual(expected_tree_spec, actual_tree_spec)
unflattened_x = pytree.tree_unflatten(actual_flat_outs, actual_tree_spec)
self.assertEqual(x, unflattened_x)
if is_torch_greater_or_equal_than_2_2:
self.assertEqual(
pytree.treespec_dumps(actual_tree_spec),
'[1, {"type": "tests.utils.test_model_output.ModelOutputTest", "context": "[\\"a\\", \\"c\\"]", "children_spec": [{"type": null, "context": null, "children_spec": []}, {"type": null, "context": null, "children_spec": []}]}]',
)
# TODO: @ydshieh
@unittest.skip(reason="CPU OOM")
@require_torch
def test_export_serialization(self):
if not is_torch_greater_or_equal_than_2_2:
self.skipTest(reason="Export serialization requires torch >= 2.2.0")
model_cls = AlbertForMaskedLM
model_config = model_cls.config_class()
model = model_cls(model_config)
input_dict = {"input_ids": torch.randint(0, 30000, (1, 512), dtype=torch.int64, requires_grad=False)}
ep = torch.export.export(model, (), input_dict)
buffer = io.BytesIO()
torch.export.save(ep, buffer)
buffer.seek(0)
loaded_ep = torch.export.load(buffer)
input_dict = {"input_ids": torch.randint(0, 30000, (1, 512), dtype=torch.int64, requires_grad=False)}
assert torch.allclose(model(**input_dict).logits, loaded_ep(**input_dict).logits)
class ModelOutputTestNoDataclass(ModelOutput):
"""Invalid test subclass of ModelOutput where @dataclass decorator is not used"""
a: float
b: Optional[float] = None
c: Optional[float] = None
class ModelOutputSubclassTester(unittest.TestCase):
def test_direct_model_output(self):
# Check that direct usage of ModelOutput instantiates without errors
ModelOutput({"a": 1.1})
def test_subclass_no_dataclass(self):
# Check that a subclass of ModelOutput without @dataclass is invalid
# A valid subclass is inherently tested other unit tests above.
with self.assertRaises(TypeError):
ModelOutputTestNoDataclass(a=1.1, b=2.2, c=3.3)
| transformers/tests/utils/test_model_output.py/0 | {
"file_path": "transformers/tests/utils/test_model_output.py",
"repo_id": "transformers",
"token_count": 3188
} |
# coding=utf-8
# Copyright 2022 The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import inspect
import re
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/check_config_docstrings.py
PATH_TO_TRANSFORMERS = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
transformers = direct_transformers_import(PATH_TO_TRANSFORMERS)
CONFIG_MAPPING = transformers.models.auto.configuration_auto.CONFIG_MAPPING
# Regex pattern used to find the checkpoint mentioned in the docstring of `config_class`.
# For example, `[google-bert/bert-base-uncased](https://huggingface.co/google-bert/bert-base-uncased)`
_re_checkpoint = re.compile(r"\[(.+?)\]\((https://huggingface\.co/.+?)\)")
CONFIG_CLASSES_TO_IGNORE_FOR_DOCSTRING_CHECKPOINT_CHECK = {
"DecisionTransformerConfig",
"EncoderDecoderConfig",
"MusicgenConfig",
"RagConfig",
"SpeechEncoderDecoderConfig",
"TimmBackboneConfig",
"TimmWrapperConfig",
"VisionEncoderDecoderConfig",
"VisionTextDualEncoderConfig",
"LlamaConfig",
"GraniteConfig",
"GraniteMoeConfig",
}
def get_checkpoint_from_config_class(config_class):
checkpoint = None
# source code of `config_class`
config_source = inspect.getsource(config_class)
checkpoints = _re_checkpoint.findall(config_source)
# Each `checkpoint` is a tuple of a checkpoint name and a checkpoint link.
# For example, `('google-bert/bert-base-uncased', 'https://huggingface.co/google-bert/bert-base-uncased')`
for ckpt_name, ckpt_link in checkpoints:
# allow the link to end with `/`
if ckpt_link.endswith("/"):
ckpt_link = ckpt_link[:-1]
# verify the checkpoint name corresponds to the checkpoint link
ckpt_link_from_name = f"https://huggingface.co/{ckpt_name}"
if ckpt_link == ckpt_link_from_name:
checkpoint = ckpt_name
break
return checkpoint
def check_config_docstrings_have_checkpoints():
configs_without_checkpoint = []
for config_class in list(CONFIG_MAPPING.values()):
# Skip deprecated models
if "models.deprecated" in config_class.__module__:
continue
checkpoint = get_checkpoint_from_config_class(config_class)
name = config_class.__name__
if checkpoint is None and name not in CONFIG_CLASSES_TO_IGNORE_FOR_DOCSTRING_CHECKPOINT_CHECK:
configs_without_checkpoint.append(name)
if len(configs_without_checkpoint) > 0:
message = "\n".join(sorted(configs_without_checkpoint))
raise ValueError(
f"The following configurations don't contain any valid checkpoint:\n{message}\n\n"
"The requirement is to include a link pointing to one of the models of this architecture in the "
"docstring of the config classes listed above. The link should have be a markdown format like "
"[myorg/mymodel](https://huggingface.co/myorg/mymodel)."
)
if __name__ == "__main__":
check_config_docstrings_have_checkpoints()
| transformers/utils/check_config_docstrings.py/0 | {
"file_path": "transformers/utils/check_config_docstrings.py",
"repo_id": "transformers",
"token_count": 1326
} |
# coding=utf-8
# Copyright 2021 The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Utility that sorts the imports in the custom inits of Transformers. Transformers uses init files that delay the
import of an object to when it's actually needed. This is to avoid the main init importing all models, which would
make the line `import transformers` very slow when the user has all optional dependencies installed. The inits with
delayed imports have two halves: one definining a dictionary `_import_structure` which maps modules to the name of the
objects in each module, and one in `TYPE_CHECKING` which looks like a normal init for type-checkers. `isort` or `ruff`
properly sort the second half which looks like traditionl imports, the goal of this script is to sort the first half.
Use from the root of the repo with:
```bash
python utils/custom_init_isort.py
```
which will auto-sort the imports (used in `make style`).
For a check only (as used in `make quality`) run:
```bash
python utils/custom_init_isort.py --check_only
```
"""
import argparse
import os
import re
from typing import Any, Callable, List, Optional
# Path is defined with the intent you should run this script from the root of the repo.
PATH_TO_TRANSFORMERS = "src/transformers"
# Pattern that looks at the indentation in a line.
_re_indent = re.compile(r"^(\s*)\S")
# Pattern that matches `"key":" and puts `key` in group 0.
_re_direct_key = re.compile(r'^\s*"([^"]+)":')
# Pattern that matches `_import_structure["key"]` and puts `key` in group 0.
_re_indirect_key = re.compile(r'^\s*_import_structure\["([^"]+)"\]')
# Pattern that matches `"key",` and puts `key` in group 0.
_re_strip_line = re.compile(r'^\s*"([^"]+)",\s*$')
# Pattern that matches any `[stuff]` and puts `stuff` in group 0.
_re_bracket_content = re.compile(r"\[([^\]]+)\]")
def get_indent(line: str) -> str:
"""Returns the indent in given line (as string)."""
search = _re_indent.search(line)
return "" if search is None else search.groups()[0]
def split_code_in_indented_blocks(
code: str, indent_level: str = "", start_prompt: Optional[str] = None, end_prompt: Optional[str] = None
) -> List[str]:
"""
Split some code into its indented blocks, starting at a given level.
Args:
code (`str`): The code to split.
indent_level (`str`): The indent level (as string) to use for identifying the blocks to split.
start_prompt (`str`, *optional*): If provided, only starts splitting at the line where this text is.
end_prompt (`str`, *optional*): If provided, stops splitting at a line where this text is.
Warning:
The text before `start_prompt` or after `end_prompt` (if provided) is not ignored, just not split. The input `code`
can thus be retrieved by joining the result.
Returns:
`List[str]`: The list of blocks.
"""
# Let's split the code into lines and move to start_index.
index = 0
lines = code.split("\n")
if start_prompt is not None:
while not lines[index].startswith(start_prompt):
index += 1
blocks = ["\n".join(lines[:index])]
else:
blocks = []
# This variable contains the block treated at a given time.
current_block = [lines[index]]
index += 1
# We split into blocks until we get to the `end_prompt` (or the end of the file).
while index < len(lines) and (end_prompt is None or not lines[index].startswith(end_prompt)):
# We have a non-empty line with the proper indent -> start of a new block
if len(lines[index]) > 0 and get_indent(lines[index]) == indent_level:
# Store the current block in the result and rest. There are two cases: the line is part of the block (like
# a closing parenthesis) or not.
if len(current_block) > 0 and get_indent(current_block[-1]).startswith(indent_level + " "):
# Line is part of the current block
current_block.append(lines[index])
blocks.append("\n".join(current_block))
if index < len(lines) - 1:
current_block = [lines[index + 1]]
index += 1
else:
current_block = []
else:
# Line is not part of the current block
blocks.append("\n".join(current_block))
current_block = [lines[index]]
else:
# Just add the line to the current block
current_block.append(lines[index])
index += 1
# Adds current block if it's nonempty.
if len(current_block) > 0:
blocks.append("\n".join(current_block))
# Add final block after end_prompt if provided.
if end_prompt is not None and index < len(lines):
blocks.append("\n".join(lines[index:]))
return blocks
def ignore_underscore_and_lowercase(key: Callable[[Any], str]) -> Callable[[Any], str]:
"""
Wraps a key function (as used in a sort) to lowercase and ignore underscores.
"""
def _inner(x):
return key(x).lower().replace("_", "")
return _inner
def sort_objects(objects: List[Any], key: Optional[Callable[[Any], str]] = None) -> List[Any]:
"""
Sort a list of objects following the rules of isort (all uppercased first, camel-cased second and lower-cased
last).
Args:
objects (`List[Any]`):
The list of objects to sort.
key (`Callable[[Any], str]`, *optional*):
A function taking an object as input and returning a string, used to sort them by alphabetical order.
If not provided, will default to noop (so a `key` must be provided if the `objects` are not of type string).
Returns:
`List[Any]`: The sorted list with the same elements as in the inputs
"""
# If no key is provided, we use a noop.
def noop(x):
return x
if key is None:
key = noop
# Constants are all uppercase, they go first.
constants = [obj for obj in objects if key(obj).isupper()]
# Classes are not all uppercase but start with a capital, they go second.
classes = [obj for obj in objects if key(obj)[0].isupper() and not key(obj).isupper()]
# Functions begin with a lowercase, they go last.
functions = [obj for obj in objects if not key(obj)[0].isupper()]
# Then we sort each group.
key1 = ignore_underscore_and_lowercase(key)
return sorted(constants, key=key1) + sorted(classes, key=key1) + sorted(functions, key=key1)
def sort_objects_in_import(import_statement: str) -> str:
"""
Sorts the imports in a single import statement.
Args:
import_statement (`str`): The import statement in which to sort the imports.
Returns:
`str`: The same as the input, but with objects properly sorted.
"""
# This inner function sort imports between [ ].
def _replace(match):
imports = match.groups()[0]
# If there is one import only, nothing to do.
if "," not in imports:
return f"[{imports}]"
keys = [part.strip().replace('"', "") for part in imports.split(",")]
# We will have a final empty element if the line finished with a comma.
if len(keys[-1]) == 0:
keys = keys[:-1]
return "[" + ", ".join([f'"{k}"' for k in sort_objects(keys)]) + "]"
lines = import_statement.split("\n")
if len(lines) > 3:
# Here we have to sort internal imports that are on several lines (one per name):
# key: [
# "object1",
# "object2",
# ...
# ]
# We may have to ignore one or two lines on each side.
idx = 2 if lines[1].strip() == "[" else 1
keys_to_sort = [(i, _re_strip_line.search(line).groups()[0]) for i, line in enumerate(lines[idx:-idx])]
sorted_indices = sort_objects(keys_to_sort, key=lambda x: x[1])
sorted_lines = [lines[x[0] + idx] for x in sorted_indices]
return "\n".join(lines[:idx] + sorted_lines + lines[-idx:])
elif len(lines) == 3:
# Here we have to sort internal imports that are on one separate line:
# key: [
# "object1", "object2", ...
# ]
if _re_bracket_content.search(lines[1]) is not None:
lines[1] = _re_bracket_content.sub(_replace, lines[1])
else:
keys = [part.strip().replace('"', "") for part in lines[1].split(",")]
# We will have a final empty element if the line finished with a comma.
if len(keys[-1]) == 0:
keys = keys[:-1]
lines[1] = get_indent(lines[1]) + ", ".join([f'"{k}"' for k in sort_objects(keys)])
return "\n".join(lines)
else:
# Finally we have to deal with imports fitting on one line
import_statement = _re_bracket_content.sub(_replace, import_statement)
return import_statement
def sort_imports(file: str, check_only: bool = True):
"""
Sort the imports defined in the `_import_structure` of a given init.
Args:
file (`str`): The path to the init to check/fix.
check_only (`bool`, *optional*, defaults to `True`): Whether or not to just check (and not auto-fix) the init.
"""
with open(file, encoding="utf-8") as f:
code = f.read()
# If the file is not a custom init, there is nothing to do.
if "_import_structure" not in code or "define_import_structure" in code:
return
# Blocks of indent level 0
main_blocks = split_code_in_indented_blocks(
code, start_prompt="_import_structure = {", end_prompt="if TYPE_CHECKING:"
)
# We ignore block 0 (everything untils start_prompt) and the last block (everything after end_prompt).
for block_idx in range(1, len(main_blocks) - 1):
# Check if the block contains some `_import_structure`s thingy to sort.
block = main_blocks[block_idx]
block_lines = block.split("\n")
# Get to the start of the imports.
line_idx = 0
while line_idx < len(block_lines) and "_import_structure" not in block_lines[line_idx]:
# Skip dummy import blocks
if "import dummy" in block_lines[line_idx]:
line_idx = len(block_lines)
else:
line_idx += 1
if line_idx >= len(block_lines):
continue
# Ignore beginning and last line: they don't contain anything.
internal_block_code = "\n".join(block_lines[line_idx:-1])
indent = get_indent(block_lines[1])
# Slit the internal block into blocks of indent level 1.
internal_blocks = split_code_in_indented_blocks(internal_block_code, indent_level=indent)
# We have two categories of import key: list or _import_structure[key].append/extend
pattern = _re_direct_key if "_import_structure = {" in block_lines[0] else _re_indirect_key
# Grab the keys, but there is a trap: some lines are empty or just comments.
keys = [(pattern.search(b).groups()[0] if pattern.search(b) is not None else None) for b in internal_blocks]
# We only sort the lines with a key.
keys_to_sort = [(i, key) for i, key in enumerate(keys) if key is not None]
sorted_indices = [x[0] for x in sorted(keys_to_sort, key=lambda x: x[1])]
# We reorder the blocks by leaving empty lines/comments as they were and reorder the rest.
count = 0
reorderded_blocks = []
for i in range(len(internal_blocks)):
if keys[i] is None:
reorderded_blocks.append(internal_blocks[i])
else:
block = sort_objects_in_import(internal_blocks[sorted_indices[count]])
reorderded_blocks.append(block)
count += 1
# And we put our main block back together with its first and last line.
main_blocks[block_idx] = "\n".join(block_lines[:line_idx] + reorderded_blocks + [block_lines[-1]])
if code != "\n".join(main_blocks):
if check_only:
return True
else:
print(f"Overwriting {file}.")
with open(file, "w", encoding="utf-8") as f:
f.write("\n".join(main_blocks))
def sort_imports_in_all_inits(check_only=True):
"""
Sort the imports defined in the `_import_structure` of all inits in the repo.
Args:
check_only (`bool`, *optional*, defaults to `True`): Whether or not to just check (and not auto-fix) the init.
"""
failures = []
for root, _, files in os.walk(PATH_TO_TRANSFORMERS):
if "__init__.py" in files:
result = sort_imports(os.path.join(root, "__init__.py"), check_only=check_only)
if result:
failures = [os.path.join(root, "__init__.py")]
if len(failures) > 0:
raise ValueError(f"Would overwrite {len(failures)} files, run `make style`.")
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--check_only", action="store_true", help="Whether to only check or fix style.")
args = parser.parse_args()
sort_imports_in_all_inits(check_only=args.check_only)
| transformers/utils/custom_init_isort.py/0 | {
"file_path": "transformers/utils/custom_init_isort.py",
"repo_id": "transformers",
"token_count": 5360
} |
import argparse
import os
past_versions_testing = {
"pytorch": {
"1.13": {
"torch": "1.13.1",
"torchvision": "0.14.1",
"torchaudio": "0.13.1",
"python": 3.9,
"cuda": "cu116",
"install": (
"python3 -m pip install --no-cache-dir -U torch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1"
" --extra-index-url https://download.pytorch.org/whl/cu116"
),
"base_image": "nvidia/cuda:11.6.2-cudnn8-devel-ubuntu20.04",
},
"1.12": {
"torch": "1.12.1",
"torchvision": "0.13.1",
"torchaudio": "0.12.1",
"python": 3.9,
"cuda": "cu113",
"install": (
"python3 -m pip install --no-cache-dir -U torch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1"
" --extra-index-url https://download.pytorch.org/whl/cu113"
),
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"1.11": {
"torch": "1.11.0",
"torchvision": "0.12.0",
"torchaudio": "0.11.0",
"python": 3.9,
"cuda": "cu113",
"install": (
"python3 -m pip install --no-cache-dir -U torch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0"
" --extra-index-url https://download.pytorch.org/whl/cu113"
),
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"1.10": {
"torch": "1.10.2",
"torchvision": "0.11.3",
"torchaudio": "0.10.2",
"python": 3.9,
"cuda": "cu113",
"install": (
"python3 -m pip install --no-cache-dir -U torch==1.10.2 torchvision==0.11.3 torchaudio==0.10.2"
" --extra-index-url https://download.pytorch.org/whl/cu113"
),
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
# torchaudio < 0.10 has no CUDA-enabled binary distributions
"1.9": {
"torch": "1.9.1",
"torchvision": "0.10.1",
"torchaudio": "0.9.1",
"python": 3.9,
"cuda": "cu111",
"install": (
"python3 -m pip install --no-cache-dir -U torch==1.9.1 torchvision==0.10.1 torchaudio==0.9.1"
" --extra-index-url https://download.pytorch.org/whl/cu111"
),
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
},
"tensorflow": {
"2.11": {
"tensorflow": "2.11.1",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.11.1",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.10": {
"tensorflow": "2.10.1",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.10.1",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.9": {
"tensorflow": "2.9.3",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.9.3",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.8": {
"tensorflow": "2.8.2",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.8.2",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.7": {
"tensorflow": "2.7.3",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.7.3",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.6": {
"tensorflow": "2.6.5",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.6.5",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
"2.5": {
"tensorflow": "2.5.3",
"install": "python3 -m pip install --no-cache-dir -U tensorflow==2.5.3",
"base_image": "nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04",
},
},
}
if __name__ == "__main__":
parser = argparse.ArgumentParser("Choose the framework and version to install")
parser.add_argument(
"--framework", help="The framework to install. Should be `torch` or `tensorflow`", type=str, required=True
)
parser.add_argument("--version", help="The version of the framework to install.", type=str, required=True)
args = parser.parse_args()
info = past_versions_testing[args.framework][args.version]
os.system(f'echo "export INSTALL_CMD=\'{info["install"]}\'" >> ~/.profile')
print(f'echo "export INSTALL_CMD=\'{info["install"]}\'" >> ~/.profile')
cuda = ""
if args.framework == "pytorch":
cuda = info["cuda"]
os.system(f"echo \"export CUDA='{cuda}'\" >> ~/.profile")
print(f"echo \"export CUDA='{cuda}'\" >> ~/.profile")
| transformers/utils/past_ci_versions.py/0 | {
"file_path": "transformers/utils/past_ci_versions.py",
"repo_id": "transformers",
"token_count": 2774
} |
- sections:
- local: index
title: TRL
- local: installation
title: Installation
- local: quickstart
title: Quickstart
title: Getting started
- sections:
- local: dataset_formats
title: Dataset Formats
- local: how_to_train
title: Training FAQ
- local: logging
title: Understanding Logs
title: Conceptual Guides
- sections:
- local: clis
title: Command Line Interface (CLI)
- local: customization
title: Customizing the Training
- local: reducing_memory_usage
title: Reducing Memory Usage
- local: speeding_up_training
title: Speeding Up Training
- local: use_model
title: Using Trained Models
title: How-to guides
- sections:
- local: deepspeed_integration
title: DeepSpeed
- local: liger_kernel_integration
title: Liger Kernel
- local: peft_integration
title: PEFT
- local: unsloth_integration
title: Unsloth
title: Integrations
- sections:
- local: example_overview
title: Example Overview
- local: community_tutorials
title: Community Tutorials
- local: sentiment_tuning
title: Sentiment Tuning
- local: using_llama_models
title: Training StackLlama
- local: detoxifying_a_lm
title: Detoxifying a Language Model
- local: learning_tools
title: Learning to Use Tools
- local: multi_adapter_rl
title: Multi Adapter RLHF
title: Examples
- sections:
- sections: # Sorted alphabetically
- local: alignprop_trainer
title: AlignProp
- local: bco_trainer
title: BCO
- local: cpo_trainer
title: CPO
- local: ddpo_trainer
title: DDPO
- local: dpo_trainer
title: DPO
- local: online_dpo_trainer
title: Online DPO
- local: gkd_trainer
title: GKD
- local: grpo_trainer
title: GRPO
- local: kto_trainer
title: KTO
- local: nash_md_trainer
title: Nash-MD
- local: orpo_trainer
title: ORPO
- local: ppo_trainer
title: PPO
- local: prm_trainer
title: PRM
- local: reward_trainer
title: Reward
- local: rloo_trainer
title: RLOO
- local: sft_trainer
title: SFT
- local: iterative_sft_trainer
title: Iterative SFT
- local: xpo_trainer
title: XPO
title: Trainers
- local: models
title: Model Classes
- local: best_of_n
title: Best of N Sampling
- local: judges
title: Judges
- local: callbacks
title: Callbacks
- local: data_utils
title: Data Utilities
- local: text_environments
title: Text Environments
- local: script_utils
title: Script Utilities
title: API
| trl/docs/source/_toctree.yml/0 | {
"file_path": "trl/docs/source/_toctree.yml",
"repo_id": "trl",
"token_count": 1007
} |
# Examples of using peft with trl to finetune 8-bit models with Low Rank Adaption (LoRA)
The notebooks and scripts in this examples show how to use Low Rank Adaptation (LoRA) to fine-tune models in a memory efficient manner. Most of PEFT methods supported in peft library but note that some PEFT methods such as Prompt tuning are not supported.
For more information on LoRA, see the [original paper](https://huggingface.co/papers/2106.09685).
Here's an overview of the `peft`-enabled notebooks and scripts in the [trl repository](https://github.com/huggingface/trl/tree/main/examples):
| File | Task | Description | Colab link |
|---|---| --- |
| [`stack_llama/rl_training.py`](https://github.com/huggingface/trl/blob/main/examples/research_projects/stack_llama/scripts/rl_training.py) | RLHF | Distributed fine-tuning of the 7b parameter LLaMA models with a learned reward model and `peft`. | |
| [`stack_llama/reward_modeling.py`](https://github.com/huggingface/trl/blob/main/examples/research_projects/stack_llama/scripts/reward_modeling.py) | Reward Modeling | Distributed training of the 7b parameter LLaMA reward model with `peft`. | |
| [`stack_llama/supervised_finetuning.py`](https://github.com/huggingface/trl/blob/main/examples/research_projects/stack_llama/scripts/supervised_finetuning.py) | SFT | Distributed instruction/supervised fine-tuning of the 7b parameter LLaMA model with `peft`. | |
## Installation
Note: peft is in active development, so we install directly from their Github page.
Peft also relies on the latest version of transformers.
```bash
pip install trl[peft]
pip install bitsandbytes loralib
pip install git+https://github.com/huggingface/transformers.git@main
#optional: wandb
pip install wandb
```
Note: if you don't want to log with `wandb` remove `log_with="wandb"` in the scripts/notebooks. You can also replace it with your favourite experiment tracker that's [supported by `accelerate`](https://huggingface.co/docs/accelerate/usage_guides/tracking).
## How to use it?
Simply declare a `PeftConfig` object in your script and pass it through `.from_pretrained` to load the TRL+PEFT model.
```python
from peft import LoraConfig
from trl import AutoModelForCausalLMWithValueHead
model_id = "edbeeching/gpt-neo-125M-imdb"
lora_config = LoraConfig(
r=16,
lora_alpha=32,
lora_dropout=0.05,
bias="none",
task_type="CAUSAL_LM",
)
model = AutoModelForCausalLMWithValueHead.from_pretrained(
model_id,
peft_config=lora_config,
)
```
And if you want to load your model in 8bit precision:
```python
pretrained_model = AutoModelForCausalLMWithValueHead.from_pretrained(
config.model_name,
load_in_8bit=True,
peft_config=lora_config,
)
```
... or in 4bit precision:
```python
pretrained_model = AutoModelForCausalLMWithValueHead.from_pretrained(
config.model_name,
peft_config=lora_config,
load_in_4bit=True,
)
```
## Launch scripts
The `trl` library is powered by `accelerate`. As such it is best to configure and launch trainings with the following commands:
```bash
accelerate config # will prompt you to define the training configuration
accelerate launch examples/scripts/ppo.py --use_peft # launch`es training
```
## Using `trl` + `peft` and Data Parallelism
You can scale up to as many GPUs as you want, as long as you are able to fit the training process in a single device. The only tweak you need to apply is to load the model as follows:
```python
from peft import LoraConfig
...
lora_config = LoraConfig(
r=16,
lora_alpha=32,
lora_dropout=0.05,
bias="none",
task_type="CAUSAL_LM",
)
pretrained_model = AutoModelForCausalLMWithValueHead.from_pretrained(
config.model_name,
peft_config=lora_config,
)
```
And if you want to load your model in 8bit precision:
```python
pretrained_model = AutoModelForCausalLMWithValueHead.from_pretrained(
config.model_name,
peft_config=lora_config,
load_in_8bit=True,
)
```
... or in 4bit precision:
```python
pretrained_model = AutoModelForCausalLMWithValueHead.from_pretrained(
config.model_name,
peft_config=lora_config,
load_in_4bit=True,
)
```
Finally, make sure that the rewards are computed on correct device as well, for that you can use `ppo_trainer.model.current_device`.
## Naive pipeline parallelism (NPP) for large models (>60B models)
The `trl` library also supports naive pipeline parallelism (NPP) for large models (>60B models). This is a simple way to parallelize the model across multiple GPUs.
This paradigm, termed as "Naive Pipeline Parallelism" (NPP) is a simple way to parallelize the model across multiple GPUs. We load the model and the adapters across multiple GPUs and the activations and gradients will be naively communicated across the GPUs. This supports `int8` models as well as other `dtype` models.
<div style="text-align: center">
<img src="https://huggingface.co/datasets/trl-lib/documentation-images/resolve/main/trl-npp.png">
</div>
### How to use NPP?
Simply load your model with a custom `device_map` argument on the `from_pretrained` to split your model across multiple devices. Check out this [nice tutorial](https://github.com/huggingface/blog/blob/main/accelerate-large-models.md) on how to properly create a `device_map` for your model.
Also make sure to have the `lm_head` module on the first GPU device as it may throw an error if it is not on the first device. As this time of writing, you need to install the `main` branch of `accelerate`: `pip install git+https://github.com/huggingface/accelerate.git@main` and `peft`: `pip install git+https://github.com/huggingface/peft.git@main`.
### Launch scripts
Although `trl` library is powered by `accelerate`, you should run your training script in a single process. Note that we do not support Data Parallelism together with NPP yet.
```bash
python PATH_TO_SCRIPT
```
## Fine-tuning Llama-2 model
You can easily fine-tune Llama2 model using `SFTTrainer` and the official script! For example to fine-tune llama2-7b on the Guanaco dataset, run (tested on a single NVIDIA T4-16GB):
```bash
python trl/scripts/sft.py --output_dir sft_openassistant-guanaco --model_name meta-llama/Llama-2-7b-hf --dataset_name timdettmers/openassistant-guanaco --load_in_4bit --use_peft --per_device_train_batch_size 4 --gradient_accumulation_steps 2
```
| trl/docs/source/peft_integration.md/0 | {
"file_path": "trl/docs/source/peft_integration.md",
"repo_id": "trl",
"token_count": 2079
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from dataclasses import dataclass, field
from typing import Optional
from datasets import load_dataset
from huggingface_hub import ModelCard
from transformers import HfArgumentParser
@dataclass
class ScriptArguments:
r"""
Arguments for the script.
Args:
push_to_hub (`bool`, *optional*, defaults to `False`):
Whether to push the dataset to the Hugging Face Hub.
repo_id (`str`, *optional*, defaults to `"trl-lib/ultrafeedback-prompt"`):
Hugging Face repository ID to push the dataset to.
dataset_num_proc (`int` or `None`, *optional*, defaults to `None`):
Number of workers to use for dataset processing.
"""
push_to_hub: bool = field(
default=False,
metadata={"help": "Whether to push the dataset to the Hugging Face Hub."},
)
repo_id: str = field(
default="trl-lib/ultrafeedback-prompt",
metadata={"help": "Hugging Face repository ID to push the dataset to."},
)
dataset_num_proc: Optional[int] = field(
default=None,
metadata={"help": "Number of workers to use for dataset processing."},
)
def to_unpaired_preference(example):
prompt = [{"role": "user", "content": example["instruction"]}]
return {"prompt": prompt}
def drop_long_prompt(example):
if len(example["prompt"][0]["content"]) > 512:
return False
else:
return True
model_card = ModelCard("""
---
tags: [trl]
---
# UltraFeedback - Prompts Dataset
## Summary
The UltraFeedback - Prompts dataset is a processed version of the [UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) dataset for model evaluation on specific aspects like helpfulness, honesty, and instruction-following.
## Data Structure
- **Format**: [Conversational](https://huggingface.co/docs/trl/main/dataset_formats#conversational)
- **Type**: [Prompt-only](https://huggingface.co/docs/trl/main/dataset_formats#prompt-only)
Column:
- `"prompt"`: The input question or instruction provided to the model.
## Generation script
The script used to generate this dataset can be found [here](https://github.com/huggingface/trl/blob/main/examples/datasets/ultrafeedback-prompt.py).
""")
if __name__ == "__main__":
parser = HfArgumentParser(ScriptArguments)
script_args = parser.parse_args_into_dataclasses()[0]
dataset = load_dataset("openbmb/UltraFeedback", split="train")
dataset = dataset.map(
to_unpaired_preference,
remove_columns=["source", "instruction", "models", "completions", "correct_answers", "incorrect_answers"],
num_proc=script_args.dataset_num_proc,
)
dataset = dataset.filter(drop_long_prompt)
dataset = dataset.train_test_split(test_size=0.05, seed=42)
if script_args.push_to_hub:
dataset.push_to_hub(script_args.repo_id)
model_card.push_to_hub(script_args.repo_id, repo_type="dataset")
| trl/examples/datasets/ultrafeedback-prompt.py/0 | {
"file_path": "trl/examples/datasets/ultrafeedback-prompt.py",
"repo_id": "trl",
"token_count": 1243
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import re
import numpy as np
import torch
from transformers import AutoTokenizer, load_tool
from trl import AutoModelForCausalLMWithValueHead, PPOConfig, PPOTrainer, TextEnvironment
def generate_data(n):
"""Generate random arithmetic tasks and answers."""
tasks, answers = [], []
for _ in range(n):
a = np.random.randint(0, 50)
b = np.random.randint(0, 50)
op = np.random.choice(["-", "+", "*"])
tasks.append(f"\n\nWhat is {a} {op} {b}?")
if op == "-":
answers.append(a - b)
elif op == "+":
answers.append(a + b)
else:
answers.append(a * b)
return tasks, answers
def exact_match_reward(responses, answers=None):
"""Reward if generated response contains correct answer."""
rewards = []
pattern = r"Result\s*=\s*(-?\d+(?:\.\d+)?)\s*<submit>" # generated by chatGPT
for response, answer in zip(responses, answers):
reward = 0.0
predicted_number = None
match_pattern = re.findall(pattern, response)
if match_pattern:
predicted_number = float(match_pattern[0])
if predicted_number is not None:
if np.abs(predicted_number - answer) < 0.01:
reward += 1.0
rewards.append(torch.tensor(reward))
return rewards
# set up models
model_id = "gpt2"
model = AutoModelForCausalLMWithValueHead.from_pretrained(model_id)
ref_model = AutoModelForCausalLMWithValueHead.from_pretrained(model_id)
tokenizer = AutoTokenizer.from_pretrained(model_id)
tokenizer.pad_token = tokenizer.eos_token
# system prompt
prompt = """\
What is 13-3?
<request><SimpleCalculatorTool>13-3<call>10.0<response>
Result=10<submit>
What is 4*3?
<request><SimpleCalculatorTool>4*3<call>12.0<response>
Result=12<submit>"""
generation_kwargs = {
"min_length": -1,
"top_k": 0.0,
"top_p": 1.0,
"do_sample": True,
"pad_token_id": tokenizer.eos_token_id,
"eos_token_id": -1,
"max_new_tokens": 32,
}
# trainer
ppo_config = PPOConfig(
batch_size=256,
learning_rate=1.41e-5,
mini_batch_size=64,
log_with="wandb",
)
ppo_trainer = PPOTrainer(ppo_config, model, ref_model, tokenizer)
# text env
text_env = TextEnvironment(
model,
tokenizer,
{"SimpleCalculatorTool": load_tool("ybelkada/simple-calculator")},
exact_match_reward,
prompt,
generation_kwargs=generation_kwargs,
)
# main training loop
for _step in range(100):
tasks, answers = generate_data(ppo_config.batch_size)
queries, responses, masks, rewards, histories = text_env.run(tasks, answers=answers)
train_stats = ppo_trainer.step(queries, responses, rewards, masks)
response_texts = [tokenizer.decode(response) for response in responses]
query_texts = [tokenizer.decode(query) for query in queries]
texts = {"query": [qt.split("<submit>")[-1].strip() for qt in query_texts], "response": response_texts}
ppo_trainer.log_stats(train_stats, texts, rewards, columns_to_log=["query", "response", "answer"])
ppo_trainer.save_pretrained(model_id + "-calculator")
| trl/examples/research_projects/tools/calculator.py/0 | {
"file_path": "trl/examples/research_projects/tools/calculator.py",
"repo_id": "trl",
"token_count": 1430
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Run the KTO training script with the commands below. In general, the optimal configuration for KTO will be similar to that of DPO.
# Full training:
python trl/scripts/kto.py \
--dataset_name trl-lib/kto-mix-14k \
--model_name_or_path=trl-lib/qwen1.5-1.8b-sft \
--per_device_train_batch_size 16 \
--num_train_epochs 1 \
--learning_rate 5e-7 \
--lr_scheduler_type=cosine \
--gradient_accumulation_steps 1 \
--logging_steps 10 \
--eval_steps 500 \
--output_dir=kto-aligned-model \
--warmup_ratio 0.1 \
--report_to wandb \
--bf16 \
--logging_first_step
# QLoRA:
python trl/scripts/kto.py \
--dataset_name trl-lib/kto-mix-14k \
--model_name_or_path=trl-lib/qwen1.5-1.8b-sft \
--per_device_train_batch_size 8 \
--num_train_epochs 1 \
--learning_rate 5e-7 \
--lr_scheduler_type=cosine \
--gradient_accumulation_steps 1 \
--logging_steps 10 \
--eval_steps 500 \
--output_dir=kto-aligned-model-lora \
--warmup_ratio 0.1 \
--report_to wandb \
--bf16 \
--logging_first_step \
--use_peft \
--load_in_4bit \
--lora_target_modules=all-linear \
--lora_r=16 \
--lora_alpha=16
"""
from datasets import load_dataset
from transformers import AutoModelForCausalLM, AutoTokenizer, HfArgumentParser
from trl import (
KTOConfig,
KTOTrainer,
ModelConfig,
ScriptArguments,
get_peft_config,
setup_chat_format,
)
if __name__ == "__main__":
parser = HfArgumentParser((ScriptArguments, KTOConfig, ModelConfig))
script_args, training_args, model_args = parser.parse_args_into_dataclasses()
# Load a pretrained model
model = AutoModelForCausalLM.from_pretrained(
model_args.model_name_or_path, trust_remote_code=model_args.trust_remote_code
)
ref_model = AutoModelForCausalLM.from_pretrained(
model_args.model_name_or_path, trust_remote_code=model_args.trust_remote_code
)
tokenizer = AutoTokenizer.from_pretrained(
model_args.model_name_or_path, trust_remote_code=model_args.trust_remote_code
)
if tokenizer.pad_token is None:
tokenizer.pad_token = tokenizer.eos_token
# If we are aligning a base model, we use ChatML as the default template
if tokenizer.chat_template is None:
model, tokenizer = setup_chat_format(model, tokenizer)
# Load the dataset
dataset = load_dataset(script_args.dataset_name, name=script_args.dataset_config)
# Initialize the KTO trainer
trainer = KTOTrainer(
model,
ref_model,
args=training_args,
train_dataset=dataset[script_args.dataset_train_split],
eval_dataset=dataset[script_args.dataset_test_split] if training_args.eval_strategy != "no" else None,
processing_class=tokenizer,
peft_config=get_peft_config(model_args),
)
# Train and push the model to the Hub
trainer.train()
# Save and push to hub
trainer.save_model(training_args.output_dir)
if training_args.push_to_hub:
trainer.push_to_hub(dataset_name=script_args.dataset_name)
| trl/examples/scripts/kto.py/0 | {
"file_path": "trl/examples/scripts/kto.py",
"repo_id": "trl",
"token_count": 1461
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import subprocess
import sys
from datetime import datetime
COPYRIGHT_HEADER = f"""# Copyright {datetime.now().year} The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
def get_tracked_python_files():
"""Get a list of all tracked Python files using git."""
try:
# Get the list of all tracked files from Git
result = subprocess.run(["git", "ls-files"], stdout=subprocess.PIPE, text=True, check=True)
# Split the result by lines to get individual file paths
files = result.stdout.splitlines()
# Filter only Python files
py_files = [f for f in files if f.endswith(".py")]
return py_files
except subprocess.CalledProcessError as e:
print(f"Error fetching tracked files: {e}")
return []
def check_and_add_copyright(file_path):
"""Check if the file contains a copyright notice, and add it if missing."""
if not os.path.isfile(file_path):
print(f"[SKIP] {file_path} does not exist.")
return
with open(file_path, encoding="utf-8") as f:
content = f.readlines()
# Check if the exact copyright header exists
if "".join(content).startswith(COPYRIGHT_HEADER):
return True
# If no copyright notice was found, prepend the header
print(f"[MODIFY] Adding copyright to {file_path}.")
with open(file_path, "w", encoding="utf-8") as f:
# Write the copyright header followed by the original content
f.write(COPYRIGHT_HEADER + "\n" + "".join(content))
return False
def main():
"""Main function to check and add copyright for all tracked Python files."""
py_files = get_tracked_python_files()
if not py_files:
print("No Python files are tracked in the repository.")
return
print(f"Checking {len(py_files)} Python files for copyright notice...")
have_copyright = [check_and_add_copyright(file_path) for file_path in py_files]
if not all(have_copyright):
print("❌ Some files were missing the required copyright and have been updated.")
sys.exit(1)
else:
print("✅ All files have the required copyright.")
sys.exit(0)
if __name__ == "__main__":
main()
| trl/scripts/add_copyrights.py/0 | {
"file_path": "trl/scripts/add_copyrights.py",
"repo_id": "trl",
"token_count": 1106
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import sys
import tempfile
import unittest
from io import StringIO
from unittest.mock import patch
@unittest.skipIf(
sys.version_info < (3, 10),
"Transformers' generation codebase uses a Python >3.10 syntax (`str | None`), which seems to cause the CLI tests "
"to fail on Python <3.10.", # let's say it's a known issue, but not expected to be fixed, because too niche
)
class TestCLI(unittest.TestCase):
def test_dpo(self):
from trl.cli import main
with tempfile.TemporaryDirectory() as tmp_dir: # Create a temporary directory
command = f"trl dpo --output_dir {tmp_dir} --model_name_or_path trl-internal-testing/tiny-Qwen2ForCausalLM-2.5 --dataset_name trl-internal-testing/zen --dataset_config standard_preference --report_to none"
with patch("sys.argv", command.split(" ")):
main()
@patch("sys.stdout", new_callable=StringIO)
def test_env(self, mock_stdout):
from trl.cli import main
command = "trl env"
with patch("sys.argv", command.split(" ")):
main()
self.assertIn("TRL version: ", mock_stdout.getvalue().strip())
def test_grpo(self):
from trl.cli import main
with tempfile.TemporaryDirectory() as tmp_dir: # Create a temporary directory
command = f"trl grpo --output_dir {tmp_dir} --model_name_or_path trl-internal-testing/tiny-Qwen2ForCausalLM-2.5 --reward_model_name_or_path trl-internal-testing/tiny-Qwen2ForSequenceClassification-2.5 --dataset_name trl-internal-testing/zen --dataset_config standard_prompt_only --num_generations 4 --max_completion_length 32 --report_to none"
with patch("sys.argv", command.split(" ")):
main()
def test_kto(self):
from trl.cli import main
with tempfile.TemporaryDirectory() as tmp_dir: # Create a temporary directory
command = f"trl kto --output_dir {tmp_dir} --model_name_or_path trl-internal-testing/tiny-Qwen2ForCausalLM-2.5 --dataset_name trl-internal-testing/zen --dataset_config standard_unpaired_preference --report_to none"
with patch("sys.argv", command.split(" ")):
main()
def test_sft(self):
from trl.cli import main
with tempfile.TemporaryDirectory() as tmp_dir: # Create a temporary directory
command = f"trl sft --output_dir {tmp_dir} --model_name_or_path trl-internal-testing/tiny-Qwen2ForCausalLM-2.5 --dataset_name trl-internal-testing/zen --dataset_config standard_language_modeling --report_to none"
with patch("sys.argv", command.split(" ")):
main()
if __name__ == "__main__":
unittest.main()
| trl/tests/test_cli.py/0 | {
"file_path": "trl/tests/test_cli.py",
"repo_id": "trl",
"token_count": 1240
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
import torch
from transformers import AutoModelForCausalLM, GenerationConfig
from trl.models.modeling_base import GeometricMixtureWrapper, create_reference_model
class TestGeometricMixtureWrapper(unittest.TestCase):
def setUp(self):
model_id = "trl-internal-testing/tiny-Qwen2ForCausalLM-2.5"
self.model = AutoModelForCausalLM.from_pretrained(model_id)
self.ref_model = create_reference_model(self.model)
self.generation_config = GenerationConfig.from_pretrained(model_id)
self.mixture_coef = 0.5
self.wrapper = GeometricMixtureWrapper(
self.model, self.ref_model, self.generation_config, mixture_coef=self.mixture_coef
)
def test_forward(self):
input_ids = torch.tensor([[1, 2, 3, 4, 5]])
attention_mask = torch.ones_like(input_ids)
output = self.wrapper(input_ids=input_ids, attention_mask=attention_mask)
self.assertIsNotNone(output)
self.assertTrue(hasattr(output, "logits"))
self.assertEqual(output.logits.shape, (1, 5, self.model.config.vocab_size))
def test_mixture_coefficient(self):
input_ids = torch.tensor([[1, 2, 3, 4, 5]])
attention_mask = torch.ones_like(input_ids)
with torch.no_grad():
model_output = self.model(input_ids=input_ids, attention_mask=attention_mask)
ref_model_output = self.ref_model(input_ids=input_ids, attention_mask=attention_mask)
wrapper_output = self.wrapper(input_ids=input_ids, attention_mask=attention_mask)
expected_logits = torch.nn.functional.log_softmax(
self.mixture_coef * ref_model_output.logits + (1 - self.mixture_coef) * model_output.logits, dim=-1
)
self.assertTrue(torch.allclose(wrapper_output.logits, expected_logits, atol=1e-5))
def test_prepare_inputs_for_generation(self):
input_ids = torch.tensor([[1, 2, 3, 4, 5]])
attention_mask = torch.ones_like(input_ids)
inputs = self.wrapper.prepare_inputs_for_generation(input_ids, attention_mask=attention_mask, use_cache=True)
self.assertIn("input_ids", inputs)
self.assertIn("attention_mask", inputs)
self.assertFalse(inputs.get("use_cache", False))
| trl/tests/test_modeling_geometric_mixture_wrapper.py/0 | {
"file_path": "trl/tests/test_modeling_geometric_mixture_wrapper.py",
"repo_id": "trl",
"token_count": 1095
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import random
import unittest
from transformers import is_bitsandbytes_available, is_comet_available, is_sklearn_available, is_wandb_available
from trl import BaseBinaryJudge, BasePairwiseJudge, is_diffusers_available, is_llm_blender_available
from trl.import_utils import is_mergekit_available
# transformers.testing_utils contains a require_bitsandbytes function, but relies on pytest markers which we don't use
# in our test suite. We therefore need to implement our own version of this function.
def require_bitsandbytes(test_case):
"""
Decorator marking a test that requires bitsandbytes. Skips the test if bitsandbytes is not available.
"""
return unittest.skipUnless(is_bitsandbytes_available(), "test requires bitsandbytes")(test_case)
def require_diffusers(test_case):
"""
Decorator marking a test that requires diffusers. Skips the test if diffusers is not available.
"""
return unittest.skipUnless(is_diffusers_available(), "test requires diffusers")(test_case)
def require_llm_blender(test_case):
"""
Decorator marking a test that requires llm-blender. Skips the test if llm-blender is not available.
"""
return unittest.skipUnless(is_llm_blender_available(), "test requires llm-blender")(test_case)
def require_mergekit(test_case):
"""
Decorator marking a test that requires mergekit. Skips the test if mergekit is not available.
"""
return unittest.skipUnless(is_mergekit_available(), "test requires mergekit")(test_case)
def require_no_wandb(test_case):
"""
Decorator marking a test that requires no wandb. Skips the test if wandb is available.
"""
return unittest.skipUnless(not is_wandb_available(), "test requires no wandb")(test_case)
def require_sklearn(test_case):
"""
Decorator marking a test that requires sklearn. Skips the test if sklearn is not available.
"""
return unittest.skipUnless(is_sklearn_available(), "test requires sklearn")(test_case)
def require_comet(test_case):
"""
Decorator marking a test that requires Comet. Skips the test if Comet is not available.
"""
return unittest.skipUnless(is_comet_available(), "test requires comet_ml")(test_case)
class RandomBinaryJudge(BaseBinaryJudge):
"""
Random binary judge, for testing purposes.
"""
def judge(self, prompts, completions, gold_completions=None, shuffle_order=True):
return [random.choice([0, 1, -1]) for _ in range(len(prompts))]
class RandomPairwiseJudge(BasePairwiseJudge):
"""
Random pairwise judge, for testing purposes.
"""
def judge(self, prompts, completions, shuffle_order=True, return_scores=False):
if not return_scores:
return [random.randint(0, len(completion) - 1) for completion in completions]
else:
return [random.random() for _ in range(len(prompts))]
| trl/tests/testing_utils.py/0 | {
"file_path": "trl/tests/testing_utils.py",
"repo_id": "trl",
"token_count": 1120
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
import torch.nn as nn
from transformers import AutoModelForCausalLM, AutoModelForSeq2SeqLM, is_torch_npu_available, is_torch_xpu_available
from .modeling_base import PreTrainedModelWrapper
class ValueHead(nn.Module):
r"""
The ValueHead class implements a head for GPT2 that returns a scalar for each output token.
"""
def __init__(self, config, **kwargs):
super().__init__()
if not hasattr(config, "summary_dropout_prob"):
summary_dropout_prob = kwargs.pop("summary_dropout_prob", 0.1)
else:
summary_dropout_prob = config.summary_dropout_prob
self.dropout = nn.Dropout(summary_dropout_prob) if summary_dropout_prob else nn.Identity()
# some models such as OPT have a projection layer before the word embeddings - e.g. OPT-350m
if hasattr(config, "hidden_size"):
hidden_size = config.hidden_size
if hasattr(config, "word_embed_proj_dim"):
hidden_size = config.word_embed_proj_dim
elif hasattr(config, "is_encoder_decoder"):
if config.is_encoder_decoder and hasattr(config, "decoder"):
if hasattr(config.decoder, "hidden_size"):
hidden_size = config.decoder.hidden_size
self.summary = nn.Linear(hidden_size, 1)
self.flatten = nn.Flatten()
def forward(self, hidden_states):
output = self.dropout(hidden_states)
# For now force upcast in fp32 if needed. Let's keep the
# output in fp32 for numerical stability.
if output.dtype != self.summary.weight.dtype:
output = output.to(self.summary.weight.dtype)
output = self.summary(output)
return output
class AutoModelForCausalLMWithValueHead(PreTrainedModelWrapper):
r"""
An autoregressive model with a value head in addition to the language model head.
This class inherits from `~trl.PreTrainedModelWrapper` and wraps a
`transformers.PreTrainedModel` class. The wrapper class supports classic functions
such as `from_pretrained`, `push_to_hub` and `generate`. To call a method of the wrapped
model, simply manipulate the `pretrained_model` attribute of this class.
Class attributes:
- **transformers_parent_class** (`transformers.PreTrainedModel`) -- The parent class of the wrapped model. This
should be set to `transformers.AutoModelForCausalLM` for this class.
- **supported_args** (`tuple`) -- A tuple of strings that are used to identify the arguments that are supported
by the `ValueHead` class. Currently, the supported args are:
- **summary_dropout_prob** (`float`, `optional`, defaults to `None`) -- The dropout probability for the
`ValueHead` class.
- **v_head_initializer_range** (`float`, `optional`, defaults to `0.2`) -- The initializer range for the
`ValueHead` if a specific initialization strategy is selected.
- **v_head_init_strategy** (`str`, `optional`, defaults to `None`) -- The initialization strategy for the
`ValueHead`. Currently, the supported strategies are:
- **`None`** -- Initializes the weights of the `ValueHead` with a random distribution. This is the default
strategy.
- **"normal"** -- Initializes the weights of the `ValueHead` with a normal distribution.
"""
transformers_parent_class = AutoModelForCausalLM
supported_args = (
"summary_dropout_prob",
"v_head_initializer_range",
"v_head_init_strategy",
)
def __init__(self, pretrained_model, **kwargs):
r"""
Initializes the model.
Args:
pretrained_model (`transformers.PreTrainedModel`):
The model to wrap. It should be a causal language model such as GPT2.
or any model mapped inside the `AutoModelForCausalLM` class.
kwargs (`dict`, `optional`):
Additional keyword arguments, that are passed to the `ValueHead` class.
"""
super().__init__(pretrained_model, **kwargs)
v_head_kwargs, _, _ = self._split_kwargs(kwargs)
self.v_head = ValueHead(self.pretrained_model.config, **v_head_kwargs)
self._init_weights(**v_head_kwargs)
def _init_weights(self, **kwargs):
r"""
Initializes the weights of the value head. The default initialization strategy is random.
Users can pass a different initialization strategy by passing the `v_head_init_strategy` argument
when calling `.from_pretrained`. Supported strategies are:
- `normal`: initializes the weights with a normal distribution.
Args:
**kwargs (`dict`, `optional`):
Additional keyword arguments, that are passed to the `ValueHead` class. These arguments
can contain the `v_head_init_strategy` argument as well as the `v_head_initializer_range`
argument.
"""
initializer_range = kwargs.pop("v_head_initializer_range", 0.2)
# random init by default
init_strategy = kwargs.pop("v_head_init_strategy", None)
if init_strategy is None:
# do nothing
pass
elif init_strategy == "normal":
self.v_head.summary.weight.data.normal_(mean=0.0, std=initializer_range)
self.v_head.summary.bias.data.zero_()
def forward(
self,
input_ids=None,
past_key_values=None,
attention_mask=None,
return_past_key_values=False,
**kwargs,
):
r"""
Applies a forward pass to the wrapped model and returns the logits of the value head.
Args:
input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary.
past_key_values (`tuple(tuple(torch.FloatTensor))`, `optional`):
Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model
(see `past_key_values` input) to speed up sequential decoding.
attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, `optional`):
Mask to avoid performing attention on padding token indices. Mask values selected in ``[0, 1]``:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
return_past_key_values (bool): A flag indicating if the computed hidden-states should be returned.
kwargs (`dict`, `optional`):
Additional keyword arguments, that are passed to the wrapped model.
"""
kwargs["output_hidden_states"] = True # this had already been set in the LORA / PEFT examples
kwargs["past_key_values"] = past_key_values
if self.is_peft_model and self.pretrained_model.active_peft_config.peft_type == "PREFIX_TUNING":
kwargs.pop("past_key_values")
base_model_output = self.pretrained_model(
input_ids=input_ids,
attention_mask=attention_mask,
**kwargs,
)
last_hidden_state = base_model_output.hidden_states[-1]
lm_logits = base_model_output.logits
loss = base_model_output.loss
if last_hidden_state.device != self.v_head.summary.weight.device:
last_hidden_state = last_hidden_state.to(self.v_head.summary.weight.device)
value = self.v_head(last_hidden_state).squeeze(-1)
# force upcast in fp32 if logits are in half-precision
if lm_logits.dtype != torch.float32:
lm_logits = lm_logits.float()
if return_past_key_values:
return (lm_logits, loss, value, base_model_output.past_key_values)
else:
return (lm_logits, loss, value)
def generate(self, *args, **kwargs):
r"""
A simple wrapper around the `generate` method of the wrapped model.
Please refer to the [`generate`](https://huggingface.co/docs/transformers/internal/generation_utils)
method of the wrapped model for more information about the supported arguments.
Args:
*args (`list`, *optional*):
Positional arguments passed to the `generate` method of the wrapped model.
**kwargs (`dict`, *optional*):
Keyword arguments passed to the `generate` method of the wrapped model.
"""
return self.pretrained_model.generate(*args, **kwargs)
def state_dict(self, *args, **kwargs):
r"""
Returns the state dictionary of the model. We add the state dictionary of the value head
to the state dictionary of the wrapped model by prepending the key with `v_head.`.
"""
if not self.is_peft_model:
pretrained_model_state_dict = self.pretrained_model.state_dict(*args, **kwargs)
else:
# if it is a peft model, only save the v_head
pretrained_model_state_dict = {}
v_head_state_dict = self.v_head.state_dict(*args, **kwargs)
for k, v in v_head_state_dict.items():
pretrained_model_state_dict[f"v_head.{k}"] = v
return pretrained_model_state_dict
def push_to_hub(self, *args, **kwargs):
self.pretrained_model.v_head = self.v_head
return self.pretrained_model.push_to_hub(*args, **kwargs)
def post_init(self, state_dict):
r"""
We add the state dictionary of the value head to the state dictionary of the wrapped model
by prepending the key with `v_head.`. This function removes the `v_head.` prefix from the
keys of the value head state dictionary.
"""
for k in list(state_dict.keys()):
if "v_head." in k:
state_dict[k.replace("v_head.", "")] = state_dict.pop(k)
self.v_head.load_state_dict(state_dict, strict=False)
del state_dict
if hasattr(self.pretrained_model, "hf_device_map"):
if (
"cpu" in self.pretrained_model.hf_device_map.values()
or "disk" in self.pretrained_model.hf_device_map.values()
):
raise ValueError(
"The model is offloaded on CPU or disk - CPU & disk offloading is not supported for ValueHead models."
)
first_device = list(set(self.pretrained_model.hf_device_map.values()))[0]
if isinstance(first_device, int):
if is_torch_npu_available():
first_device = f"npu:{first_device}"
elif is_torch_xpu_available():
first_device = f"xpu:{first_device}"
else:
first_device = f"cuda:{first_device}"
self.v_head = self.v_head.to(first_device)
def set_device_hook(module, input, outputs):
new_output = ()
for output in outputs:
if isinstance(output, torch.Tensor):
new_output += (output.to(first_device),)
else:
new_output += (output,)
return new_output
self.register_forward_hook(set_device_hook)
self.is_sequential_parallel = True
class AutoModelForSeq2SeqLMWithValueHead(PreTrainedModelWrapper):
r"""
A seq2seq model with a value head in addition to the language model head.
This class inherits from `~trl.PreTrainedModelWrapper` and wraps a
`transformers.PreTrainedModel` class. The wrapper class supports classic functions
such as `from_pretrained` and `push_to_hub` and also provides some additional
functionalities such as `generate`.
Args:
pretrained_model (`transformers.PreTrainedModel`):
The model to wrap. It should be a causal language model such as GPT2.
or any model mapped inside the `AutoModelForSeq2SeqLM` class.
kwargs:
Additional keyword arguments passed along to the `ValueHead` class.
"""
transformers_parent_class = AutoModelForSeq2SeqLM
lm_head_namings = ["lm_head", "embed_out", "output_projection"]
supported_args = (
"summary_dropout_prob",
"v_head_initializer_range",
"v_head_init_strategy",
)
def __init__(self, pretrained_model, **kwargs):
super().__init__(pretrained_model, **kwargs)
v_head_kwargs, _, _ = self._split_kwargs(kwargs)
self.is_encoder_decoder = True
if not self._has_lm_head():
raise ValueError("The model does not have a language model head, please use a model that has one.")
self.v_head = ValueHead(self.pretrained_model.config, **v_head_kwargs)
self._init_weights(**v_head_kwargs)
def _has_lm_head(self):
# check module names of all modules inside `pretrained_model` to find the language model head
for name, _module in self.pretrained_model.named_modules():
if any(attribute in name for attribute in self.lm_head_namings):
return True
return False
def post_init(self, state_dict):
r"""
We add the state dictionary of the value head to the state dictionary of the wrapped model
by prepending the key with `v_head.`. This function removes the `v_head.` prefix from the
keys of the value head state dictionary.
"""
for k in list(state_dict.keys()):
if "v_head." in k:
state_dict[k.replace("v_head.", "")] = state_dict.pop(k)
self.v_head.load_state_dict(state_dict, strict=False)
del state_dict
if hasattr(self.pretrained_model, "hf_device_map"):
if (
"cpu" in self.pretrained_model.hf_device_map.values()
or "disk" in self.pretrained_model.hf_device_map.values()
):
raise ValueError(
"The model is offloaded on CPU or disk - CPU & disk offloading is not supported for ValueHead models."
)
# get the lm_head device
for name, module in self.pretrained_model.named_modules():
if any(attribute in name for attribute in self.lm_head_namings):
lm_head_device = module.weight.device
break
# put v_head on the same device as the lm_head to avoid issues
self.v_head = self.v_head.to(lm_head_device)
def set_device_hook(module, input, outputs):
r"""
A hook that sets the device of the output of the model to the device of the first
parameter of the model.
Args:
module (`nn.Module`):
The module to which the hook is attached.
input (`tuple`):
The input to the module.
outputs (`tuple`):
The output of the module.
"""
new_output = ()
for output in outputs:
if isinstance(output, torch.Tensor):
new_output += (output.to(lm_head_device),)
else:
new_output += (output,)
return new_output
self.register_forward_hook(set_device_hook)
self.is_sequential_parallel = True
def state_dict(self, *args, **kwargs):
r"""
Returns the state dictionary of the model. We add the state dictionary of the value head
to the state dictionary of the wrapped model by prepending the key with `v_head.`.
"""
if not self.is_peft_model:
pretrained_model_state_dict = self.pretrained_model.state_dict(*args, **kwargs)
else:
# if it is a peft model, only save the v_head
pretrained_model_state_dict = {}
v_head_state_dict = self.v_head.state_dict(*args, **kwargs)
for k, v in v_head_state_dict.items():
pretrained_model_state_dict[f"v_head.{k}"] = v
return pretrained_model_state_dict
def push_to_hub(self, *args, **kwargs):
self.pretrained_model.v_head = self.v_head
return self.pretrained_model.push_to_hub(*args, **kwargs)
def _init_weights(self, **kwargs):
r"""
We initialize the weights of the value head.
"""
initializer_range = kwargs.pop("v_head_initializer_range", 0.2)
# random init by default
init_strategy = kwargs.pop("v_head_init_strategy", None)
if init_strategy is None:
# do nothing
pass
elif init_strategy == "normal":
self.v_head.summary.weight.data.normal_(mean=0.0, std=initializer_range)
self.v_head.summary.bias.data.zero_()
def forward(
self,
input_ids=None,
past_key_values=None,
attention_mask=None,
return_past_key_values=False,
**kwargs,
):
kwargs["past_key_values"] = past_key_values
if self.is_peft_model and self.pretrained_model.active_peft_config.peft_type == "PREFIX_TUNING":
kwargs.pop("past_key_values")
base_model_output = self.pretrained_model(
input_ids=input_ids,
attention_mask=attention_mask,
output_hidden_states=True, # We force the model to output hidden states
**kwargs,
)
last_hidden_state = base_model_output.decoder_hidden_states[-1]
lm_logits = base_model_output.logits
loss = base_model_output.loss
value = self.v_head(last_hidden_state).squeeze(-1)
# force upcast in fp32 if logits are in half-precision
if lm_logits.dtype != torch.float32:
lm_logits = lm_logits.float()
if return_past_key_values:
return (lm_logits, loss, value, base_model_output.past_key_values)
else:
return (lm_logits, loss, value)
def generate(self, *args, **kwargs):
r"""
We call `generate` on the wrapped model.
"""
return self.pretrained_model.generate(*args, **kwargs)
| trl/trl/models/modeling_value_head.py/0 | {
"file_path": "trl/trl/models/modeling_value_head.py",
"repo_id": "trl",
"token_count": 8204
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import inspect
import os
import random
import textwrap
import warnings
from collections import defaultdict
from contextlib import contextmanager, nullcontext
from copy import deepcopy
from operator import itemgetter
from typing import TYPE_CHECKING, Any, Callable, Literal, Optional, Union
import numpy as np
import pandas as pd
import torch
import torch.amp as amp
import torch.nn as nn
import torch.nn.functional as F
import transformers
from accelerate import PartialState
from accelerate.utils import is_deepspeed_available, tqdm
from datasets import Dataset
from packaging import version
from torch.utils.data import DataLoader, SequentialSampler
from transformers import (
AutoModelForCausalLM,
BaseImageProcessor,
DataCollator,
FeatureExtractionMixin,
PreTrainedModel,
PreTrainedTokenizerBase,
ProcessorMixin,
Trainer,
TrainingArguments,
is_comet_available,
is_sklearn_available,
is_wandb_available,
)
from transformers.trainer_callback import TrainerCallback
from transformers.trainer_utils import EvalLoopOutput, has_length
from transformers.utils import is_peft_available
from ..data_utils import maybe_apply_chat_template
from ..models import PreTrainedModelWrapper, create_reference_model
from .bco_config import BCOConfig
from .utils import (
DPODataCollatorWithPadding,
RunningMoments,
disable_dropout_in_model,
generate_model_card,
get_comet_experiment_url,
log_table_to_comet_experiment,
pad_to_length,
peft_module_casting_to_bf16,
selective_log_softmax,
)
if is_peft_available():
from peft import PeftModel, get_peft_model, prepare_model_for_kbit_training
if is_wandb_available():
import wandb
if is_sklearn_available():
from sklearn.linear_model import LogisticRegression
if is_deepspeed_available():
import deepspeed
if TYPE_CHECKING:
from transformers import PreTrainedModel, PreTrainedTokenizer
RUNNING_NAME = "running.json"
CLF_NAME = "clf.pt"
def _tokenize(
batch: dict[str, list[Any]],
tokenizer: "PreTrainedTokenizer",
embedding_tokenizer: Optional["PreTrainedTokenizer"] = None,
) -> dict[str, list[Any]]:
"""Tokenize a batch from a BCO specific dataset."""
prompt_tokenized = tokenizer(batch["prompt"], add_special_tokens=False)
prompt_input_ids = prompt_tokenized["input_ids"]
prompt_attention_mask = prompt_tokenized["attention_mask"]
prompt_and_completion = [prompt + completion for prompt, completion in zip(batch["prompt"], batch["completion"])]
full_tokenized = tokenizer(prompt_and_completion, add_special_tokens=False)
full_input_ids = full_tokenized["input_ids"]
full_attention_mask = full_tokenized["attention_mask"]
answer_input_ids = [f[len(p) :] for f, p in zip(full_input_ids, prompt_input_ids)]
answer_attention_mask = [f[len(p) :] for f, p in zip(full_attention_mask, prompt_attention_mask)]
# Concat tokens to form `enc(a) + enc(a + b)[len(enc(a)):]`
full_concat_input_ids = [np.concatenate([p, a]) for p, a in zip(prompt_input_ids, answer_input_ids)]
# Prepare input tokens for token by token comparison
full_input_ids = [np.array(f) for f in full_input_ids]
for full, concat in zip(full_input_ids, full_concat_input_ids):
if len(full) != len(concat):
raise ValueError(
"The elements in 'full_input_ids' and 'full_concat_input_ids' must have the same pairwise length."
)
# On some tokenizers, like Llama-2 tokenizer, there are occasions where tokens
# can be merged together when tokenizing prompt+answer. This could result
# on the last token from the prompt being different when tokenized on its own
# vs when done as prompt+answer.
response_token_ids_start_idx = [len(p) for p in prompt_input_ids]
# If tokenized prompt is different than both prompt+answer, then it means the
# last token has changed due to merging.
for idx, (p, f, r) in enumerate(zip(prompt_input_ids, full_input_ids, response_token_ids_start_idx)):
if not np.array_equal(p, f[:r]):
response_token_ids_start_idx[idx] -= 1
prompt_input_ids = [f[:r] for f, r in zip(full_input_ids, response_token_ids_start_idx)]
prompt_attention_mask = [f[:r] for f, r in zip(full_attention_mask, response_token_ids_start_idx)]
for p, m in zip(prompt_input_ids, prompt_attention_mask):
if len(p) != len(m):
raise ValueError("Prompt input ids and attention mask should have the same length.")
answer_input_ids = [f[r:] for f, r in zip(full_input_ids, response_token_ids_start_idx)]
answer_attention_mask = [f[r:] for f, r in zip(full_attention_mask, response_token_ids_start_idx)]
output = dict(
prompt_input_ids=prompt_input_ids,
prompt_attention_mask=prompt_attention_mask,
answer_input_ids=answer_input_ids,
answer_attention_mask=answer_attention_mask,
)
if embedding_tokenizer is not None:
embedding_tokenized = embedding_tokenizer(batch["prompt"], truncation=True, add_special_tokens=False)
output.update(
{
"embedding_input_ids": embedding_tokenized["input_ids"],
"embedding_attention_mask": embedding_tokenized["attention_mask"],
}
)
return output
def _process_tokens(example: dict[str, Any], model: "PreTrainedModel" = None, **kwargs) -> dict:
"""Process tokens of a BCO specific dataset.
At this stage, we don't convert to PyTorch tensors yet; we just handle the truncation
in case the prompt + completion responses is/are too long. First
we truncate the prompt; if we're still too long, we truncate the completion.
We also create the labels for the completion responses, which are of length equal to
the sum of the length of the prompt and the completion response, with
label_pad_token_id for the prompt tokens.
"""
prompt = example["prompt"]
completion = example["completion"]
batch = {
f"{kwargs['prefix']}prompt": prompt,
f"{kwargs['prefix']}completion": completion,
f"{kwargs['prefix']}label": example["label"],
}
if not kwargs["is_encoder_decoder"]:
# Check issues below for more details
# 1. https://github.com/huggingface/trl/issues/907
# 2. https://github.com/EleutherAI/lm-evaluation-harness/pull/531#issuecomment-1595586257
# 3. https://github.com/LianjiaTech/BELLE/issues/337
if not isinstance(prompt, str):
raise ValueError(f"prompt should be an str but got {type(prompt)}")
if not isinstance(completion, str):
raise ValueError(f"completion should be an str but got {type(completion)}")
# keys of format prompt_* refers to just the prompt and answer_* refers to just the answer
all_tokens = {
"prompt_input_ids": example["prompt_input_ids"],
"prompt_attention_mask": example["prompt_attention_mask"],
"answer_input_ids": example["answer_input_ids"],
"answer_attention_mask": example["answer_attention_mask"],
}
# calculate max length by checking if BOS/EOS is already there
max_length = kwargs["max_length"]
bos_token_id = kwargs["tokenizer"].bos_token_id
eos_token_id = kwargs["tokenizer"].eos_token_id
if bos_token_id != all_tokens["prompt_input_ids"][0]:
max_length -= 1
if eos_token_id != all_tokens["answer_input_ids"][-1]:
max_length -= 1
# if combined sequence is too long (> max_length - 1 for BOS token - 1 for EOS), truncate the prompt
if len(all_tokens["prompt_input_ids"]) + len(all_tokens["answer_input_ids"]) > max_length:
for k in ["prompt_input_ids", "prompt_attention_mask"]:
if kwargs["truncation_mode"] == "keep_start":
all_tokens[k] = all_tokens[k][: kwargs["max_prompt_length"]]
elif kwargs["truncation_mode"] == "keep_end":
all_tokens[k] = all_tokens[k][-kwargs["max_prompt_length"] :]
else:
raise ValueError(f"Unknown truncation mode: {kwargs['truncation_mode']}")
# if that's still too long, truncate the response
if len(all_tokens["prompt_input_ids"]) + len(all_tokens["answer_input_ids"]) > max_length:
for k in ["answer_input_ids", "answer_attention_mask"]:
all_tokens[k] = all_tokens[k][: max_length - kwargs["max_prompt_length"]]
# all input_ids and attention mask as is. We then check if we need to add BOS/EOS tokens
batch[f"{kwargs['prefix']}prompt_input_ids"] = all_tokens["prompt_input_ids"]
batch[f"{kwargs['prefix']}prompt_attention_mask"] = all_tokens["prompt_attention_mask"]
batch[f"{kwargs['prefix']}completion_input_ids"] = (
all_tokens["prompt_input_ids"] + all_tokens["answer_input_ids"]
)
batch[f"{kwargs['prefix']}completion_attention_mask"] = (
all_tokens["prompt_attention_mask"] + all_tokens["answer_attention_mask"]
)
# add BOS, which affects both prompt and the full completion
if bos_token_id is not None:
if len(all_tokens["prompt_input_ids"]) == 0 or bos_token_id != all_tokens["prompt_input_ids"][0]:
batch[f"{kwargs['prefix']}prompt_input_ids"] = [bos_token_id] + batch[
f"{kwargs['prefix']}prompt_input_ids"
]
batch[f"{kwargs['prefix']}prompt_attention_mask"] = [1] + batch[
f"{kwargs['prefix']}prompt_attention_mask"
]
batch[f"{kwargs['prefix']}completion_input_ids"] = [bos_token_id] + batch[
f"{kwargs['prefix']}completion_input_ids"
]
batch[f"{kwargs['prefix']}completion_attention_mask"] = [1] + batch[
f"{kwargs['prefix']}completion_attention_mask"
]
# add EOS, which affects only the full completion
if len(all_tokens["answer_input_ids"]) == 0 or eos_token_id != all_tokens["answer_input_ids"][-1]:
batch[f"{kwargs['prefix']}completion_input_ids"] = batch[f"{kwargs['prefix']}completion_input_ids"] + [
eos_token_id
]
batch[f"{kwargs['prefix']}completion_attention_mask"] = batch[
f"{kwargs['prefix']}completion_attention_mask"
] + [1]
batch[f"{kwargs['prefix']}completion_labels"] = batch[f"{kwargs['prefix']}completion_input_ids"][:]
batch[f"{kwargs['prefix']}completion_labels"][: len(batch[f"{kwargs['prefix']}prompt_input_ids"])] = [
kwargs["label_pad_token_id"]
] * len(batch[f"{kwargs['prefix']}prompt_input_ids"])
else:
completion_tokens = kwargs["tokenizer"](
completion, truncation=True, max_length=kwargs["max_completion_length"], add_special_tokens=True
)
prompt_tokens = kwargs["tokenizer"](
prompt, truncation=True, max_length=kwargs["max_prompt_length"], add_special_tokens=True
)
batch[f"{kwargs['prefix']}prompt_input_ids"] = prompt_tokens["input_ids"]
batch[f"{kwargs['prefix']}prompt_attention_mask"] = prompt_tokens["attention_mask"]
batch[f"{kwargs['prefix']}completion_labels"] = completion_tokens["input_ids"]
batch[f"{kwargs['prefix']}completion_attention_mask"] = completion_tokens["attention_mask"]
if model is not None and hasattr(model, "prepare_decoder_input_ids_from_labels"):
batch[f"{kwargs['prefix']}completion_decoder_input_ids"] = model.prepare_decoder_input_ids_from_labels(
labels=torch.tensor(batch["completion_labels"])
)
return batch
class BCOTrainer(Trainer):
r"""
Initialize BCOTrainer from [BCO](https://huggingface.co/papers/2404.04656) paper.
Args:
model (`transformers.PreTrainedModel`):
The model to train, preferably an `AutoModelForSequenceClassification`.
ref_model (`PreTrainedModelWrapper`):
Hugging Face transformer model with a casual language modelling head. Used for implicit reward computation and loss. If no
reference model is provided, the trainer will create a reference model with the same architecture as the model to be optimized.
args (`BCOConfig`):
The arguments to use for training.
train_dataset (`datasets.Dataset`):
The dataset to use for training.
eval_dataset (`datasets.Dataset`):
The dataset to use for evaluation.
processing_class (`PreTrainedTokenizerBase` or `BaseImageProcessor` or `FeatureExtractionMixin` or `ProcessorMixin`, *optional*):
Processing class used to process the data. If provided, will be used to automatically process the inputs
for the model, and it will be saved along the model to make it easier to rerun an interrupted training or
reuse the fine-tuned model.
data_collator (`transformers.DataCollator`, *optional*, defaults to `None`):
The data collator to use for training. If None is specified, the default data collator (`DPODataCollatorWithPadding`) will be used
which will pad the sequences to the maximum length of the sequences in the batch, given a dataset of paired sequences.
model_init (`Callable[[], transformers.PreTrainedModel]`):
The model initializer to use for training. If None is specified, the default model initializer will be used.
callbacks (`list[transformers.TrainerCallback]`):
The callbacks to use for training.
optimizers (`tuple[torch.optim.Optimizer, torch.optim.lr_scheduler.LambdaLR]`):
The optimizer and scheduler to use for training.
preprocess_logits_for_metrics (`Callable[[torch.Tensor, torch.Tensor], torch.Tensor]`):
The function to use to preprocess the logits before computing the metrics.
peft_config (`dict`, defaults to `None`):
The PEFT configuration to use for training. If you pass a PEFT configuration, the model will be wrapped in a PEFT model.
compute_metrics (`Callable[[EvalPrediction], dict]`, *optional*):
The function to use to compute the metrics. Must take a `EvalPrediction` and return
a dictionary string to metric values.
model_adapter_name (`str`, defaults to `None`):
Name of the train target PEFT adapter, when using LoRA with multiple adapters.
ref_adapter_name (`str`, defaults to `None`):
Name of the reference PEFT adapter, when using LoRA with multiple adapters.
"""
_tag_names = ["trl", "bco"]
def __init__(
self,
model: Union[PreTrainedModel, nn.Module, str] = None,
ref_model: Optional[Union[PreTrainedModel, nn.Module, str]] = None,
args: BCOConfig = None,
train_dataset: Optional[Dataset] = None,
eval_dataset: Optional[Union[Dataset, dict[str, Dataset]]] = None,
processing_class: Optional[
Union[PreTrainedTokenizerBase, BaseImageProcessor, FeatureExtractionMixin, ProcessorMixin]
] = None,
data_collator: Optional[DataCollator] = None,
model_init: Optional[Callable[[], PreTrainedModel]] = None,
callbacks: Optional[list[TrainerCallback]] = None,
optimizers: tuple[torch.optim.Optimizer, torch.optim.lr_scheduler.LambdaLR] = (None, None),
preprocess_logits_for_metrics: Optional[Callable[[torch.Tensor, torch.Tensor], torch.Tensor]] = None,
peft_config: Optional[dict] = None,
compute_metrics: Optional[Callable[[EvalLoopOutput], dict]] = None,
model_adapter_name: Optional[str] = None,
ref_adapter_name: Optional[str] = None,
embedding_func: Optional[Callable] = None,
embedding_tokenizer: Optional[PreTrainedTokenizerBase] = None,
):
if not is_sklearn_available():
raise ImportError(
"BCOTrainer requires the scikit-learn library. Please install it with `pip install scikit-learn`."
)
if type(args) is TrainingArguments:
raise ValueError("Please use `BCOConfig` instead `TrainingArguments`.")
if not isinstance(model, str) and ref_model is model:
raise ValueError(
"`model` and `ref_model` cannot be the same object. If you want `ref_model` to be the "
"same as `model`, you must mass a copy of it, or `None` if you use peft."
)
if args.model_init_kwargs is None:
model_init_kwargs = {}
elif not isinstance(model, str):
raise ValueError("You passed model_kwargs to the BCOTrainer. But your model is already instantiated.")
else:
model_init_kwargs = args.model_init_kwargs
torch_dtype = model_init_kwargs.get("torch_dtype")
if torch_dtype is not None:
# Convert to `torch.dtype` if an str is passed
if isinstance(torch_dtype, str) and torch_dtype != "auto":
torch_dtype = getattr(torch, torch_dtype)
if torch_dtype != "auto" and not isinstance(torch_dtype, torch.dtype):
raise ValueError(
f"Invalid `torch_dtype` passed to the BCOConfig. Expected a string with either `torch.dtype` or 'auto', but got {torch_dtype}."
)
model_init_kwargs["torch_dtype"] = torch_dtype
if args.ref_model_init_kwargs is None:
ref_model_init_kwargs = {}
elif not isinstance(ref_model, str):
raise ValueError(
"You passed ref_model_kwargs to the BCOTrainer. But your ref_model is already instantiated."
)
else:
ref_model_init_kwargs = args.ref_model_init_kwargs
torch_dtype = ref_model_init_kwargs.get("torch_dtype")
if torch_dtype is not None:
# Convert to `torch.dtype` if an str is passed
if isinstance(torch_dtype, str) and torch_dtype != "auto":
torch_dtype = getattr(torch, torch_dtype)
if torch_dtype != "auto" and not isinstance(torch_dtype, torch.dtype):
raise ValueError(
f"Invalid `torch_dtype` passed to the BCOConfig. Expected a string with either `torch.dtype` or 'auto', but got {torch_dtype}."
)
ref_model_init_kwargs["torch_dtype"] = torch_dtype
if isinstance(model, str):
model = AutoModelForCausalLM.from_pretrained(model, **model_init_kwargs)
if isinstance(ref_model, str):
ref_model = AutoModelForCausalLM.from_pretrained(ref_model, **ref_model_init_kwargs)
# Initialize this variable to False. This helps tracking the case when `peft_module_casting_to_bf16`
# has been called in order to properly call autocast if needed.
self._peft_has_been_casted_to_bf16 = False
if not is_peft_available() and peft_config is not None:
raise ValueError(
"PEFT is not installed and you passed a `peft_config` in the trainer's kwargs, please install it with `pip install peft` to use the PEFT models"
)
elif is_peft_available() and peft_config is not None:
# if model is a peft model and we have a peft_config, we merge and unload it first
if isinstance(model, PeftModel):
model = model.merge_and_unload()
if getattr(model, "is_loaded_in_8bit", False) or getattr(model, "is_loaded_in_4bit", False):
_support_gc_kwargs = hasattr(
args, "gradient_checkpointing_kwargs"
) and "gradient_checkpointing_kwargs" in list(
inspect.signature(prepare_model_for_kbit_training).parameters
)
prepare_model_kwargs = {"use_gradient_checkpointing": args.gradient_checkpointing}
if _support_gc_kwargs:
prepare_model_kwargs["gradient_checkpointing_kwargs"] = args.gradient_checkpointing_kwargs
model = prepare_model_for_kbit_training(model, **prepare_model_kwargs)
elif getattr(args, "gradient_checkpointing", False):
# For backward compatibility with older versions of transformers
if hasattr(model, "enable_input_require_grads"):
model.enable_input_require_grads()
else:
def make_inputs_require_grad(module, input, output):
output.requires_grad_(True)
model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
# get peft model with the given config
model = get_peft_model(model, peft_config)
if args.bf16 and getattr(model, "is_loaded_in_4bit", False):
peft_module_casting_to_bf16(model)
# If args.bf16 we need to explicitly call `generate` with torch amp autocast context manager
self._peft_has_been_casted_to_bf16 = True
# For models that use gradient_checkpointing, we need to attach a hook that enables input
# to explicitly have `requires_grad=True`, otherwise training will either silently
# fail or completely fail.
elif getattr(args, "gradient_checkpointing", False):
# For backward compatibility with older versions of transformers
if hasattr(model, "enable_input_require_grads"):
model.enable_input_require_grads()
else:
def make_inputs_require_grad(module, input, output):
output.requires_grad_(True)
model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
if args.generate_during_eval and not (is_wandb_available() or is_comet_available()):
raise ValueError(
"`generate_during_eval=True` requires Weights and Biases or Comet to be installed."
" Please install `wandb` or `comet-ml` to resolve."
)
if model is not None:
self.is_encoder_decoder = model.config.is_encoder_decoder
elif args.is_encoder_decoder is None:
raise ValueError("When no model is provided, you need to pass the parameter is_encoder_decoder.")
else:
self.is_encoder_decoder = args.is_encoder_decoder
self.is_peft_model = is_peft_available() and isinstance(model, PeftModel)
self.model_adapter_name = model_adapter_name
self.ref_adapter_name = ref_adapter_name
if ref_model:
self.ref_model = ref_model
elif self.is_peft_model or args.precompute_ref_log_probs:
# The `model` with adapters turned off will be used as the reference model
self.ref_model = None
else:
self.ref_model = create_reference_model(model)
if processing_class is None:
raise ValueError(
"max_length or a processing_class must be specified when using the default DPODataCollatorWithPadding"
)
if args.max_length is None:
warnings.warn(
"When using DPODataCollatorWithPadding, you should set `max_length` in the `BCOConfig`. "
"It will be set to `512` by default, but you should do it yourself in the future.",
UserWarning,
)
max_length = 512
if args.max_length is not None:
max_length = args.max_length
if args.max_prompt_length is None:
warnings.warn(
"When using DPODataCollatorWithPadding, you should set `max_prompt_length` in the `BCOConfig`. "
"It will be set to `128` by default, but you should do it yourself in the future.",
UserWarning,
)
max_prompt_length = 128
if args.max_prompt_length is not None:
max_prompt_length = args.max_prompt_length
max_completion_length = None
if args.max_completion_length is None and self.is_encoder_decoder:
warnings.warn(
"When using DPODataCollatorWithPadding with an encoder decoder architecture, you should set `max_completion_length` in the BCOTrainer's init"
" it will be set to `128` by default, but you should do it yourself in the future.",
UserWarning,
)
max_completion_length = 128
if args.max_completion_length is not None and self.is_encoder_decoder:
max_completion_length = args.max_completion_length
if data_collator is None:
data_collator = DPODataCollatorWithPadding(
pad_token_id=processing_class.pad_token_id,
label_pad_token_id=args.label_pad_token_id,
is_encoder_decoder=self.is_encoder_decoder,
)
if args.remove_unused_columns:
args.remove_unused_columns = False
# warn users
warnings.warn(
"When using DPODataCollatorWithPadding, you should set `remove_unused_columns=False` in your BCOConfig"
" we have set it for you, but you should do it yourself in the future.",
UserWarning,
)
self.use_dpo_data_collator = True
else:
self.use_dpo_data_collator = False
# Disable dropout in the model and reference model
if args.disable_dropout:
disable_dropout_in_model(model)
if self.ref_model is not None:
disable_dropout_in_model(self.ref_model)
self.max_length = max_length
self.generate_during_eval = args.generate_during_eval
self.label_pad_token_id = args.label_pad_token_id
self.padding_value = args.padding_value if args.padding_value is not None else processing_class.pad_token_id
self.max_prompt_length = max_prompt_length
self.truncation_mode = args.truncation_mode
self.max_completion_length = max_completion_length
self.precompute_ref_log_probs = args.precompute_ref_log_probs
# Since ref_logs are precomputed on the first call to get_train/eval_dataloader
# keep track of first called to avoid computation of future calls
self._precomputed_train_ref_log_probs = False
self._precomputed_eval_ref_log_probs = False
# metric
self._stored_metrics = defaultdict(lambda: defaultdict(list))
# BCO parameter
self.beta = args.beta
self.aux_loss_enabled = getattr(model.config, "output_router_logits", False)
self.aux_loss_coef = getattr(model.config, "router_aux_loss_coef", 0.0)
if self.aux_loss_enabled and self.aux_loss_coef == 0.0:
warnings.warn(
"You set `output_router_logits` to `True` in the model config, but `router_aux_loss_coef` is set to "
"`0.0`, meaning the auxiliary loss will not be used. Either set `router_aux_loss_coef` to a value "
"greater than `0.0`, or set `output_router_logits` to `False` if you don't want to use the auxiliary "
"loss.",
UserWarning,
)
# Underlying Distribution Matching argument
self.embedding_func = embedding_func
self.embedding_tokenizer = embedding_tokenizer
# The trainer estimates the number of FLOPs (floating-point operations) using the number of elements in the
# input tensor associated with the key "input_ids". However, in BCO, the sampled data does not include the
# "input_ids" key. Instead, the available keys are "prompt_input_ids" and "completion_input_ids". As a result,
# the trainer issues the warning: "Could not estimate the number of tokens of the input, floating-point
# operations will not be computed." To suppress this warning, we set the "estimate_tokens" key in the model's
# "warnings_issued" dictionary to True. This acts as a flag to indicate that the warning has already been
# issued.
model.warnings_issued["estimate_tokens"] = True
with PartialState().local_main_process_first():
# Apply the chat template if needed
train_dataset = train_dataset.map(
maybe_apply_chat_template, fn_kwargs={"tokenizer": processing_class}, num_proc=args.dataset_num_proc
)
if eval_dataset is not None:
eval_dataset = eval_dataset.map(
maybe_apply_chat_template,
fn_kwargs={"tokenizer": processing_class},
num_proc=args.dataset_num_proc,
)
# Shuffle the datasets
train_dataset = train_dataset.shuffle(seed=args.data_seed)
if eval_dataset is not None:
eval_dataset = eval_dataset.shuffle(seed=args.data_seed)
# Tokenize and prepare the training datasets
train_dataset = train_dataset.map(
_tokenize,
batched=True,
fn_kwargs={"tokenizer": processing_class, "embedding_tokenizer": self.embedding_tokenizer},
num_proc=args.dataset_num_proc,
desc="Tokenizing train dataset",
)
# Prepare the datasets
fn_kwargs = {
"prefix": "",
"is_encoder_decoder": self.is_encoder_decoder,
"tokenizer": processing_class,
"max_length": self.max_length,
"truncation_mode": self.truncation_mode,
"label_pad_token_id": self.label_pad_token_id,
"max_prompt_length": self.max_prompt_length,
"max_completion_length": self.max_completion_length,
}
train_dataset = train_dataset.map(
_process_tokens,
fn_kwargs=fn_kwargs,
num_proc=args.dataset_num_proc,
desc="Processing tokenized train dataset",
)
if eval_dataset is not None:
# Tokenize
eval_dataset = eval_dataset.map(
_tokenize,
fn_kwargs={"tokenizer": processing_class, "embedding_tokenizer": self.embedding_tokenizer},
batched=True,
num_proc=args.dataset_num_proc,
desc="Tokenizing eval dataset",
)
# Process
fn_kwargs = {
"prefix": "",
"is_encoder_decoder": self.is_encoder_decoder,
"tokenizer": processing_class,
"max_length": self.max_length,
"truncation_mode": self.truncation_mode,
"label_pad_token_id": self.label_pad_token_id,
"max_prompt_length": self.max_prompt_length,
"max_completion_length": self.max_completion_length,
}
eval_dataset = eval_dataset.map(
_process_tokens,
fn_kwargs=fn_kwargs,
num_proc=args.dataset_num_proc,
desc="Processing tokenized eval dataset",
)
desirable = train_dataset.filter(
lambda x: x["label"], num_proc=args.dataset_num_proc, desc="Filtering desirable examples"
)
undesirable = train_dataset.filter(
lambda x: not x["label"], num_proc=args.dataset_num_proc, desc="Filtering undesirable examples"
)
desirable = desirable.shuffle(seed=args.data_seed)
undesirable = undesirable.shuffle(seed=args.data_seed)
super().__init__(
model=model,
args=args,
data_collator=data_collator,
train_dataset=train_dataset,
eval_dataset=eval_dataset,
processing_class=processing_class,
model_init=model_init,
compute_metrics=compute_metrics,
callbacks=callbacks,
optimizers=optimizers,
preprocess_logits_for_metrics=preprocess_logits_for_metrics,
)
# Gradient accumulation requires scaled loss. Normally, loss scaling in the parent class depends on whether the
# model accepts loss-related kwargs. Since we compute our own loss, this check is irrelevant. We set
# self.model_accepts_loss_kwargs to False to enable scaling.
self.model_accepts_loss_kwargs = False
# Add tags for models that have been loaded with the correct transformers version
if hasattr(self.model, "add_model_tags"):
self.model.add_model_tags(self._tag_names)
if not hasattr(self, "accelerator"):
raise AttributeError(
"Your `Trainer` does not have an `accelerator` object. Consider upgrading `transformers`."
)
# Deepspeed Zero-3 does not support precompute_ref_log_probs
if self.is_deepspeed_enabled:
if self.accelerator.state.deepspeed_plugin.zero_stage == 3 and self.precompute_ref_log_probs:
raise ValueError(
"You cannot use `precompute_ref_log_probs=True` with Deepspeed ZeRO-3. Please set `precompute_ref_log_probs=False`."
)
if self.ref_model is None:
if not (self.is_peft_model or self.precompute_ref_log_probs):
raise ValueError(
"No reference model and model is not a Peft model. Try setting `precompute_ref_log_probs=True`"
)
else:
if self.is_deepspeed_enabled:
self.ref_model = self._prepare_deepspeed(self.ref_model)
else:
self.ref_model = self.accelerator.prepare_model(self.ref_model, evaluation_mode=True)
self.running = RunningMoments(accelerator=self.accelerator)
if self.embedding_func is None:
return
chosen_embeddings = self._get_sample_prompt_embeddings(desirable, sample_size=self.args.prompt_sample_size)
rejected_embeddings = self._get_sample_prompt_embeddings(undesirable, sample_size=self.args.prompt_sample_size)
embeddings = torch.cat((chosen_embeddings, rejected_embeddings), dim=0)
labels = torch.cat(
(torch.ones_like(chosen_embeddings[:, 0]), torch.zeros_like(rejected_embeddings[:, 0])), dim=0
)
self.clf = LogisticRegression(class_weight="balanced").fit(
embeddings.cpu().float().numpy(), labels.cpu().numpy()
)
@property
def match_underlying_distribution(self):
return self.embedding_func is not None and self.embedding_tokenizer is not None
def _get_chosen_prob(self, prompt_embeddings: torch.FloatTensor) -> torch.FloatTensor:
"""
Calculates the probability if the given prompt embedding is from desirable dataset.
This function calculates the probability in the process and ensemble across processes.
"""
dtype = prompt_embeddings.dtype
device = prompt_embeddings.device
rank = self.accelerator.process_index
padded_prompt_embeddings = self.accelerator.pad_across_processes(
prompt_embeddings, pad_index=self.embedding_tokenizer.pad_token_id
)
sample_size = padded_prompt_embeddings.shape[0]
nonzero = padded_prompt_embeddings.mean(dim=1) != self.embedding_tokenizer.pad_token_id
prompt_embeddings = self.accelerator.gather(padded_prompt_embeddings)
# cannot predict for all empty values
if prompt_embeddings.shape[0] == 0:
return torch.tensor([], device=device, dtype=dtype)
prob = self.clf.predict_proba(prompt_embeddings.cpu().float().numpy())[:, 1]
prob = torch.as_tensor(prob, dtype=dtype, device=device)
prob = self.accelerator.reduce(prob, reduction="mean")
prob = prob[sample_size * rank : sample_size * (rank + 1)]
prob = prob[nonzero]
return prob
def _vectorize_prompt(self, input_ids: torch.LongTensor, attention_mask: torch.LongTensor) -> torch.FloatTensor:
"""
Replaces processing_class.pad_token_id to embedding_tokenizer.pad_token_id
and applies self.embedding_func
"""
input_ids = torch.where(
input_ids == self.processing_class.pad_token_id,
self.embedding_tokenizer.pad_token_id,
input_ids,
)
with torch.no_grad():
embeddings = self.embedding_func(
input_ids=input_ids,
attention_mask=attention_mask,
)
return embeddings
def _get_prompt_embeddings(
self, batch: dict[str, Union[list, torch.LongTensor]]
) -> tuple[torch.FloatTensor, torch.FloatTensor]:
"""Extract embeddings from frozen embedding model"""
if not self.match_underlying_distribution:
return None, None
embeddings = self._vectorize_prompt(
input_ids=batch["embedding_input_ids"],
attention_mask=batch["embedding_attention_mask"],
)
chosen_idx = [i for i in range(len(batch["label"])) if batch["label"][i] is True]
rejected_idx = [i for i in range(len(batch["label"])) if batch["label"][i] is False]
chosen_embeddings = embeddings[chosen_idx, ...]
rejected_embeddings = embeddings[rejected_idx, ...]
return (chosen_embeddings, rejected_embeddings)
def _get_sample_prompt_embeddings(self, dataset: Dataset, sample_size: int = 512) -> torch.FloatTensor:
"""
Sample instances from dataset and get prompt embeddings.
Used for density ratio classifier training.
"""
n_samples = min(len(dataset), sample_size)
rand_indices = np.random.choice(len(dataset), size=(n_samples,))
embedding_dataset = dataset.select(rand_indices)
dataloader_params = {
"batch_size": self.args.per_device_train_batch_size,
"collate_fn": self.data_collator,
"num_workers": self.args.dataloader_num_workers,
"pin_memory": self.args.dataloader_pin_memory,
"shuffle": False,
}
# prepare dataloader
data_loader = self.accelerator.prepare(DataLoader(embedding_dataset, **dataloader_params))
with torch.no_grad():
all_embeddings = torch.empty(0)
for padded_batch in tqdm(iterable=data_loader, desc="Building sample prompt embeddings"):
embeddings = self._vectorize_prompt(
input_ids=padded_batch["embedding_input_ids"],
attention_mask=padded_batch["embedding_attention_mask"],
)
embeddings = self.accelerator.gather_for_metrics(embeddings)
all_embeddings = torch.cat((all_embeddings, embeddings.cpu()))
return all_embeddings
def _prepare_deepspeed(self, model: PreTrainedModelWrapper):
# Adapted from accelerate: https://github.com/huggingface/accelerate/blob/739b135f8367becb67ffaada12fe76e3aa60fefd/src/accelerate/accelerator.py#L1473
deepspeed_plugin = self.accelerator.state.deepspeed_plugin
config_kwargs = deepcopy(deepspeed_plugin.deepspeed_config)
if model is not None:
if hasattr(model, "config"):
hidden_size = (
max(model.config.hidden_sizes)
if getattr(model.config, "hidden_sizes", None)
else getattr(model.config, "hidden_size", None)
)
if hidden_size is not None and config_kwargs["zero_optimization"]["stage"] == 3:
# Note that `stage3_prefetch_bucket_size` can produce DeepSpeed messages like: `Invalidate trace cache @ step 0: expected module 1, but got module 0`
# This is expected and is not an error, see: https://github.com/microsoft/DeepSpeed/discussions/4081
config_kwargs.update(
{
"zero_optimization.reduce_bucket_size": hidden_size * hidden_size,
"zero_optimization.stage3_param_persistence_threshold": 10 * hidden_size,
"zero_optimization.stage3_prefetch_bucket_size": 0.9 * hidden_size * hidden_size,
}
)
# If ZeRO-3 is used, we shard both the active and reference model.
# Otherwise, we assume the reference model fits in memory and is initialized on each device with ZeRO disabled (stage 0)
if config_kwargs["zero_optimization"]["stage"] != 3:
config_kwargs["zero_optimization"]["stage"] = 0
model, *_ = deepspeed.initialize(model=model, config=config_kwargs)
model.eval()
return model
def _save_optimizer_and_scheduler(self, output_dir):
super()._save_optimizer_and_scheduler(output_dir)
# When saving optimizer and scheduler to checkpoint, save also the running delta object.
output_dir = output_dir if output_dir is not None else self.args.output_dir
self.running.save_to_json(os.path.join(output_dir, RUNNING_NAME))
if self.match_underlying_distribution:
torch.save(self.clf.get_params(), os.path.join(output_dir, CLF_NAME))
def _load_optimizer_and_scheduler(self, checkpoint):
super()._load_optimizer_and_scheduler(checkpoint)
if checkpoint is None:
return
# when loading optimizer and scheduler from checkpoint, also load the running delta object.
running_file = os.path.join(checkpoint, RUNNING_NAME)
if os.path.isfile(running_file):
self.running = RunningMoments.load_from_json(self.accelerator, running_file)
if self.match_underlying_distribution:
clf_file = os.path.join(checkpoint, CLF_NAME)
if os.path.isfile(running_file):
self.clf.set_params(**torch.load(clf_file, weights_only=True, map_location="cpu"))
@contextmanager
def null_ref_context(self):
"""Context manager for handling null reference model (that is, peft adapter manipulation)."""
with (
self.accelerator.unwrap_model(self.model).disable_adapter()
if self.is_peft_model and not self.ref_adapter_name
else nullcontext()
):
if self.ref_adapter_name:
self.model.set_adapter(self.ref_adapter_name)
yield
if self.ref_adapter_name:
self.model.set_adapter(self.model_adapter_name or "default")
def get_train_dataloader(self) -> DataLoader:
"""
Returns the training [`~torch.utils.data.DataLoader`].
Subclass of transformers.src.transformers.trainer.get_train_dataloader to precompute `ref_log_probs`.
"""
if self.precompute_ref_log_probs and not self._precomputed_train_ref_log_probs:
dataloader_params = {
"batch_size": self.args.per_device_train_batch_size,
"collate_fn": self.data_collator,
"num_workers": self.args.dataloader_num_workers,
"pin_memory": self.args.dataloader_pin_memory,
"shuffle": False,
}
# prepare dataloader
data_loader = self.accelerator.prepare(DataLoader(self.train_dataset, **dataloader_params))
reference_completion_logps = []
for padded_batch in tqdm(iterable=data_loader, desc="Train dataset reference log probs"):
reference_completion_logp = self.compute_reference_log_probs(padded_batch)
reference_completion_logp = self.accelerator.gather_for_metrics(reference_completion_logp)
reference_completion_logps.append(reference_completion_logp.cpu())
self.train_dataset = self.train_dataset.add_column(
name="reference_logps", column=torch.cat(reference_completion_logps).float().numpy()
)
self._precomputed_train_ref_log_probs = True
return super().get_train_dataloader()
def get_eval_dataloader(self, eval_dataset: Optional[Dataset] = None) -> DataLoader:
"""
Returns the evaluation [`~torch.utils.data.DataLoader`].
Subclass of transformers.src.transformers.trainer.get_eval_dataloader to precompute `ref_log_probs`.
Args:
eval_dataset (`torch.utils.data.Dataset`, *optional*):
If provided, will override `self.eval_dataset`. If it is a [`~datasets.Dataset`], columns not accepted
by the `model.forward()` method are automatically removed. It must implement `__len__`.
"""
if eval_dataset is None and self.eval_dataset is None:
raise ValueError("Trainer: evaluation requires an eval_dataset.")
eval_dataset = eval_dataset if eval_dataset is not None else self.eval_dataset
if self.precompute_ref_log_probs and not self._precomputed_eval_ref_log_probs:
dataloader_params = {
"batch_size": self.args.per_device_eval_batch_size,
"collate_fn": self.data_collator,
"num_workers": self.args.dataloader_num_workers,
"pin_memory": self.args.dataloader_pin_memory,
"shuffle": False,
}
# prepare dataloader
data_loader = self.accelerator.prepare(DataLoader(eval_dataset, **dataloader_params))
reference_completion_logps = []
for padded_batch in tqdm(iterable=data_loader, desc="Eval dataset reference log probs"):
reference_completion_logp = self.compute_reference_log_probs(padded_batch)
reference_completion_logp = self.accelerator.gather_for_metrics(reference_completion_logp)
reference_completion_logps.append(reference_completion_logp.cpu())
eval_dataset = eval_dataset.add_column(
name="reference_logps", column=torch.cat(reference_completion_logps).float().numpy()
)
# Save calculated reference_chosen_logps and reference_rejected_logps to the eval_dataset for subsequent runs
if self.eval_dataset is not None:
self.eval_dataset = eval_dataset
self._precomputed_eval_ref_log_probs = True
return super().get_eval_dataloader(eval_dataset=eval_dataset)
def compute_reference_log_probs(self, padded_batch: dict) -> dict:
"""Computes log probabilities of the reference model for a single padded batch of a BCO specific dataset."""
with torch.no_grad():
if self.ref_model is None:
with self.null_ref_context():
if self.is_encoder_decoder:
completion_logits = self.model(
padded_batch["prompt_input_ids"],
attention_mask=padded_batch["prompt_attention_mask"],
decoder_input_ids=padded_batch.get("completion_decoder_input_ids"),
labels=padded_batch["completion_labels"],
).logits
else:
completion_logits = self.model(
padded_batch["completion_input_ids"],
attention_mask=padded_batch["completion_attention_mask"],
).logits
else:
if self.is_encoder_decoder:
completion_logits = self.ref_model(
padded_batch["prompt_input_ids"],
attention_mask=padded_batch["prompt_attention_mask"],
decoder_input_ids=padded_batch.get("completion_decoder_input_ids"),
labels=padded_batch["completion_labels"],
).logits
else:
completion_logits = self.ref_model(
padded_batch["completion_input_ids"], attention_mask=padded_batch["completion_attention_mask"]
).logits
completion_logps = self.get_batch_logps(
completion_logits,
padded_batch["completion_labels"],
average_log_prob=False,
is_encoder_decoder=self.is_encoder_decoder,
label_pad_token_id=self.label_pad_token_id,
)
return completion_logps
@staticmethod
def get_batch_logps(
logits: torch.FloatTensor,
labels: torch.LongTensor,
average_log_prob: bool = False,
label_pad_token_id: int = -100,
is_encoder_decoder: bool = False,
) -> torch.FloatTensor:
"""Compute the log probabilities of the given labels under the given logits.
Args:
logits: Logits of the model (unnormalized). Shape: (batch_size, sequence_length, vocab_size)
labels: Labels for which to compute the log probabilities. Label tokens with a value of label_pad_token_id are ignored. Shape: (batch_size, sequence_length)
average_log_prob: If True, return the average log probability per (non-masked) token. Otherwise, return the sum of the log probabilities of the (non-masked) tokens.
Returns:
A tensor of shape (batch_size,) containing the average/sum log probabilities of the given labels under the given logits.
"""
if logits.shape[:-1] != labels.shape:
raise ValueError("Logits (batch and sequence length dim) and labels must have the same shape.")
if not is_encoder_decoder:
labels = labels[:, 1:].clone()
logits = logits[:, :-1, :]
else:
# Fixes end-dec RuntimeError
labels = labels.clone()
loss_mask = labels != label_pad_token_id
# dummy token; we'll ignore the losses on these tokens later
labels[labels == label_pad_token_id] = 0
per_token_logps = selective_log_softmax(logits, labels)
if average_log_prob:
return (per_token_logps * loss_mask).sum(-1) / loss_mask.sum(-1)
else:
return (per_token_logps * loss_mask).sum(-1)
def forward(
self, model: nn.Module, batch: dict[str, Union[list, torch.LongTensor]]
) -> tuple[torch.FloatTensor, torch.FloatTensor, torch.FloatTensor, torch.FloatTensor]:
model_kwargs = (
{
"labels": batch["completion_labels"],
"decoder_input_ids": batch.get("completion_decoder_input_ids"),
}
if self.is_encoder_decoder
else {}
)
if self.aux_loss_enabled:
model_kwargs["output_router_logits"] = True
outputs = model(
batch["completion_input_ids"],
attention_mask=batch["completion_attention_mask"],
**model_kwargs,
)
completion_logits = outputs.logits
completion_logps = self.get_batch_logps(
completion_logits,
batch["completion_labels"],
average_log_prob=False,
is_encoder_decoder=self.is_encoder_decoder,
label_pad_token_id=self.label_pad_token_id,
)
if completion_logps.shape[0] != len(batch["label"]):
raise ValueError(
"There is a mismatch between the number of examples in this batch and the number of "
"examples for which an output sequence was predicted."
)
chosen_idx = [i for i in range(completion_logps.shape[0]) if batch["label"][i] is True]
rejected_idx = [i for i in range(completion_logps.shape[0]) if batch["label"][i] is False]
chosen_logps = completion_logps[chosen_idx, ...]
rejected_logps = completion_logps[rejected_idx, ...]
chosen_logits = completion_logits[chosen_idx, ...]
rejected_logits = completion_logits[rejected_idx, ...]
if self.aux_loss_enabled:
return (chosen_logps, rejected_logps, chosen_logits, rejected_logits, outputs.aux_loss)
else:
return (chosen_logps, rejected_logps, chosen_logits, rejected_logits)
def _get_udm_weight(self, rejected_embeddings: torch.FloatTensor) -> torch.FloatTensor:
prob_desirable = self._get_chosen_prob(rejected_embeddings)
min_ratio = self.args.min_density_ratio
max_ratio = self.args.max_density_ratio
weight = (prob_desirable / (1 - prob_desirable + 1e-8)).clamp(min=min_ratio, max=max_ratio)
return weight
def bco_loss(
self,
policy_chosen_logps: torch.FloatTensor,
policy_rejected_logps: torch.FloatTensor,
reference_chosen_logps: torch.FloatTensor,
reference_rejected_logps: torch.FloatTensor,
chosen_embeddings: Optional[torch.FloatTensor],
rejected_embeddings: Optional[torch.FloatTensor],
) -> tuple[torch.FloatTensor, torch.FloatTensor, torch.FloatTensor, torch.FloatTensor]:
"""Compute the BCO loss for a batch of policy and reference model log probabilities.
Args:
policy_chosen_logps: Log probabilities of the policy model for the chosen responses. Shape: (num(chosen) in batch_size,)
policy_rejected_logps: Log probabilities of the policy model for the rejected responses. Shape: (num(rejected) in batch_size,)
reference_chosen_logps: Log probabilities of the reference model for the chosen responses. Shape: (num(chosen) in batch_size,)
reference_rejected_logps: Log probabilities of the reference model for the rejected responses. Shape: (num(rejected) in batch_size,)
chosen_embeddings: embeddings of desirable prompts
rejected_embeddings: embeddings of undesirable prompts
Returns:
A tuple of four tensors: (losses, chosen_rewards, rejected_rewards, delta).
The losses tensor contains the BCO loss for each example in the batch.
The chosen_rewards and rejected_rewards tensors contain the rewards for the chosen and rejected responses, respectively.
The delta value contains the moving average of all implicit rewards.
"""
if policy_chosen_logps.shape[0] != 0 or reference_chosen_logps.shape[0] != 0:
chosen_logratios = policy_chosen_logps - reference_chosen_logps
chosen_rewards = self.beta * chosen_logratios
else:
# lists can't be empty -- if they are, then accelerate.gather will hang
chosen_losses = torch.Tensor([]).to(self.accelerator.device)
chosen_rewards = torch.Tensor([]).to(self.accelerator.device)
if policy_rejected_logps.shape[0] != 0 or reference_rejected_logps.shape[0] != 0:
rejected_logratios = policy_rejected_logps - reference_rejected_logps
rejected_rewards = self.beta * rejected_logratios
else:
# lists can't be empty -- if they are, then accelerate.gather will hang
rejected_losses = torch.Tensor([]).to(self.accelerator.device)
rejected_rewards = torch.Tensor([]).to(self.accelerator.device)
rewards = torch.cat((chosen_rewards, rejected_rewards), 0).mean().detach()
self.running.update(rewards)
delta = self.running.mean
if policy_chosen_logps.shape[0] != 0 or reference_chosen_logps.shape[0] != 0:
chosen_losses = -F.logsigmoid(chosen_rewards - delta)
if policy_rejected_logps.shape[0] != 0 or reference_rejected_logps.shape[0] != 0:
rejected_losses = -F.logsigmoid(-(rejected_rewards - delta))
if self.match_underlying_distribution:
chosen_weight = torch.ones_like(chosen_losses)
rejected_weight = self._get_udm_weight(rejected_embeddings)
losses = torch.cat((chosen_weight * chosen_losses, rejected_weight * rejected_losses), dim=0)
else:
losses = torch.cat((chosen_losses, rejected_losses), dim=0)
return losses, chosen_rewards, rejected_rewards, torch.as_tensor(delta)
def get_batch_loss_metrics(
self,
model,
batch: dict[str, Union[list, torch.LongTensor]],
):
"""Compute the BCO loss and other metrics for the given batch of inputs for train or test."""
metrics = {}
batch = {k: (v.to(self.accelerator.device) if isinstance(v, torch.Tensor) else v) for k, v in batch.items()}
forward_output = self.forward(model, batch)
(
policy_chosen_logps,
policy_rejected_logps,
policy_chosen_logits,
policy_rejected_logits,
) = forward_output[:4]
if self.aux_loss_enabled:
aux_loss = forward_output[4]
# if reference_logps in batch use them, otherwise use the reference model
if "reference_logps" in batch:
chosen_idx = [i for i in range(batch["reference_logps"].shape[0]) if batch["label"][i] is True]
rejected_idx = [i for i in range(batch["reference_logps"].shape[0]) if batch["label"][i] is False]
reference_chosen_logps = batch["reference_logps"][chosen_idx, ...]
reference_rejected_logps = batch["reference_logps"][rejected_idx, ...]
else:
with torch.no_grad():
if self.ref_model is None:
with self.null_ref_context():
(
reference_chosen_logps,
reference_rejected_logps,
_,
_,
) = self.forward(self.model, batch)[:4]
else:
(
reference_chosen_logps,
reference_rejected_logps,
_,
_,
) = self.forward(self.ref_model, batch)[:4]
chosen_embeddings, rejected_embeddings = self._get_prompt_embeddings(batch)
losses, chosen_rewards, rejected_rewards, delta = self.bco_loss(
policy_chosen_logps,
policy_rejected_logps,
reference_chosen_logps,
reference_rejected_logps,
chosen_embeddings,
rejected_embeddings,
)
metrics["delta"] = self.accelerator.gather_for_metrics(delta).mean().item()
num_chosen = torch.Tensor([len(chosen_rewards)]).to(self.accelerator.device)
num_rejected = torch.Tensor([len(rejected_rewards)]).to(self.accelerator.device)
all_num_chosen = self.accelerator.gather_for_metrics(num_chosen).sum().item()
all_num_rejected = self.accelerator.gather_for_metrics(num_rejected).sum().item()
if all_num_chosen > 0:
metrics["rewards/chosen_sum"] = (
self.accelerator.gather_for_metrics(chosen_rewards.nansum()).nansum().item()
)
metrics["logps/chosen_sum"] = (
self.accelerator.gather_for_metrics(policy_chosen_logps.nansum()).nansum().item()
)
metrics["logits/chosen_sum"] = (
self.accelerator.gather_for_metrics(policy_chosen_logits.nansum()).nansum().item()
)
metrics["count/chosen"] = all_num_chosen
if all_num_rejected > 0:
metrics["rewards/rejected_sum"] = (
self.accelerator.gather_for_metrics(rejected_rewards.nansum()).nansum().item()
)
metrics["logps/rejected_sum"] = (
self.accelerator.gather_for_metrics(policy_rejected_logps.nansum()).nansum().item()
)
metrics["logits/rejected_sum"] = (
self.accelerator.gather_for_metrics(policy_rejected_logits.nansum()).nansum().item()
)
metrics["count/rejected"] = all_num_rejected
loss = losses.nanmean()
if self.aux_loss_enabled:
loss += self.aux_loss_coef * aux_loss
return loss, metrics
def compute_loss(
self,
model: Union[PreTrainedModel, nn.Module],
inputs: dict[str, Union[torch.Tensor, Any]],
return_outputs=False,
num_items_in_batch=None,
) -> Union[torch.Tensor, tuple[torch.Tensor, dict[str, torch.Tensor]]]:
compute_loss_context_manager = amp.autocast("cuda") if self._peft_has_been_casted_to_bf16 else nullcontext()
with compute_loss_context_manager:
loss, metrics = self.get_batch_loss_metrics(model, inputs)
# Make sure to move the loss to the device the original accumulating loss is at back in the `Trainer` class:
loss = loss.to(self.args.device)
# force log the metrics
if self.accelerator.is_main_process:
self.store_metrics(metrics, train_eval="train")
if return_outputs:
return (loss, metrics)
return loss
def store_metrics(self, metrics: dict[str, float], train_eval: Literal["train", "eval"] = "train") -> None:
for key, value in metrics.items():
self._stored_metrics[train_eval][key].append(value)
def _get_train_sampler(self) -> Optional[torch.utils.data.Sampler]:
if self.train_dataset is None or not has_length(self.train_dataset):
return None
return SequentialSampler(self.train_dataset)
def generate_from_model_and_ref(self, model, batch: dict[str, torch.LongTensor]) -> tuple[str, str]:
"""Generate samples from the model and reference model for the given batch of inputs."""
# If one uses `generate_during_eval` with peft + bf16, we need to explicitly call generate with
# the torch cuda amp context manager as some hidden states are silently casted to full precision.
generate_context_manager = amp.autocast("cuda") if self._peft_has_been_casted_to_bf16 else nullcontext()
with generate_context_manager:
policy_output = model.generate(
input_ids=batch["prompt_input_ids"],
attention_mask=batch["prompt_attention_mask"],
max_length=self.max_length,
do_sample=True,
pad_token_id=self.processing_class.pad_token_id,
)
# if reference_output in batch use that otherwise use the reference model
if "reference_output" in batch:
reference_output = batch["reference_output"]
else:
if self.ref_model is None:
with self.null_ref_context():
reference_output = self.model.generate(
input_ids=batch["prompt_input_ids"],
attention_mask=batch["prompt_attention_mask"],
max_length=self.max_length,
do_sample=True,
pad_token_id=self.processing_class.pad_token_id,
)
else:
reference_output = self.ref_model.generate(
input_ids=batch["prompt_input_ids"],
attention_mask=batch["prompt_attention_mask"],
max_length=self.max_length,
do_sample=True,
pad_token_id=self.processing_class.pad_token_id,
)
policy_output = pad_to_length(policy_output, self.max_length, self.processing_class.pad_token_id)
policy_output_decoded = self.processing_class.batch_decode(policy_output, skip_special_tokens=True)
reference_output = pad_to_length(reference_output, self.max_length, self.processing_class.pad_token_id)
reference_output_decoded = self.processing_class.batch_decode(reference_output, skip_special_tokens=True)
return policy_output_decoded, reference_output_decoded
def prediction_step(
self,
model: Union[PreTrainedModel, nn.Module],
inputs: dict[str, Union[torch.Tensor, Any]],
prediction_loss_only: bool,
ignore_keys: Optional[list[str]] = None,
):
if ignore_keys is None:
if hasattr(model, "config"):
ignore_keys = getattr(model.config, "keys_to_ignore_at_inference", [])
else:
ignore_keys = []
prediction_context_manager = amp.autocast("cuda") if self._peft_has_been_casted_to_bf16 else nullcontext()
with torch.no_grad(), prediction_context_manager:
loss, metrics = self.get_batch_loss_metrics(model, inputs)
# force log the metrics
if self.accelerator.is_main_process:
self.store_metrics(metrics, train_eval="eval")
if prediction_loss_only:
return (loss.detach(), None, None)
# logits for the chosen and rejected samples from model
logits_dict = {
"eval_logits/chosen": metrics["logits/chosen"],
"eval_logits/rejected": metrics["logits/rejected"],
}
logits = tuple(v.unsqueeze(dim=0) for k, v in logits_dict.items() if k not in ignore_keys)
logits = torch.stack(logits).mean(axis=1).to(self.accelerator.device)
labels = torch.zeros(logits.shape[0], device=self.accelerator.device)
return (loss.detach(), logits, labels)
def evaluation_loop(
self,
dataloader: DataLoader,
description: str,
prediction_loss_only: Optional[bool] = None,
ignore_keys: Optional[list[str]] = None,
metric_key_prefix: str = "eval",
) -> EvalLoopOutput:
"""
Overriding built-in evaluation loop to store metrics for each batch.
Prediction/evaluation loop, shared by `Trainer.evaluate()` and `Trainer.predict()`.
Works both with or without labels.
"""
# Sample and save to game log if requested (for one batch to save time)
if self.generate_during_eval:
# Generate random indices within the range of the total number of samples
num_samples = len(dataloader.dataset)
random_indices = random.sample(range(num_samples), k=self.args.eval_batch_size)
# Use dataloader.dataset.select to get the random batch without iterating over the DataLoader
random_batch_dataset = dataloader.dataset.select(random_indices)
random_batch = self.data_collator(random_batch_dataset)
random_batch = self._prepare_inputs(random_batch)
target_indicies = [i for i in range(len(random_batch["label"])) if random_batch["label"][i] is False]
target_batch = {
"prompt_input_ids": random_batch["prompt_input_ids"][target_indicies],
"prompt_attention_mask": random_batch["prompt_attention_mask"][target_indicies],
"prompt": itemgetter(*target_indicies)(random_batch["prompt"]),
}
policy_output_decoded, ref_output_decoded = self.generate_from_model_and_ref(self.model, target_batch)
table = pd.DataFrame(
columns=["Prompt", "Policy", "Ref Model"],
data=[
[prompt, pol[len(prompt) :], ref[len(prompt) :]]
for prompt, pol, ref in zip(target_batch["prompt"], policy_output_decoded, ref_output_decoded)
],
)
if "wandb" in self.args.report_to:
wandb.log({"game_log": wandb.Table(data=table)})
if "comet_ml" in self.args.report_to:
log_table_to_comet_experiment(
name="game_log.csv",
table=table,
)
# Base evaluation
initial_output = super().evaluation_loop(
dataloader, description, prediction_loss_only, ignore_keys, metric_key_prefix
)
return initial_output
def log(self, logs: dict[str, float], start_time: Optional[float] = None) -> None:
"""
Log `logs` on the various objects watching training, including stored metrics.
Args:
logs (`dict[str, float]`):
The values to log.
start_time (`float` or `None`, *optional*, defaults to `None`):
Start time of the training.
"""
# logs either has 'loss' or 'eval_loss'
train_eval = "train" if "loss" in logs else "eval"
# train metrics should have no prefix, eval should have 'eval_'
prefix = "eval_" if train_eval == "eval" else ""
# accumulate average metrics from sums and lengths
for split in ["chosen", "rejected"]:
if f"count/{split}" in self._stored_metrics[train_eval]:
count_sum = torch.Tensor(self._stored_metrics[train_eval][f"count/{split}"]).sum().item()
for metric in ["rewards", "logps", "logits"]:
logs[f"{prefix}{metric}/{split}"] = (
torch.Tensor(self._stored_metrics[train_eval][f"{metric}/{split}_sum"]).sum().item()
/ count_sum
)
# delete obsolete metric
del self._stored_metrics[train_eval][f"{metric}/{split}_sum"]
del self._stored_metrics[train_eval][f"count/{split}"]
# calculate reward margin
if f"{prefix}rewards/chosen" in logs and f"{prefix}rewards/rejected" in logs:
logs[f"{prefix}rewards/margins"] = logs[f"{prefix}rewards/chosen"] - logs[f"{prefix}rewards/rejected"]
# Add averaged stored metrics to logs
for key, metrics in self._stored_metrics[train_eval].items():
logs[f"{prefix}{key}"] = torch.Tensor(metrics).mean().item()
del self._stored_metrics[train_eval]
if version.parse(transformers.__version__) >= version.parse("4.47.0.dev0"):
return super().log(logs, start_time)
else: # transformers<=4.46
return super().log(logs)
def create_model_card(
self,
model_name: Optional[str] = None,
dataset_name: Optional[str] = None,
tags: Union[str, list[str], None] = None,
):
"""
Creates a draft of a model card using the information available to the `Trainer`.
Args:
model_name (`str` or `None`, *optional*, defaults to `None`):
Name of the model.
dataset_name (`str` or `None`, *optional*, defaults to `None`):
Name of the dataset used for training.
tags (`str`, `list[str]` or `None`, *optional*, defaults to `None`):
Tags to be associated with the model card.
"""
if not self.is_world_process_zero():
return
if hasattr(self.model.config, "_name_or_path") and not os.path.isdir(self.model.config._name_or_path):
base_model = self.model.config._name_or_path
else:
base_model = None
tags = tags or []
if isinstance(tags, str):
tags = [tags]
if hasattr(self.model.config, "unsloth_version"):
tags.append("unsloth")
citation = textwrap.dedent("""\
@article{jung2024binary,
title = {{Binary Classifier Optimization for Large Language Model Alignment}},
author = {Seungjae Jung and Gunsoo Han and Daniel Wontae Nam and Kyoung{-}Woon On},
year = 2024,
eprint = {arXiv:2404.04656}
}""")
model_card = generate_model_card(
base_model=base_model,
model_name=model_name,
hub_model_id=self.hub_model_id,
dataset_name=dataset_name,
tags=tags,
wandb_url=wandb.run.get_url() if is_wandb_available() and wandb.run is not None else None,
comet_url=get_comet_experiment_url(),
trainer_name="BCO",
trainer_citation=citation,
paper_title="Binary Classifier Optimization for Large Language Model Alignment",
paper_id="2404.04656",
)
model_card.save(os.path.join(self.args.output_dir, "README.md"))
| trl/trl/trainer/bco_trainer.py/0 | {
"file_path": "trl/trl/trainer/bco_trainer.py",
"repo_id": "trl",
"token_count": 32834
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from dataclasses import dataclass, field
from typing import Optional
@dataclass
class ModelConfig:
"""
Configuration class for the models.
Using [`~transformers.HfArgumentParser`] we can turn this class into
[argparse](https://docs.python.org/3/library/argparse#module-argparse) arguments that can be specified on the
command line.
Parameters:
model_name_or_path (`str` or `None`, *optional*, defaults to `None`):
Model checkpoint for weights initialization.
model_revision (`str`, *optional*, defaults to `"main"`):
Specific model version to use. It can be a branch name, a tag name, or a commit id.
torch_dtype (`Literal["auto", "bfloat16", "float16", "float32"]` or `None`, *optional*, defaults to `None`):
Override the default `torch.dtype` and load the model under this dtype. Possible values are
- `"bfloat16"`: `torch.bfloat16`
- `"float16"`: `torch.float16`
- `"float32"`: `torch.float32`
- `"auto"`: Automatically derive the dtype from the model's weights.
trust_remote_code (`bool`, *optional*, defaults to `False`):
Whether to allow for custom models defined on the Hub in their own modeling files. This option should only
be set to `True` for repositories you trust and in which you have read the code, as it will execute code
present on the Hub on your local machine.
attn_implementation (`str` or `None`, *optional*, defaults to `None`):
Which attention implementation to use. You can run `--attn_implementation=flash_attention_2`, in which case
you must install this manually by running `pip install flash-attn --no-build-isolation`.
use_peft (`bool`, *optional*, defaults to `False`):
Whether to use PEFT for training.
lora_r (`int`, *optional*, defaults to `16`):
LoRA R value.
lora_alpha (`int`, *optional*, defaults to `32`):
LoRA alpha.
lora_dropout (`float`, *optional*, defaults to `0.05`):
LoRA dropout.
lora_target_modules (`Union[str, list[str]]` or `None`, *optional*, defaults to `None`):
LoRA target modules.
lora_modules_to_save (`list[str]` or `None`, *optional*, defaults to `None`):
Model layers to unfreeze & train.
lora_task_type (`str`, *optional*, defaults to `"CAUSAL_LM"`):
Task type to pass for LoRA (use `"SEQ_CLS"` for reward modeling).
use_rslora (`bool`, *optional*, defaults to `False`):
Whether to use Rank-Stabilized LoRA, which sets the adapter scaling factor to `lora_alpha/√r`, instead of
the original default value of `lora_alpha/r`.
load_in_8bit (`bool`, *optional*, defaults to `False`):
Whether to use 8 bit precision for the base model. Works only with LoRA.
load_in_4bit (`bool`, *optional*, defaults to `False`):
Whether to use 4 bit precision for the base model. Works only with LoRA.
bnb_4bit_quant_type (`str`, *optional*, defaults to `"nf4"`):
Quantization type (`"fp4"` or `"nf4"`).
use_bnb_nested_quant (`bool`, *optional*, defaults to `False`):
Whether to use nested quantization.
"""
model_name_or_path: Optional[str] = field(
default=None,
metadata={"help": "Model checkpoint for weights initialization."},
)
model_revision: str = field(
default="main",
metadata={"help": "Specific model version to use. It can be a branch name, a tag name, or a commit id."},
)
torch_dtype: Optional[str] = field(
default=None,
metadata={
"help": "Override the default `torch.dtype` and load the model under this dtype.",
"choices": ["auto", "bfloat16", "float16", "float32"],
},
)
trust_remote_code: bool = field(
default=False,
metadata={
"help": "Whether to allow for custom models defined on the Hub in their own modeling files. This option "
"should only be set to `True` for repositories you trust and in which you have read the code, as it will "
"execute code present on the Hub on your local machine."
},
)
attn_implementation: Optional[str] = field(
default=None,
metadata={
"help": "Which attention implementation to use. You can run `--attn_implementation=flash_attention_2`, in "
"which case you must install this manually by running `pip install flash-attn --no-build-isolation`."
},
)
use_peft: bool = field(
default=False,
metadata={"help": "Whether to use PEFT for training."},
)
lora_r: int = field(
default=16,
metadata={"help": "LoRA R value."},
)
lora_alpha: int = field(
default=32,
metadata={"help": "LoRA alpha."},
)
lora_dropout: float = field(
default=0.05,
metadata={"help": "LoRA dropout."},
)
lora_target_modules: Optional[list[str]] = field(
default=None,
metadata={"help": "LoRA target modules."},
)
lora_modules_to_save: Optional[list[str]] = field(
default=None,
metadata={"help": "Model layers to unfreeze & train."},
)
lora_task_type: str = field(
default="CAUSAL_LM",
metadata={"help": "Task type to pass for LoRA (use 'SEQ_CLS' for reward modeling)."},
)
use_rslora: bool = field(
default=False,
metadata={
"help": "Whether to use Rank-Stabilized LoRA, which sets the adapter scaling factor to `lora_alpha/√r`, "
"instead of the original default value of `lora_alpha/r`."
},
)
load_in_8bit: bool = field(
default=False,
metadata={"help": "Whether to use 8 bit precision for the base model. Works only with LoRA."},
)
load_in_4bit: bool = field(
default=False,
metadata={"help": "Whether to use 4 bit precision for the base model. Works only with LoRA."},
)
bnb_4bit_quant_type: str = field(
default="nf4",
metadata={"help": "Quantization type.", "choices": ["fp4", "nf4"]},
)
use_bnb_nested_quant: bool = field(
default=False,
metadata={"help": "Whether to use nested quantization."},
)
def __post_init__(self):
if self.load_in_8bit and self.load_in_4bit:
raise ValueError("You can't use 8 bit and 4 bit precision at the same time")
if hasattr(self.lora_target_modules, "__len__") and len(self.lora_target_modules) == 1:
self.lora_target_modules = self.lora_target_modules[0]
| trl/trl/trainer/model_config.py/0 | {
"file_path": "trl/trl/trainer/model_config.py",
"repo_id": "trl",
"token_count": 2929
} |
# Copyright 2025 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import dataclasses
import os
import warnings
from collections import defaultdict
from typing import Any, Callable, Optional, Type, Union
import torch
import torch.nn as nn
import transformers
from accelerate import PartialState
from datasets import Dataset, IterableDataset
from packaging import version
from transformers import (
AutoModelForCausalLM,
AutoTokenizer,
BaseImageProcessor,
DataCollator,
DataCollatorForLanguageModeling,
FeatureExtractionMixin,
PreTrainedModel,
PreTrainedTokenizerBase,
ProcessorMixin,
Trainer,
TrainingArguments,
is_wandb_available,
)
from transformers.trainer_callback import TrainerCallback
from transformers.trainer_utils import EvalPrediction
from transformers.utils import is_liger_kernel_available, is_peft_available
from transformers.utils.deprecation import deprecate_kwarg
from ..data_utils import is_conversational, maybe_apply_chat_template, pack_examples
from .sft_config import SFTConfig
from .utils import (
ConstantLengthDataset,
compute_token_accuracy,
generate_model_card,
get_comet_experiment_url,
peft_module_casting_to_bf16,
)
if is_peft_available():
import peft
from peft import PeftConfig, PeftModel, get_peft_model, prepare_model_for_kbit_training
if is_liger_kernel_available():
from liger_kernel.transformers import AutoLigerKernelForCausalLM
if is_wandb_available():
import wandb
class SFTTrainer(Trainer):
"""
Trainer for Supervised Fine-Tuning (SFT) method.
This class is a wrapper around the [`transformers.Trainer`] class and inherits all of its attributes and methods.
Example:
```python
from datasets import load_dataset
from trl import SFTTrainer
dataset = load_dataset("roneneldan/TinyStories", split="train[:1%]")
trainer = SFTTrainer(model="Qwen/Qwen2-0.5B-Instruct", train_dataset=dataset)
trainer.train()
```
Args:
model (`Union[str, PreTrainedModel]`):
Model to be trained. Can be either:
- A string, being the *model id* of a pretrained model hosted inside a model repo on huggingface.co, or
a path to a *directory* containing model weights saved using
[`~transformers.PreTrainedModel.save_pretrained`], e.g., `'./my_model_directory/'`. The model is
loaded using [`~transformers.AutoModelForCausalLM.from_pretrained`] with the keywork arguments
in `args.model_init_kwargs`.
- A [`~transformers.PreTrainedModel`] object. Only causal language models are supported.
args ([`SFTConfig`], *optional*, defaults to `None`):
Configuration for this trainer. If `None`, a default configuration is used.
data_collator (`DataCollator`, *optional*):
Function to use to form a batch from a list of elements of the prcessed `train_dataset` or `eval_dataset`.
Will default to [`~transformers.default_data_collator`] if no `processing_class` is provided, an instance
of [`~transformers.DataCollatorWithPadding`] otherwise if the processing_class is a feature extractor or
tokenizer.
train_dataset ([`~datasets.Dataset`] or [`~datasets.IterableDataset`]):
Dataset to use for training. SFT supports both [language modeling](#language-modeling) type and
[prompt-completion](#prompt-completion) type. The format of the samples can be either:
- [Standard](dataset_formats#standard): Each sample contains plain text.
- [Conversational](dataset_formats#conversational): Each sample contains structured messages (e.g., role
and content).
eval_dataset ([`~datasets.Dataset`], [`~datasets.IterableDataset`] or `dict[str, Union[Dataset, IterableDataset]]`):
Dataset to use for evaluation. It must meet the same requirements as `train_dataset`.
processing_class ([`~transformers.PreTrainedTokenizerBase`], *optional*, defaults to `None`):
Processing class used to process the data. If `None`, the processing class is loaded from the model's name
with [`~transformers.AutoTokenizer.from_pretrained`].
callbacks (list of [`~transformers.TrainerCallback`], *optional*, defaults to `None`):
List of callbacks to customize the training loop. Will add those to the list of default callbacks
detailed in [here](https://huggingface.co/docs/transformers/main_classes/callback).
If you want to remove one of the default callbacks used, use the [`~transformers.Trainer.remove_callback`]
method.
optimizers (`tuple[torch.optim.Optimizer, torch.optim.lr_scheduler.LambdaLR]`, *optional*, defaults to `(None, None)`):
A tuple containing the optimizer and the scheduler to use. Will default to an instance of [`AdamW`] on your
model and a scheduler given by [`get_linear_schedule_with_warmup`] controlled by `args`.
optimizer_cls_and_kwargs (`Tuple[Type[torch.optim.Optimizer], Dict[str, Any]]`, *optional*, defaults to `None`):
A tuple containing the optimizer class and keyword arguments to use.
Overrides `optim` and `optim_args` in `args`. Incompatible with the `optimizers` argument.
Unlike `optimizers`, this argument avoids the need to place model parameters on the correct devices before initializing the Trainer.
preprocess_logits_for_metrics (`Callable[[torch.Tensor, torch.Tensor], torch.Tensor]`, *optional*, defaults to `None`):
A function that preprocess the logits right before caching them at each evaluation step. Must take two
tensors, the logits and the labels, and return the logits once processed as desired. The modifications made
by this function will be reflected in the predictions received by `compute_metrics`.
Note that the labels (second parameter) will be `None` if the dataset does not have them.
peft_config ([`~peft.PeftConfig`], *optional*, defaults to `None`):
PEFT configuration used to wrap the model. If `None`, the model is not wrapped.
formatting_func (`Optional[Callable]`):
Formatting function applied to the dataset before tokenization.
"""
_tag_names = ["trl", "sft"]
@deprecate_kwarg(
"tokenizer", "0.16.0", "processing_class", warn_if_greater_or_equal_version=True, raise_if_both_names=True
)
def __init__(
self,
model: Union[str, nn.Module, PreTrainedModel],
args: Optional[Union[SFTConfig, TrainingArguments]] = None,
data_collator: Optional[DataCollator] = None, # type: ignore
train_dataset: Optional[Union[Dataset, IterableDataset]] = None,
eval_dataset: Optional[Union[Dataset, dict[str, Dataset]]] = None,
processing_class: Optional[
Union[PreTrainedTokenizerBase, BaseImageProcessor, FeatureExtractionMixin, ProcessorMixin]
] = None,
compute_loss_func: Optional[Callable] = None,
compute_metrics: Optional[Callable[[EvalPrediction], dict]] = None,
callbacks: Optional[list[TrainerCallback]] = None,
optimizers: tuple[Optional[torch.optim.Optimizer], Optional[torch.optim.lr_scheduler.LambdaLR]] = (None, None),
optimizer_cls_and_kwargs: Optional[tuple[Type[torch.optim.Optimizer], dict[str, Any]]] = None,
preprocess_logits_for_metrics: Optional[Callable[[torch.Tensor, torch.Tensor], torch.Tensor]] = None,
peft_config: Optional["PeftConfig"] = None,
formatting_func: Optional[Union[Callable[[dict], str], Callable[[dict], list[str]]]] = None,
):
# Args
if args is None:
model_name = model if isinstance(model, str) else model.config._name_or_path
model_name = model_name.split("/")[-1]
args = SFTConfig(f"{model_name}-SFT")
elif isinstance(args, TrainingArguments) and not isinstance(args, SFTConfig):
dict_args = args.to_dict()
dict_args["hub_token"] = args.hub_token # to_dict hides the hub_token
dict_args.pop("push_to_hub_token")
args = SFTConfig(**dict_args)
# Model
if args.model_init_kwargs is not None and not isinstance(model, str):
warnings.warn(
"You passed model_init_kwargs to the `SFTConfig`, but your model is already instantiated. "
"The `model_init_kwargs` will be ignored."
)
if isinstance(model, str):
model = self._create_model_from_path(model, args)
# PEFT configuration and model wrapping
if peft_config is not None:
model = self._prepare_peft_model(model, peft_config, args)
# 3. Handle the tokenizer
if processing_class is None:
processing_class = AutoTokenizer.from_pretrained(model.config._name_or_path)
if processing_class.pad_token is None:
processing_class.pad_token = processing_class.eos_token # required for padding when collating data
# Dataset
preprocess_dataset = args.dataset_kwargs is None or not args.dataset_kwargs.get("skip_prepare_dataset", False)
if preprocess_dataset:
train_dataset = self._prepare_dataset(
train_dataset, processing_class, args, args.packing, formatting_func, "train"
)
if eval_dataset is not None:
packing = args.packing if args.eval_packing is None else args.eval_packing
if isinstance(eval_dataset, dict):
eval_dataset = {
key: self._prepare_dataset(dataset, processing_class, args, packing, formatting_func, key)
for key, dataset in eval_dataset.items()
}
else:
eval_dataset = self._prepare_dataset(
eval_dataset, processing_class, args, packing, formatting_func, "eval"
)
# Data collator
if data_collator is None:
data_collator = DataCollatorForLanguageModeling(tokenizer=processing_class, mlm=False)
# Initialize the metrics
self._metrics = defaultdict(list)
# Initialize the Trainer. Parent class will handle:
# - DeepSpeed configuration (through create_accelerator_and_postprocess)
# - FSDP setup
# - Distributed training setup
# - Optimizer and scheduler creation
# Some arguments are only available for transformers>=4.47.0. Can be removed when the min version is bumped.
super_init_kwargs = {}
if version.parse(transformers.__version__) >= version.parse("4.47.0.dev0"):
super_init_kwargs["optimizer_cls_and_kwargs"] = optimizer_cls_and_kwargs
else:
if optimizer_cls_and_kwargs is not None:
warnings.warn(
"The `optimizer_cls_and_kwargs` argument is only available for `transformers>=4.47.0`. "
"The default optimizer will be used. "
"Remove the `optimizer_cls_and_kwargs` or upgrade to `transformers>=4.47.0`."
)
super().__init__(
model=model,
args=args,
data_collator=data_collator,
train_dataset=train_dataset,
eval_dataset=eval_dataset,
processing_class=processing_class,
compute_loss_func=compute_loss_func,
compute_metrics=compute_metrics,
callbacks=callbacks,
optimizers=optimizers,
preprocess_logits_for_metrics=preprocess_logits_for_metrics,
**super_init_kwargs,
)
# Add tags for models that have been loaded with the correct transformers version
if hasattr(self.model, "add_model_tags"):
self.model.add_model_tags(self._tag_names)
def _create_model_from_path(self, model_path: str, args: SFTConfig) -> PreTrainedModel:
"""Creates a model from a path or model identifier."""
model_init_kwargs = args.model_init_kwargs or {}
# Handle torch dtype
torch_dtype = model_init_kwargs.get("torch_dtype")
if isinstance(torch_dtype, torch.dtype) or torch_dtype == "auto" or torch_dtype is None:
pass # torch_dtype is already a torch.dtype or "auto" or None
elif isinstance(torch_dtype, str): # it's a str, but not "auto"
torch_dtype = getattr(torch, torch_dtype)
model_init_kwargs["torch_dtype"] = torch_dtype
else:
raise ValueError(
"Invalid `torch_dtype` passed to `SFTConfig`. Expected either 'auto' or a string representing "
f"a `torch.dtype` (e.g., 'float32'), but got {torch_dtype}."
)
# Disable caching if gradient checkpointing is enabled (not supported)
if args.gradient_checkpointing:
model_init_kwargs["use_cache"] = False
# Create model
if args.use_liger:
if not is_liger_kernel_available():
raise ImportError("Please install Liger-kernel for use_liger=True")
return AutoLigerKernelForCausalLM.from_pretrained(model_path, **model_init_kwargs)
return AutoModelForCausalLM.from_pretrained(model_path, **model_init_kwargs)
def _prepare_peft_model(self, model: PreTrainedModel, peft_config: Any, args: SFTConfig) -> PreTrainedModel:
"""Prepares a model for PEFT training."""
if not is_peft_available():
raise ImportError("To use PeftModel, you need to install the `peft` library.")
if not isinstance(peft_config, PeftConfig):
raise ValueError(
f"Expected PeftConfig object but got {type(peft_config)}. If you want to use the PeftModel, you need "
"to pass a PeftConfig object to the SFTTrainer."
)
if isinstance(model, PeftModel):
return model
# Handle quantized models (QLoRA)
is_qlora = getattr(model, "is_loaded_in_4bit", False) or getattr(model, "is_loaded_in_8bit", False)
is_sharded_qlora = False
if getattr(model, "is_loaded_in_4bit", False):
# Check if model is sharded (FSDP/DS-Zero3)
for _, param in model.named_parameters():
if param.__class__.__name__ == "Params4bit":
is_sharded_qlora = param.data.device.type in {"cpu", "meta"}
break
# Prepare model for kbit training if needed
if is_qlora and not is_sharded_qlora:
model = self._prepare_model_for_kbit_training(model, args)
# Disable gradient checkpointing as it's handled by prepare_model_for_kbit_training
args = dataclasses.replace(args, gradient_checkpointing=False)
elif args.gradient_checkpointing:
model = self._enable_gradient_checkpointing(model, args)
# Create PEFT model
if (
version.parse(peft.__version__) >= version.parse("0.12") # autocast_adapter_dtype introduced in 0.12
and getattr(model, "is_loaded_in_4bit", False)
and is_sharded_qlora
):
model = get_peft_model(model, peft_config, autocast_adapter_dtype=False)
else:
model = get_peft_model(model, peft_config)
# Handle bf16 casting for 4-bit models
if args.bf16 and getattr(model, "is_loaded_in_4bit", False) and not is_sharded_qlora:
peft_module_casting_to_bf16(model)
return model
def _prepare_model_for_kbit_training(self, model: PreTrainedModel, args: SFTConfig) -> PreTrainedModel:
"""Prepares a quantized model for kbit training."""
prepare_model_kwargs = {
"use_gradient_checkpointing": args.gradient_checkpointing,
"gradient_checkpointing_kwargs": args.gradient_checkpointing_kwargs or {},
}
return prepare_model_for_kbit_training(model, **prepare_model_kwargs)
def _enable_gradient_checkpointing(self, model: PreTrainedModel, args: SFTConfig) -> PreTrainedModel:
"""Enables gradient checkpointing for the model."""
gradient_checkpointing_kwargs = args.gradient_checkpointing_kwargs or {}
use_reentrant = (
"use_reentrant" not in gradient_checkpointing_kwargs or gradient_checkpointing_kwargs["use_reentrant"]
)
if use_reentrant:
if hasattr(model, "enable_input_require_grads"):
model.enable_input_require_grads()
else:
def make_inputs_require_grad(module, input, output):
output.requires_grad_(True)
model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
return model
def _prepare_dataset(
self,
dataset: Union[Dataset, IterableDataset],
processing_class: Union[PreTrainedTokenizerBase, BaseImageProcessor, FeatureExtractionMixin, ProcessorMixin],
args: SFTConfig,
packing: bool,
formatting_func: Optional[Callable[[dict], str]],
dataset_name: str,
) -> Union[Dataset, IterableDataset]:
# Convert the dataset to an IterableDataset if it is a ConstantLengthDataset
if isinstance(dataset, ConstantLengthDataset):
return dataset
# Build the kwargs for the `map` function
map_kwargs = {}
if isinstance(dataset, Dataset): # IterableDataset does not support num_proc
map_kwargs["num_proc"] = args.dataset_num_proc
with PartialState().local_main_process_first():
# Apply the formatting function if any
if formatting_func is not None:
if isinstance(dataset, Dataset): # `IterableDataset.map` does not support `desc`
map_kwargs["desc"] = f"Applying formatting function to {dataset_name} dataset"
batched = isinstance(formatting_func(next(iter(dataset))), list)
def _func(example):
return {"text": formatting_func(example)}
dataset = dataset.map(_func, batched=batched, **map_kwargs)
# If the dataset is prompt-completion, convert it to language modeling type
if "prompt" in dataset.column_names and "completion" in dataset.column_names:
key = "messages" if is_conversational(dataset[0]) else "text"
def concat_prompt_completion(example):
return {key: example["prompt"] + example["completion"]}
dataset = dataset.map(concat_prompt_completion, remove_columns=["prompt", "completion"])
# Apply the chat template if needed
if isinstance(dataset, Dataset): # `IterableDataset.map` does not support `desc`
map_kwargs["desc"] = f"Applying chat template to {dataset_name} dataset"
dataset = dataset.map(
maybe_apply_chat_template,
fn_kwargs={"tokenizer": processing_class},
remove_columns="messages" if "messages" in dataset.column_names else None, # renamed to "text"
**map_kwargs,
)
# Tokenize the dataset
if isinstance(dataset, Dataset): # `IterableDataset.map` does not support `desc`
map_kwargs["desc"] = f"Tokenizing {dataset_name} dataset"
dataset = dataset.map(lambda ex: processing_class(ex[args.dataset_text_field]), **map_kwargs)
# Pack or truncate
if packing:
if args.max_seq_length is None:
raise ValueError("When packing is enabled, `max_seq_length` can't be `None`.")
if isinstance(dataset, Dataset): # `IterableDataset.map` does not support `desc`
map_kwargs["desc"] = f"Packing {dataset_name} dataset"
dataset = dataset.select_columns("input_ids")
dataset = dataset.map(
pack_examples, batched=True, fn_kwargs={"seq_length": args.max_seq_length}, **map_kwargs
)
elif args.max_seq_length is not None:
dataset = dataset.map(
lambda ex: {key: ex[key][: args.max_seq_length] for key in ["input_ids", "attention_mask"]},
**map_kwargs,
)
# For Liger kernel, ensure only input_ids is present
if args.use_liger:
dataset = dataset.select_columns("input_ids")
return dataset
def compute_loss(self, model, inputs, return_outputs=False, num_items_in_batch=None):
"""
Compute training loss and additionally compute token accuracies
"""
(loss, outputs) = super().compute_loss(
model, inputs, return_outputs=True, num_items_in_batch=num_items_in_batch
)
# Compute token accuracy if we have labels and if the model is not using Liger (no logits)
if "labels" in inputs and not self.args.use_liger:
shift_logits = outputs.logits[..., :-1, :].contiguous()
shift_labels = inputs["labels"][..., 1:].contiguous()
# Gather logits and labels from all GPUs first
shift_logits = self.accelerator.gather_for_metrics(shift_logits)
shift_labels = self.accelerator.gather_for_metrics(shift_labels)
# Then compute accuracy on the gathered tensors
if self.accelerator.is_main_process:
accuracy = compute_token_accuracy(shift_logits, shift_labels)
self._metrics["mean_token_accuracy"].append(accuracy)
return (loss, outputs) if return_outputs else loss
def log(self, logs: dict[str, float], start_time: Optional[float] = None) -> None:
metrics = {key: sum(val) / len(val) for key, val in self._metrics.items()} # average the metrics
# This method can be called both in training and evaluation. When called in evaluation, the keys in `logs`
# start with "eval_". We need to add the prefix "eval_" to the keys in `metrics` to match the format.
if next(iter(logs.keys())).startswith("eval_"):
metrics = {f"eval_{key}": val for key, val in metrics.items()}
logs = {**logs, **metrics}
if version.parse(transformers.__version__) >= version.parse("4.47.0.dev0"):
super().log(logs, start_time)
else: # transformers<=4.46
super().log(logs)
self._metrics.clear()
def create_model_card(
self,
model_name: Optional[str] = None,
dataset_name: Optional[str] = None,
tags: Union[str, list[str], None] = None,
):
"""
Creates a draft of a model card using the information available to the `Trainer`.
Args:
model_name (`str` or `None`, *optional*, defaults to `None`):
Name of the model.
dataset_name (`str` or `None`, *optional*, defaults to `None`):
Name of the dataset used for training.
tags (`str`, `list[str]` or `None`, *optional*, defaults to `None`):
Tags to be associated with the model card.
"""
if not self.is_world_process_zero():
return
if hasattr(self.model.config, "_name_or_path") and not os.path.isdir(self.model.config._name_or_path):
base_model = self.model.config._name_or_path
else:
base_model = None
tags = tags or []
if isinstance(tags, str):
tags = [tags]
if hasattr(self.model.config, "unsloth_version"):
tags.append("unsloth")
model_card = generate_model_card(
base_model=base_model,
model_name=model_name,
hub_model_id=self.hub_model_id,
dataset_name=dataset_name,
tags=tags,
wandb_url=wandb.run.get_url() if is_wandb_available() and wandb.run is not None else None,
comet_url=get_comet_experiment_url(),
trainer_name="SFT",
)
model_card.save(os.path.join(self.args.output_dir, "README.md"))
| trl/trl/trainer/sft_trainer.py/0 | {
"file_path": "trl/trl/trainer/sft_trainer.py",
"repo_id": "trl",
"token_count": 10503
} |
# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This script tests to ensure that `accelerate` performs at the same level as raw `MS-AMP`.
This particular script verifies this for single GPU training.
"""
import evaluate
import msamp
import torch
from fp8_utils import evaluate_model, get_training_utilities
from accelerate import Accelerator
from accelerate.state import AcceleratorState
from accelerate.utils import FP8RecipeKwargs, get_grad_scaler, set_seed
MODEL_NAME = "bert-base-cased"
METRIC = evaluate.load("glue", "mrpc")
def train_baseline(opt_level="O2"):
set_seed(42)
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = get_training_utilities(MODEL_NAME)
model, optimizer = msamp.initialize(model, optimizer, opt_level=opt_level)
model.to("cuda")
base_model_results = evaluate_model(model, eval_dataloader, METRIC)
model.train()
scaler = get_grad_scaler()
for batch in train_dataloader:
batch = batch.to("cuda")
with torch.autocast(device_type="cuda", dtype=torch.bfloat16):
outputs = model(**batch)
loss = outputs.loss
loss = scaler.scale(loss)
loss.backward()
optimizer.step()
optimizer.zero_grad()
lr_scheduler.step()
trained_model_results = evaluate_model(model, eval_dataloader, METRIC)
assert (
trained_model_results["accuracy"] > base_model_results["accuracy"]
), f'Accuracy should be higher for the trained model: {trained_model_results["accuracy"]} > {base_model_results["accuracy"]}'
assert (
trained_model_results["f1"] > base_model_results["f1"]
), f'F1 score should be higher for the trained model: {trained_model_results["f1"]} > {base_model_results["f1"]}'
return base_model_results, trained_model_results
def train_integration(opt_level="O2"):
kwargs_handlers = [FP8RecipeKwargs(backend="msamp", opt_level=opt_level)]
AcceleratorState()._reset_state(True)
accelerator = Accelerator(mixed_precision="fp8", kwargs_handlers=kwargs_handlers)
set_seed(42)
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = get_training_utilities(
MODEL_NAME, accelerator=accelerator
)
model, optimizer, lr_scheduler = accelerator.prepare(model, optimizer, lr_scheduler)
base_model_results = evaluate_model(model, eval_dataloader, METRIC)
model.train()
for batch in train_dataloader:
outputs = model(**batch)
loss = outputs.loss
accelerator.backward(loss)
optimizer.step()
optimizer.zero_grad()
lr_scheduler.step()
trained_model_results = evaluate_model(model, eval_dataloader, METRIC)
assert (
trained_model_results["accuracy"] > base_model_results["accuracy"]
), f'Accuracy should be higher for the trained model: {trained_model_results["accuracy"]} > {base_model_results["accuracy"]}'
assert (
trained_model_results["f1"] > base_model_results["f1"]
), f'F1 score should be higher for the trained model: {trained_model_results["f1"]} > {base_model_results["f1"]}'
return base_model_results, trained_model_results
if __name__ == "__main__":
for opt_level in ["O1", "O2"]:
baseline_not_trained, baseline_trained = train_baseline(opt_level)
accelerator_not_trained, accelerator_trained = train_integration(opt_level)
assert (
baseline_not_trained["accuracy"] == accelerator_not_trained["accuracy"]
), f'Accuracy should be the same for the baseline and accelerator: {baseline_not_trained["accuracy"]} == {accelerator_not_trained["accuracy"]}'
assert (
baseline_not_trained["f1"] == accelerator_not_trained["f1"]
), f'F1 score should be the same for the baseline and accelerator: {baseline_not_trained["f1"]} == {accelerator_not_trained["f1"]}'
assert (
baseline_trained["accuracy"] == accelerator_trained["accuracy"]
), f'Accuracy should be the same for the baseline and accelerator: {baseline_trained["accuracy"]} == {accelerator_trained["accuracy"]}'
assert (
baseline_trained["f1"] == accelerator_trained["f1"]
), f'F1 score should be the same for the baseline and accelerator: {baseline_trained["f1"]} == {accelerator_trained["f1"]}'
| accelerate/benchmarks/fp8/ms_amp/non_distributed.py/0 | {
"file_path": "accelerate/benchmarks/fp8/ms_amp/non_distributed.py",
"repo_id": "accelerate",
"token_count": 1773
} |
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Installation
Before you start, you will need to setup your environment, install the appropriate packages, and configure Accelerate. Accelerate is tested on **Python 3.8+**.
Accelerate is available on pypi and conda, as well as on GitHub. Details to install from each are below:
## pip
To install Accelerate from pypi, perform:
```bash
pip install accelerate
```
## conda
Accelerate can also be installed with conda with:
```bash
conda install -c conda-forge accelerate
```
## Source
New features are added every day that haven't been released yet. To try them out yourself, install
from the GitHub repository:
```bash
pip install git+https://github.com/huggingface/accelerate
```
If you're working on contributing to the library or wish to play with the source code and see live
results as you run the code, an editable version can be installed from a locally-cloned version of the
repository:
```bash
git clone https://github.com/huggingface/accelerate
cd accelerate
pip install -e .
```
## Configuration
After installing, you need to configure Accelerate for how the current system is setup for training.
To do so run the following and answer the questions prompted to you:
```bash
accelerate config
```
To write a barebones configuration that doesn't include options such as DeepSpeed configuration or running on TPUs, you can quickly run:
```bash
python -c "from accelerate.utils import write_basic_config; write_basic_config(mixed_precision='fp16')"
```
Accelerate will automatically utilize the maximum number of GPUs available and set the mixed precision mode.
To check that your configuration looks fine, run:
```bash
accelerate env
```
An example output is shown below, which describes two GPUs on a single machine with no mixed precision being used:
```bash
- `Accelerate` version: 1.2.0.dev0
- Platform: Linux-6.8.0-47-generic-x86_64-with-glibc2.35
- `accelerate` bash location: /home/zach/miniconda3/envs/accelerate/bin/accelerate
- Python version: 3.10.13
- Numpy version: 1.26.4
- PyTorch version (GPU?): 2.5.1+cu124 (True)
- PyTorch XPU available: False
- PyTorch NPU available: False
- PyTorch MLU available: False
- PyTorch MUSA available: False
- System RAM: 187.91 GB
- GPU type: NVIDIA GeForce RTX 4090
- `Accelerate` default config:
- compute_environment: LOCAL_MACHINE
- distributed_type: MULTI_GPU
- mixed_precision: no
- use_cpu: False
- debug: False
- num_processes: 2
- machine_rank: 0
- num_machines: 1
- gpu_ids: all
- rdzv_backend: static
- same_network: True
- main_training_function: main
- enable_cpu_affinity: False
- downcast_bf16: no
- tpu_use_cluster: False
- tpu_use_sudo: False
- tpu_env: []
```
| accelerate/docs/source/basic_tutorials/install.md/0 | {
"file_path": "accelerate/docs/source/basic_tutorials/install.md",
"repo_id": "accelerate",
"token_count": 1129
} |
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Model memory estimator
One very difficult aspect when exploring potential models to use on your machine is knowing just how big of a model will *fit* into memory with your current graphics card (such as loading the model onto CUDA).
To help alleviate this, Accelerate has a CLI interface through `accelerate estimate-memory`. This tutorial will
help walk you through using it, what to expect, and at the end link to the interactive demo hosted on the Hub which will
even let you post those results directly on the model repo!
Currently we support searching for models that can be used in `timm` and `transformers`.
<Tip>
This API will load the model into memory on the `meta` device, so we are not actually downloading
and loading the full weights of the model into memory, nor do we need to. As a result it's
perfectly fine to measure 8 billion parameter models (or more), without having to worry about
if your CPU can handle it!
</Tip>
## Gradio Demos
Below are a few gradio demos related to what was described above. The first is the official Hugging Face memory estimation space, utilizing Accelerate directly:
<div class="block dark:hidden">
<iframe
src="https://hf-accelerate-model-memory-usage.hf.space?__theme=light"
width="850"
height="1600"
></iframe>
</div>
<div class="hidden dark:block">
<iframe
src="https://hf-accelerate-model-memory-usage.hf.space?__theme=dark"
width="850"
height="1600"
></iframe>
</div>
A community member has taken the idea and expanded it further, allowing you to filter models directly and see if you can run a particular LLM given GPU constraints and LoRA configurations. To play with it, see [here](https://huggingface.co/spaces/Vokturz/can-it-run-llm) for more details.
## The Command
When using `accelerate estimate-memory`, you need to pass in the name of the model you want to use, potentially the framework
that model utilizing (if it can't be found automatically), and the data types you want the model to be loaded in with.
For example, here is how we can calculate the memory footprint for `bert-base-cased`:
```bash
accelerate estimate-memory bert-base-cased
```
This will download the `config.json` for `bert-based-cased`, load the model on the `meta` device, and report back how much space
it will use:
Memory Usage for loading `bert-base-cased`:
| dtype | Largest Layer | Total Size | Training using Adam |
|---------|---------------|------------|---------------------|
| float32 | 84.95 MB | 418.18 MB | 1.61 GB |
| float16 | 42.47 MB | 206.59 MB | 826.36 MB |
| int8 | 21.24 MB | 103.29 MB | 413.18 MB |
| int4 | 10.62 MB | 51.65 MB | 206.59 MB |
By default it will return all the supported dtypes (`int4` through `float32`), but if you are interested in specific ones these can be filtered.
### Specific libraries
If the source library cannot be determined automatically (like it could in the case of `bert-base-cased`), a library name can
be passed in.
```bash
accelerate estimate-memory HuggingFaceM4/idefics-80b-instruct --library_name transformers
```
Memory Usage for loading `HuggingFaceM4/idefics-80b-instruct`:
| dtype | Largest Layer | Total Size | Training using Adam |
|---------|---------------|------------|---------------------|
| float32 | 3.02 GB | 297.12 GB | 1.16 TB |
| float16 | 1.51 GB | 148.56 GB | 594.24 GB |
| int8 | 772.52 MB | 74.28 GB | 297.12 GB |
| int4 | 386.26 MB | 37.14 GB | 148.56 GB |
```bash
accelerate estimate-memory timm/resnet50.a1_in1k --library_name timm
```
Memory Usage for loading `timm/resnet50.a1_in1k`:
| dtype | Largest Layer | Total Size | Training using Adam |
|---------|---------------|------------|---------------------|
| float32 | 9.0 MB | 97.7 MB | 390.78 MB |
| float16 | 4.5 MB | 48.85 MB | 195.39 MB |
| int8 | 2.25 MB | 24.42 MB | 97.7 MB |
| int4 | 1.12 MB | 12.21 MB | 48.85 MB |
### Specific dtypes
As mentioned earlier, while we return `int4` through `float32` by default, any dtype can be used from `float32`, `float16`, `int8`, and `int4`.
To do so, pass them in after specifying `--dtypes`:
```bash
accelerate estimate-memory bert-base-cased --dtypes float32 float16
```
Memory Usage for loading `bert-base-cased`:
| dtype | Largest Layer | Total Size | Training using Adam |
|---------|---------------|------------|---------------------|
| float32 | 84.95 MB | 413.18 MB | 1.61 GB |
| float16 | 42.47 MB | 206.59 MB | 826.36 MB |
## Caveats with this calculator
This calculator will tell you how much memory is needed to purely load the model in, *not* to perform inference.
This calculation is accurate within a few % of the actual value, so it is a very good view of just how much memory it will take. For instance loading `bert-base-cased` actually takes `413.68 MB` when loaded on CUDA in full precision, and the calculator estimates `413.18 MB`.
When performing inference you can expect to add up to an additional 20% as found by [EleutherAI](https://blog.eleuther.ai/transformer-math/). We'll be conducting research into finding a more accurate estimate to these values, and will update
this calculator once done.
| accelerate/docs/source/usage_guides/model_size_estimator.md/0 | {
"file_path": "accelerate/docs/source/usage_guides/model_size_estimator.md",
"repo_id": "accelerate",
"token_count": 2017
} |
# This config template is for a multi-node setup. This assumes DDP, but can be interop'd with the other configs in this folder
# Generally it's recommended to look at the SLURM config template for a more robust multi-node setup
distributed_type: MULTI_GPU
# We need to specify the current machine's rank
machine_rank: 0
# We then need to specify the IP address and port of the main process
main_process_ip: '1234'
main_process_port: 9999
# We need to specify the number of machines
num_machines: 2
# We need to specify the *total* number of processes
num_processes: 8
# And then we need to specify how rdvz comms will be handled
rdzv_backend: static # or c10d
# If the compute nodes are on the same network (cloud will more than likely be false)
same_network: false
| accelerate/examples/config_yaml_templates/multi_node.yaml/0 | {
"file_path": "accelerate/examples/config_yaml_templates/multi_node.yaml",
"repo_id": "accelerate",
"token_count": 216
} |
# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import time
import torch
from transformers import AutoModelForMaskedLM
from accelerate import PartialState, prepare_pippy
from accelerate.utils import set_seed
# Set the random seed to have reproducable outputs
set_seed(42)
# Create an example model
model = AutoModelForMaskedLM.from_pretrained("bert-base-uncased")
model.eval()
# Input configs
# Create example inputs for the model
input = torch.randint(
low=0,
high=model.config.vocab_size,
size=(1, 512), # bs x seq_len
device="cpu",
dtype=torch.int64,
requires_grad=False,
)
# Create a pipeline stage from the model
# Using `auto` is equivalent to letting `device_map="auto"` figure
# out device mapping and will also split the model according to the
# number of total GPUs available if it fits on one GPU
model = prepare_pippy(model, split_points="auto", example_args=(input,))
# You can pass `gather_output=True` to have the output from the model
# available on all GPUs
# model = prepare_pippy(model, split_points="auto", example_args=(input,), gather_output=True)
# Create new inputs of the expected size (n_processes)
input = torch.randint(
low=0,
high=model.config.vocab_size,
size=(2, 512), # bs x seq_len
device="cpu",
dtype=torch.int64,
requires_grad=False,
)
# Move the inputs to the first device
input = input.to("cuda:0")
# Take an average of 5 times
# Measure first batch
torch.cuda.synchronize()
start_time = time.time()
with torch.no_grad():
output = model(input)
torch.cuda.synchronize()
end_time = time.time()
first_batch = end_time - start_time
# Now that CUDA is init, measure after
torch.cuda.synchronize()
start_time = time.time()
for i in range(5):
with torch.no_grad():
output = model(input)
torch.cuda.synchronize()
end_time = time.time()
# The outputs are only on the final process by default
if PartialState().is_last_process:
output = torch.stack(tuple(output[0]))
print(f"Time of first pass: {first_batch}")
print(f"Average time per batch: {(end_time - start_time) / 5}")
PartialState().destroy_process_group()
| accelerate/examples/inference/pippy/bert.py/0 | {
"file_path": "accelerate/examples/inference/pippy/bert.py",
"repo_id": "accelerate",
"token_count": 871
} |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from manim import *
class Stage4(Scene):
def construct(self):
mem = Rectangle(height=0.5,width=0.5)
fill = Rectangle(height=0.46,width=0.46).set_stroke(width=0)
meta_mem = Rectangle(height=0.25,width=0.25)
cpu_left_col_base = [mem.copy() for i in range(6)]
cpu_right_col_base = [mem.copy() for i in range(6)]
cpu_left_col = VGroup(*cpu_left_col_base).arrange(UP, buff=0)
cpu_right_col = VGroup(*cpu_right_col_base).arrange(UP, buff=0)
cpu_rects = VGroup(cpu_left_col,cpu_right_col).arrange(RIGHT, buff=0)
cpu_text = Text("CPU", font_size=24)
cpu = Group(cpu_rects,cpu_text).arrange(DOWN, buff=0.5, aligned_edge=DOWN)
cpu.move_to([-2.5,-.5,0])
self.add(cpu)
gpu_base = [mem.copy() for i in range(4)]
gpu_rect = VGroup(*gpu_base).arrange(UP,buff=0)
gpu_text = Text("GPU", font_size=24)
gpu = Group(gpu_rect,gpu_text).arrange(DOWN, buff=0.5, aligned_edge=DOWN)
gpu.move_to([-1,-1,0])
self.add(gpu)
model_base = [mem.copy() for i in range(6)]
model_rect = VGroup(*model_base).arrange(RIGHT,buff=0)
model_text = Text("Model", font_size=24)
model = Group(model_rect,model_text).arrange(DOWN, buff=0.5, aligned_edge=DOWN)
model.move_to([3, -1., 0])
self.add(model)
model_cpu_arr = []
model_meta_arr = []
for i,rect in enumerate(model_base):
rect.set_stroke(YELLOW)
cpu_target = Rectangle(height=0.46/4,width=0.46/3).set_stroke(width=0.).set_fill(YELLOW, opacity=0.7)
if i == 0:
cpu_target.next_to(cpu_left_col_base[0].get_corner(DOWN+LEFT), buff=0.02, direction=UP)
cpu_target.set_x(cpu_target.get_x()+0.1)
elif i == 3:
cpu_target.next_to(model_cpu_arr[0], direction=UP, buff=0.)
else:
cpu_target.next_to(model_cpu_arr[i-1], direction=RIGHT, buff=0.)
self.add(cpu_target)
model_cpu_arr.append(cpu_target)
self.add(*model_cpu_arr, *model_meta_arr)
disk_left_col_base = [meta_mem.copy() for i in range(6)]
disk_right_col_base = [meta_mem.copy() for i in range(6)]
disk_left_col = VGroup(*disk_left_col_base).arrange(UP, buff=0)
disk_right_col = VGroup(*disk_right_col_base).arrange(UP, buff=0)
disk_rects = VGroup(disk_left_col,disk_right_col).arrange(RIGHT, buff=0)
disk_text = Text("Disk", font_size=24)
disk = Group(disk_rects,disk_text).arrange(DOWN, buff=0.5, aligned_edge=DOWN)
disk.move_to([-4.,-1.25,0])
self.add(disk_text, disk_rects)
cpu_disk_arr = []
for i in range(6):
target = fill.copy().set_fill(BLUE, opacity=0.8)
target.move_to(disk_left_col_base[i]).scale(0.5)
cpu_disk_arr.append(target)
self.add(*cpu_disk_arr)
key = Square(side_length=2.2)
key.move_to([-5, 2, 0])
key_text = MarkupText(
f"<b>Key:</b>\n\n<span fgcolor='{YELLOW}'>●</span> Empty Model",
font_size=18,
)
key_text.move_to([-5, 2.4, 0])
self.add(key_text, key)
blue_text = MarkupText(
f"<span fgcolor='{BLUE}'>●</span> Checkpoint",
font_size=18,
)
blue_text.next_to(key_text, DOWN*2.4, aligned_edge=key_text.get_left())
self.add(blue_text)
step_5 = MarkupText(
f'The offloaded weights are all sent to the CPU.',
font_size=24
)
step_5.move_to([2, 2, 0])
self.play(Write(step_5, run_time=3))
for i in range(6):
rect = cpu_disk_arr[i]
cp2 = rect.copy().set_fill(BLUE, opacity=0.8).scale(2.0)
cp2.generate_target()
cp2.target.move_to(model_base[i])
if i == 0:
rect.set_fill(BLUE, opacity=0.8)
rect.generate_target()
rect.target.move_to(cpu_left_col_base[0]).scale(2.0)
self.remove(*model_meta_arr,
*model_cpu_arr,
)
else:
rect.generate_target()
rect.target.move_to(cpu_left_col_base[i]).scale(2.0)
self.play(
MoveToTarget(rect),
MoveToTarget(cp2),
model_base[i].animate.set_stroke(WHITE)
)
self.play(FadeOut(step_5))
step_5 = MarkupText(
f'Finally, hooks are added to each weight in the model\nto transfer the weights from CPU to GPU\n\t\tand back when needed.',
font_size=24
)
step_5.move_to([2, 2, 0])
self.play(Write(step_5, run_time=3))
arrows = []
animations = []
for i in range(6):
a = Arrow(start=UP, end=DOWN, color=RED, buff=.5)
a.next_to(model_base[i].get_left(), UP, buff=0.2)
arrows.append(a)
animations.append(Write(a))
self.play(*animations)
self.wait() | accelerate/manim_animations/big_model_inference/stage_4.py/0 | {
"file_path": "accelerate/manim_animations/big_model_inference/stage_4.py",
"repo_id": "accelerate",
"token_count": 2919
} |
# Copyright 2022 The HuggingFace Team and Brian Chao. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This file contains utilities for handling input from the user and registering specific keys to specific functions,
based on https://github.com/bchao1/bullet
"""
from typing import List
from .keymap import KEYMAP, get_character
def mark(key: str):
"""
Mark the function with the key code so it can be handled in the register
"""
def decorator(func):
handle = getattr(func, "handle_key", [])
handle += [key]
func.handle_key = handle
return func
return decorator
def mark_multiple(*keys: List[str]):
"""
Mark the function with the key codes so it can be handled in the register
"""
def decorator(func):
handle = getattr(func, "handle_key", [])
handle += keys
func.handle_key = handle
return func
return decorator
class KeyHandler(type):
"""
Metaclass that adds the key handlers to the class
"""
def __new__(cls, name, bases, attrs):
new_cls = super().__new__(cls, name, bases, attrs)
if not hasattr(new_cls, "key_handler"):
new_cls.key_handler = {}
new_cls.handle_input = KeyHandler.handle_input
for value in attrs.values():
handled_keys = getattr(value, "handle_key", [])
for key in handled_keys:
new_cls.key_handler[key] = value
return new_cls
@staticmethod
def handle_input(cls):
"Finds and returns the selected character if it exists in the handler"
char = get_character()
if char != KEYMAP["undefined"]:
char = ord(char)
handler = cls.key_handler.get(char)
if handler:
cls.current_selection = char
return handler(cls)
else:
return None
def register(cls):
"""Adds KeyHandler metaclass to the class"""
return KeyHandler(cls.__name__, cls.__bases__, cls.__dict__.copy())
| accelerate/src/accelerate/commands/menu/input.py/0 | {
"file_path": "accelerate/src/accelerate/commands/menu/input.py",
"repo_id": "accelerate",
"token_count": 947
} |
# Copyright 2021 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import annotations
import logging
import os
import threading
import warnings
from contextlib import contextmanager
from functools import partial
from typing import Any, Callable, Optional
import torch
from .utils import (
DistributedType,
DynamoBackend,
GradientAccumulationPlugin,
check_cuda_p2p_ib_support,
check_fp8_capability,
deepspeed_required,
get_ccl_version,
get_cpu_distributed_information,
get_int_from_env,
is_ccl_available,
is_datasets_available,
is_deepspeed_available,
is_fp8_available,
is_ipex_available,
is_mlu_available,
is_mps_available,
is_musa_available,
is_npu_available,
is_torch_xla_available,
is_xpu_available,
parse_choice_from_env,
parse_flag_from_env,
set_numa_affinity,
)
from .utils.dataclasses import SageMakerDistributedType
if is_torch_xla_available():
import torch_xla.core.xla_model as xm
if is_mlu_available(check_device=False):
import torch_mlu # noqa: F401
if is_musa_available(check_device=False):
import torch_musa # noqa: F401
if is_npu_available(check_device=False):
import torch_npu # noqa: F401
logger = logging.getLogger(__name__)
def is_initialized() -> bool:
"""
Checks if the `AcceleratorState` has been initialized from `Accelerator`. Same as `AcceleratorState.initialized`,
but works as a module method.
"""
return AcceleratorState._shared_state != {}
# Lambda function that does nothing
def do_nothing(*args, **kwargs):
return None
class ThreadLocalSharedDict(threading.local):
"""
Descriptor that holds a dict shared between instances of a class in the same thread.
Note: Descriptors have slightly different semantics than just a dict field on its own.
`PartialState(...)._shared_state` and `PartialState._shared_state` (instance vs class) give the same value: the
underlying _storage dict. Likewise, `PartialState(...)._shared_state = {...}` overrides the _storage dict inside
the descriptor as you would expect. However, `PartialState._shared_state = {}` actually replaces the descriptor
object with a dict instead Thus, you should modify the _storage dict in-place (e.g. `_shared_state.clear()`).
See Python documentation for an explanation of descriptors: https://docs.python.org/3/howto/descriptor.html
This is required for using PyTorch/XLA with PJRT in multithreaded mode (required for TPU v2 and v3).
See https://github.com/pytorch/xla/blob/r2.0/docs/pjrt.md#multithreading-on-tpu-v2v3
"""
def __init__(self, thread_local: bool = False):
self._storage = {}
def __get__(self, obj, objtype=None):
return self._storage
def __set__(self, obj, value):
self._storage = value
# Prefer global shared dictionary, except when using TPU.
SharedDict = dict if not is_torch_xla_available() else ThreadLocalSharedDict
# Inspired by Alex Martelli's 'Borg'.
class PartialState:
"""
Singleton class that has information about the current training environment and functions to help with process
control. Designed to be used when only process control and device execution states are needed. Does *not* need to
be initialized from `Accelerator`.
Args:
cpu (`bool`, *optional*):
Whether or not to force the script to execute on CPU. Will ignore any accelerators available if set to
`True` and force the execution on the CPU.
kwargs (additional keyword arguments, *optional*):
Additional keyword arguments to pass to the relevent `init_process_group` function. Valid `kwargs` can be
found in [`utils.InitProcessGroupKwargs`]. See the example section for detailed usage.
**Available attributes:**
- **device** (`torch.device`) -- The device to use.
- **distributed_type** ([`~accelerate.state.DistributedType`]) -- The type of distributed environment currently
in use.
- **local_process_index** (`int`) -- The index of the current process on the current server.
- **mixed_precision** (`str`) -- Whether or not the current script will use mixed precision, and if so the type
of mixed precision being performed. (Choose from 'no','fp16','bf16 or 'fp8').
- **num_processes** (`int`) -- The number of processes currently launched in parallel.
- **process_index** (`int`) -- The index of the current process.
- **is_last_process** (`bool`) -- Whether or not the current process is the last one.
- **is_main_process** (`bool`) -- Whether or not the current process is the main one.
- **is_local_main_process** (`bool`) -- Whether or not the current process is the main one on the local node.
- **debug** (`bool`) -- Whether or not the current script is being run in debug mode.
Example:
```python
from accelerate.utils import InitProcessGroupKwargs
# To include `InitProcessGroupKwargs`, init then call `.to_kwargs()`
kwargs = InitProcessGroupKwargs(...).to_kwargs()
state = PartialState(**kwargs)
```
"""
_shared_state = SharedDict()
_known_attrs = [
"_cpu",
"_mixed_precision",
"_shared_state",
"backend",
"debug",
"device",
"distributed_type",
"fork_launched",
"local_process_index",
"num_processes",
"process_index",
]
def __init__(self, cpu: bool = False, **kwargs):
self.__dict__ = self._shared_state
if not self.initialized:
self._cpu = cpu
self.backend = None
env_device = os.environ.get("ACCELERATE_TORCH_DEVICE", None)
self.device = torch.device(env_device) if env_device is not None else None
self.debug = parse_flag_from_env("ACCELERATE_DEBUG_MODE")
use_sagemaker_dp = kwargs.pop("_use_sagemaker_dp", None)
dist_information = None
if use_sagemaker_dp is None:
use_sagemaker_dp = (
os.environ.get("ACCELERATE_USE_SAGEMAKER", "false") == "true"
and os.environ.get("ACCELERATE_SAGEMAKER_DISTRIBUTED_TYPE") != SageMakerDistributedType.NO
)
# Sets up self.backend + imports
original_backend = kwargs.pop("backend", None)
backend, distributed_type = self._prepare_backend(cpu, use_sagemaker_dp, original_backend)
if original_backend is not None and backend != original_backend:
raise ValueError(f"Your assigned backend {original_backend} is not avaliable, please use {backend}")
self.backend = backend
self.distributed_type = distributed_type
use_deepspeed = False
if not cpu and self.backend != "xla":
if int(os.environ.get("LOCAL_RANK", -1)) != -1:
# Deal with spawning deepspeed
if os.environ.get("ACCELERATE_USE_DEEPSPEED", "false") == "true":
if not is_deepspeed_available():
raise ImportError(
"DeepSpeed is not available => install it using `pip3 install deepspeed` or build it from source"
)
from deepspeed import comm as dist
if not dist.is_initialized():
dist.init_distributed(dist_backend=self.backend, auto_mpi_discovery=False, **kwargs)
# We need to flag to `use_deepspeed` to be True to override `distributed_type` later
use_deepspeed = True
# Deal with all other backends but XPU and CPU, that gets handled special later
elif (
self.distributed_type not in (DistributedType.MULTI_XPU, DistributedType.MULTI_CPU)
and not torch.distributed.is_initialized()
):
torch.distributed.init_process_group(backend=self.backend, **kwargs)
# XPU and CPU require special env configs to be set
if self.distributed_type in (DistributedType.MULTI_XPU, DistributedType.MULTI_CPU):
dist_information = get_cpu_distributed_information()
os.environ["RANK"] = str(dist_information.rank)
os.environ["WORLD_SIZE"] = str(dist_information.world_size)
os.environ["LOCAL_RANK"] = str(dist_information.local_rank)
os.environ["LOCAL_WORLD_SIZE"] = str(dist_information.local_world_size)
if not os.environ.get("MASTER_PORT", None):
os.environ["MASTER_PORT"] = "29500"
if (
not os.environ.get("MASTER_ADDR", None)
and dist_information.local_world_size != dist_information.world_size
and self.backend != "mpi"
):
raise ValueError(
"Tried to launch on distributed with multinode, but `MASTER_ADDR` env was not set, "
"please try exporting rank 0's hostname as `MASTER_ADDR`"
)
kwargs["rank"] = dist_information.rank
kwargs["world_size"] = dist_information.world_size
if (
self.distributed_type == DistributedType.MULTI_CPU
and get_int_from_env(["OMP_NUM_THREADS"], 0) == 0
):
import psutil
num_cpu_threads_per_process = int(
psutil.cpu_count(logical=False) / dist_information.local_world_size
)
if num_cpu_threads_per_process == 0:
num_cpu_threads_per_process = 1
torch.set_num_threads(num_cpu_threads_per_process)
warnings.warn(
f"OMP_NUM_THREADS/MKL_NUM_THREADS unset, we set it at {num_cpu_threads_per_process} to improve oob"
" performance."
)
if not torch.distributed.is_initialized():
torch.distributed.init_process_group(backend=self.backend, **kwargs)
# No backend == no distributed training
if self.backend is None:
self.distributed_type = DistributedType.NO
self.num_processes = 1
self.process_index = 0
self.local_process_index = 0
elif self.backend == "xla":
# XLA needs device setting first for `set_replication`
self.set_device()
xm.set_replication(self.device, xm.get_xla_supported_devices())
self.num_processes = xm.xrt_world_size()
self.process_index = xm.get_ordinal()
if is_torch_xla_available(check_is_tpu=True):
self.local_process_index = xm.get_local_ordinal()
else:
self.local_process_index = int(os.environ.get("LOCAL_RANK", -1))
else:
self.num_processes = torch.distributed.get_world_size()
self.process_index = torch.distributed.get_rank()
self.local_process_index = (
int(os.environ.get("LOCAL_RANK", -1)) if dist_information is None else dist_information.local_rank
)
self.set_device()
# Now we can change to deepseed
if use_deepspeed:
self.distributed_type = DistributedType.DEEPSPEED
# Set CPU affinity if enabled
if parse_flag_from_env("ACCELERATE_CPU_AFFINITY", False):
set_numa_affinity(self.local_process_index)
# Check for old RTX 4000's that can't use P2P or IB and are on old drivers
if self.device.type == "cuda" and not check_cuda_p2p_ib_support():
if "NCCL_P2P_DISABLE" not in os.environ or "NCCL_IB_DISABLE" not in os.environ:
raise NotImplementedError(
"Using RTX 4000 series doesn't support faster communication broadband via P2P or IB. "
'Please set `NCCL_P2P_DISABLE="1"` and `NCCL_IB_DISABLE="1" or use `accelerate launch` which '
"will do this automatically."
)
# Important: This should be the *only* code outside of `self.initialized!`
self.fork_launched = parse_flag_from_env("FORK_LAUNCHED", 0)
def __repr__(self) -> str:
return (
f"Distributed environment: {self.distributed_type}{(' Backend: ' + self.backend) if self.backend else ''}\n"
f"Num processes: {self.num_processes}\n"
f"Process index: {self.process_index}\n"
f"Local process index: {self.local_process_index}\n"
f"Device: {self.device}\n"
)
@staticmethod
def _reset_state():
"Resets `_shared_state`, is used internally and should not be called"
PartialState._shared_state.clear()
@property
def initialized(self) -> bool:
"Returns whether the `PartialState` has been initialized"
return self._shared_state != {}
@property
def use_distributed(self):
"""
Whether the Accelerator is configured for distributed training
"""
return self.distributed_type != DistributedType.NO and self.num_processes > 1
@property
def is_last_process(self) -> bool:
"Returns whether the current process is the last one"
return self.process_index == self.num_processes - 1
@property
def is_main_process(self) -> bool:
"Returns whether the current process is the main process"
return (
self.process_index == 0 if self.distributed_type != DistributedType.MEGATRON_LM else self.is_last_process
)
@property
def is_local_main_process(self) -> bool:
"Returns whether the current process is the main process on the local node"
return (
self.local_process_index == 0
if self.distributed_type != DistributedType.MEGATRON_LM
else self.is_last_process
)
def wait_for_everyone(self):
"""
Will stop the execution of the current process until every other process has reached that point (so this does
nothing when the script is only run in one process). Useful to do before saving a model.
Example:
```python
>>> # Assuming two GPU processes
>>> import time
>>> from accelerate.state import PartialState
>>> state = PartialState()
>>> if state.is_main_process:
... time.sleep(2)
>>> else:
... print("I'm waiting for the main process to finish its sleep...")
>>> state.wait_for_everyone()
>>> # Should print on every process at the same time
>>> print("Everyone is here")
```
"""
if self.distributed_type in (
DistributedType.MULTI_GPU,
DistributedType.MULTI_MLU,
DistributedType.MULTI_MUSA,
DistributedType.MULTI_NPU,
DistributedType.MULTI_XPU,
DistributedType.MULTI_CPU,
DistributedType.DEEPSPEED,
DistributedType.FSDP,
):
torch.distributed.barrier()
elif self.distributed_type == DistributedType.XLA:
xm.rendezvous("accelerate.utils.wait_for_everyone")
def _goes_first(self, is_main: bool):
if not is_main:
self.wait_for_everyone()
yield
if is_main:
self.wait_for_everyone()
@contextmanager
def split_between_processes(self, inputs: list | tuple | dict | torch.Tensor, apply_padding: bool = False):
"""
Splits `input` between `self.num_processes` quickly and can be then used on that process. Useful when doing
distributed inference, such as with different prompts.
Note that when using a `dict`, all keys need to have the same number of elements.
Args:
inputs (`list`, `tuple`, `torch.Tensor`, `dict` of `list`/`tuple`/`torch.Tensor`, or `datasets.Dataset`):
The input to split between processes.
apply_padding (`bool`, `optional`, defaults to `False`):
Whether to apply padding by repeating the last element of the input so that all processes have the same
number of elements. Useful when trying to perform actions such as `gather()` on the outputs or passing
in less inputs than there are processes. If so, just remember to drop the padded elements afterwards.
Example:
```python
# Assume there are two processes
from accelerate import PartialState
state = PartialState()
with state.split_between_processes(["A", "B", "C"]) as inputs:
print(inputs)
# Process 0
["A", "B"]
# Process 1
["C"]
with state.split_between_processes(["A", "B", "C"], apply_padding=True) as inputs:
print(inputs)
# Process 0
["A", "B"]
# Process 1
["C", "C"]
```
"""
if self.num_processes == 1:
yield inputs
return
length = len(inputs)
# Nested dictionary of any types
if isinstance(inputs, dict):
length = len(inputs[list(inputs.keys())[0]])
if not all(len(v) == length for v in inputs.values()):
raise ValueError("All values in the dictionary must have the same length")
num_samples_per_process, num_extras = divmod(length, self.num_processes)
start_index = self.process_index * num_samples_per_process + min(self.process_index, num_extras)
end_index = start_index + num_samples_per_process + (1 if self.process_index < num_extras else 0)
def _split_values(inputs, start_index, end_index):
if isinstance(inputs, (list, tuple, torch.Tensor)):
if start_index >= len(inputs):
result = inputs[-1:]
else:
result = inputs[start_index:end_index]
if apply_padding:
if isinstance(result, torch.Tensor):
from accelerate.utils import pad_across_processes, send_to_device
# The tensor needs to be on the device before we can pad it
tensorized_result = send_to_device(result, self.device)
result = pad_across_processes(tensorized_result, pad_index=inputs[-1])
else:
result += [result[-1]] * (num_samples_per_process + 1 - len(result))
return result
elif isinstance(inputs, dict):
for key in inputs.keys():
inputs[key] = _split_values(inputs[key], start_index, end_index)
return inputs
else:
if is_datasets_available():
from datasets import Dataset
if isinstance(inputs, Dataset):
if start_index >= len(inputs):
start_index = len(inputs) - 1
if end_index > len(inputs):
end_index = len(inputs)
result_idcs = list(range(start_index, end_index))
if apply_padding:
result_idcs += [end_index - 1] * (num_samples_per_process + 1 - len(result_idcs))
return inputs.select(result_idcs)
return inputs
yield _split_values(inputs, start_index, end_index)
@contextmanager
def main_process_first(self):
"""
Lets the main process go first inside a with block.
The other processes will enter the with block after the main process exits.
Example:
```python
>>> from accelerate import Accelerator
>>> accelerator = Accelerator()
>>> with accelerator.main_process_first():
... # This will be printed first by process 0 then in a seemingly
... # random order by the other processes.
... print(f"This will be printed by process {accelerator.process_index}")
```
"""
yield from self._goes_first(self.is_main_process)
@contextmanager
def local_main_process_first(self):
"""
Lets the local main process go inside a with block.
The other processes will enter the with block after the main process exits.
Example:
```python
>>> from accelerate.state import PartialState
>>> state = PartialState()
>>> with state.local_main_process_first():
... # This will be printed first by local process 0 then in a seemingly
... # random order by the other processes.
... print(f"This will be printed by process {state.local_process_index}")
```
"""
yield from self._goes_first(self.is_local_main_process)
def on_main_process(self, function: Callable[..., Any] = None):
"""
Decorator that only runs the decorated function on the main process.
Args:
function (`Callable`): The function to decorate.
Example:
```python
>>> from accelerate.state import PartialState
>>> state = PartialState()
>>> @state.on_main_process
... def print_something():
... print("This will be printed by process 0 only.")
>>> print_something()
"This will be printed by process 0 only"
```
"""
if not self.initialized:
raise ValueError("The `PartialState` or `Accelerator` must be initialized before calling this function.")
if self.is_main_process or not self.use_distributed:
return function
return do_nothing
def on_local_main_process(self, function: Callable[..., Any] = None):
"""
Decorator that only runs the decorated function on the local main process.
Args:
function (`Callable`): The function to decorate.
Example:
```python
# Assume we have 2 servers with 4 processes each.
from accelerate.state import PartialState
state = PartialState()
@state.on_local_main_process
def print_something():
print("This will be printed by process 0 only on each server.")
print_something()
# On server 1:
"This will be printed by process 0 only"
# On server 2:
"This will be printed by process 0 only"
```
"""
if self.is_local_main_process or not self.use_distributed:
return function
return do_nothing
def on_last_process(self, function: Callable[..., Any]):
"""
Decorator that only runs the decorated function on the last process.
Args:
function (`Callable`): The function to decorate.
Example:
```python
# Assume we have 4 processes.
from accelerate.state import PartialState
state = PartialState()
@state.on_last_process
def print_something():
print(f"Printed on process {state.process_index}")
print_something()
"Printed on process 3"
```
"""
if self.is_last_process or not self.use_distributed:
return function
return do_nothing
def on_process(self, function: Callable[..., Any] = None, process_index: int = None):
"""
Decorator that only runs the decorated function on the process with the given index.
Args:
function (`Callable`, `optional`):
The function to decorate.
process_index (`int`, `optional`):
The index of the process on which to run the function.
Example:
```python
# Assume we have 4 processes.
from accelerate.state import PartialState
state = PartialState()
@state.on_process(process_index=2)
def print_something():
print(f"Printed on process {state.process_index}")
print_something()
"Printed on process 2"
```
"""
if function is None:
return partial(self.on_process, process_index=process_index)
if (self.process_index == process_index) or (not self.use_distributed):
return function
return do_nothing
def on_local_process(self, function: Callable[..., Any] = None, local_process_index: int = None):
"""
Decorator that only runs the decorated function on the process with the given index on the current node.
Args:
function (`Callable`, *optional*):
The function to decorate.
local_process_index (`int`, *optional*):
The index of the local process on which to run the function.
Example:
```python
# Assume we have 2 servers with 4 processes each.
from accelerate import Accelerator
accelerator = Accelerator()
@accelerator.on_local_process(local_process_index=2)
def print_something():
print(f"Printed on process {accelerator.local_process_index}")
print_something()
# On server 1:
"Printed on process 2"
# On server 2:
"Printed on process 2"
```
"""
if function is None:
return partial(self.on_local_process, local_process_index=local_process_index)
if (self.local_process_index == local_process_index) or (not self.use_distributed):
return function
return do_nothing
def print(self, *args, **kwargs):
if self.is_local_main_process:
print(*args, **kwargs)
@property
def default_device(self) -> torch.device:
"""
Returns the default device which is:
- MPS if `torch.backends.mps.is_available()` and `torch.backends.mps.is_built()` both return True.
- CUDA if `torch.cuda.is_available()`
- MLU if `is_mlu_available()`
- MUSA if `is_musa_available()`
- NPU if `is_npu_available()`
- CPU otherwise
"""
if is_mps_available():
os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
return torch.device("mps")
elif is_mlu_available():
return torch.device("mlu")
elif is_musa_available():
return torch.device("musa")
# NPU should be checked before CUDA when using `transfer_to_npu`
# See issue #3020: https://github.com/huggingface/accelerate/issues/3020
elif is_npu_available():
return torch.device("npu")
elif torch.cuda.is_available():
return torch.device("cuda")
elif is_xpu_available():
return torch.device("xpu")
else:
return torch.device("cpu")
def _prepare_backend(
self, cpu: bool = False, sagemaker_dp=False, backend: str = None
) -> tuple[str, DistributedType]:
"Prepares any imports needed before initializing the distributed backend and sets `self.backend` properly"
distributed_type = None
if sagemaker_dp:
import smdistributed.dataparallel.torch.torch_smddp # noqa
backend = "smddp"
distributed_type = DistributedType.MULTI_GPU
elif is_torch_xla_available():
backend = "xla"
distributed_type = DistributedType.XLA
elif int(os.environ.get("LOCAL_RANK", -1)) != -1 and not cpu:
if is_mlu_available():
backend = "cncl"
distributed_type = DistributedType.MULTI_MLU
elif is_musa_available():
backend = "mccl"
distributed_type = DistributedType.MULTI_MUSA
# NPU should be checked before CUDA when using `transfer_to_npu`
# See issue #3020: https://github.com/huggingface/accelerate/issues/3020
elif is_npu_available():
backend = "hccl"
distributed_type = DistributedType.MULTI_NPU
elif torch.cuda.is_available():
if backend is None:
backend = "nccl"
distributed_type = DistributedType.MULTI_GPU
if distributed_type is None and (
int(os.environ.get("LOCAL_RANK", -1)) != -1
or get_int_from_env(["PMI_SIZE", "OMPI_COMM_WORLD_SIZE", "MV2_COMM_WORLD_SIZE", "WORLD_SIZE"], 1) > 1
):
if not cpu and is_xpu_available():
distributed_type = DistributedType.MULTI_XPU
else:
distributed_type = DistributedType.MULTI_CPU
if (
backend in (None, "ccl")
and is_ccl_available()
and (get_int_from_env(["CCL_WORKER_COUNT"], 0) > 0 or distributed_type == DistributedType.MULTI_XPU)
):
if get_ccl_version() >= "1.12":
import oneccl_bindings_for_pytorch # noqa: F401
else:
import torch_ccl # noqa: F401
backend = "ccl"
elif backend in (None, "mpi") and torch.distributed.is_mpi_available():
backend = "mpi"
else:
backend = "gloo"
if distributed_type is None:
distributed_type = DistributedType.NO
return backend, distributed_type
def set_device(self):
"""
Sets the device in `self.device` to the current distributed environment.
"""
if self.device is not None:
return
if self.distributed_type == DistributedType.NO:
self.device = torch.device("cpu") if self._cpu else self.default_device
return
device = str(self.distributed_type).split(".")[-1].replace("MULTI_", "").lower()
if device not in ("cpu", "gpu", "mlu", "musa", "npu", "xpu", "xla"):
raise ValueError(
f"Can't set device for {self.distributed_type} ({device}), verify we should be calling `_set_device()` for it!"
)
if device == "xla":
self.device = xm.xla_device()
else:
if device == "gpu":
device = "cuda"
device_module = getattr(torch, device)
device_index = self.local_process_index % device_module.device_count()
self.device = torch.device(device, device_index)
device_module.set_device(self.device)
def destroy_process_group(self, group=None):
"""
Destroys the process group. If one is not specified, the default process group is destroyed.
"""
if self.fork_launched and group is None:
return
# needed when using torch.distributed.init_process_group
if torch.distributed.is_initialized():
torch.distributed.destroy_process_group(group)
def __getattr__(self, name: str):
# By this point we know that no attributes of `self` contain `name`,
# so we just modify the error message
if name in self._known_attrs:
raise AttributeError(
f"`PartialState` object has no attribute `{name}`. "
"This happens if `PartialState._reset_state()` was called and "
"an `Accelerator` or `PartialState` was not reinitialized."
)
# Raise a typical AttributeError
raise AttributeError(f"'PartialState' object has no attribute '{name}'")
class AcceleratorState:
"""
Singleton class that has information about the current training environment.
**Available attributes:**
- **device** (`torch.device`) -- The device to use.
- **distributed_type** ([`~accelerate.state.DistributedType`]) -- The type of distributed environment currently
in use.
- **initialized** (`bool`) -- Whether or not the `AcceleratorState` has been initialized from `Accelerator`.
- **local_process_index** (`int`) -- The index of the current process on the current server.
- **mixed_precision** (`str`) -- Whether or not the current script will use mixed precision, and if so the type
of mixed precision being performed. (Choose from 'no','fp16','bf16 or 'fp8').
- **num_processes** (`int`) -- The number of processes currently launched in parallel.
- **process_index** (`int`) -- The index of the current process.
- **is_last_process** (`bool`) -- Whether or not the current process is the last one.
- **is_main_process** (`bool`) -- Whether or not the current process is the main one.
- **is_local_main_process** (`bool`) -- Whether or not the current process is the main one on the local node.
- **debug** (`bool`) -- Whether or not the current script is being run in debug mode.
"""
_shared_state = SharedDict()
_known_attrs = PartialState._known_attrs + [
"deepspeed_plugin",
"use_ipex",
"fsdp_plugin",
"megatron_lm_plugin",
"dynamo_plugin",
]
def __init__(
self,
mixed_precision: str = None,
cpu: bool = False,
dynamo_plugin=None,
deepspeed_plugin=None,
fsdp_plugin=None,
torch_tp_plugin=None,
megatron_lm_plugin=None,
_from_accelerator: bool = False,
**kwargs,
):
self.__dict__ = self._shared_state
if parse_flag_from_env("ACCELERATE_USE_CPU"):
cpu = True
if PartialState._shared_state == {}:
PartialState(cpu, **kwargs)
self.__dict__.update(PartialState._shared_state)
self._check_initialized(mixed_precision, cpu)
if not self.initialized:
self.deepspeed_plugins = None
self.use_ipex = None
self.torch_tp_plugin = torch_tp_plugin
mixed_precision = (
parse_choice_from_env("ACCELERATE_MIXED_PRECISION", "no")
if mixed_precision is None
else mixed_precision.lower()
)
if mixed_precision == "fp8":
if not is_fp8_available():
raise ValueError(
"Using `fp8` precision requires `transformer_engine` or `MS-AMP` to be installed."
)
elif not check_fp8_capability():
logger.warning(
f"The current device has compute capability of {torch.cuda.get_device_capability()} which is "
"insufficient for FP8 mixed precision training (requires a GPU Hopper/Ada Lovelace "
"or higher, compute capability of 8.9 or higher). Will use FP16 instead."
)
mixed_precision = "fp16"
self.dynamo_plugin = dynamo_plugin
if not _from_accelerator:
raise ValueError(
"Please make sure to properly initialize your accelerator via `accelerator = Accelerator()` "
"before using any functionality from the `accelerate` library."
)
# deepspeed handles mixed_precision using deepspeed_config
self._mixed_precision = "no" if self.distributed_type == DistributedType.DEEPSPEED else mixed_precision
if self.distributed_type == DistributedType.XLA and is_torch_xla_available(check_is_tpu=True):
if mixed_precision == "bf16":
if os.environ.get("ACCELERATE_DOWNCAST_BF16"):
os.environ["XLA_USE_BF16"] = str(0)
os.environ["XLA_DOWNCAST_BF16"] = str(1)
self.downcast_bfloat = True
else:
os.environ["XLA_USE_BF16"] = str(1)
os.environ["XLA_DOWNCAST_BF16"] = str(0)
self.downcast_bfloat = False
elif os.environ.get("ACCELERATE_USE_DEEPSPEED", "false") == "true" and not cpu:
self.deepspeed_plugins = deepspeed_plugin
self.distributed_type = DistributedType.DEEPSPEED
elif self.distributed_type in [
DistributedType.MULTI_GPU,
DistributedType.MULTI_MLU,
DistributedType.MULTI_MUSA,
DistributedType.MULTI_NPU,
DistributedType.MULTI_XPU,
]:
if os.environ.get("ACCELERATE_USE_FSDP", "false") == "true" or fsdp_plugin is not None:
self.distributed_type = DistributedType.FSDP
if self._mixed_precision != "no":
fsdp_plugin.set_mixed_precision(self._mixed_precision)
self.fsdp_plugin = fsdp_plugin
if os.environ.get("ACCELERATE_USE_MEGATRON_LM", "false") == "true" and self.distributed_type not in [
DistributedType.MULTI_XPU,
]:
self.distributed_type = DistributedType.MEGATRON_LM
megatron_lm_plugin.set_mixed_precision(self._mixed_precision)
self.megatron_lm_plugin = megatron_lm_plugin
if os.environ.get("ACCELERATE_USE_TP", "false") == "true" or self.torch_tp_plugin is not None:
self.distributed_type = DistributedType.TP
elif self.distributed_type in [DistributedType.MULTI_CPU, DistributedType.MULTI_XPU, DistributedType.NO]:
if is_ipex_available():
# check if user disables it explicitly
self.use_ipex = parse_flag_from_env("ACCELERATE_USE_IPEX", default=True)
else:
self.use_ipex = False
if (
self.dynamo_plugin.backend != DynamoBackend.NO
and self._mixed_precision == "no"
and self.device.type == "cuda"
):
torch.backends.cuda.matmul.allow_tf32 = True
if (
self.dynamo_plugin.backend != DynamoBackend.NO
and self._mixed_precision == "no"
and self.device.type == "musa"
):
torch.backends.musa.matmul.allow_tf32 = True
PartialState._shared_state["distributed_type"] = self.distributed_type
@property
def initialized(self) -> bool:
return self._shared_state != PartialState._shared_state
def __repr__(self):
repr = PartialState().__repr__() + f"\nMixed precision type: {self.mixed_precision}\n"
if self.distributed_type == DistributedType.DEEPSPEED:
repr += f"ds_config: {self.deepspeed_plugin.deepspeed_config}\n"
return repr
def _check_initialized(self, mixed_precision=None, cpu=None):
"Checks if a modification is trying to be made and the `AcceleratorState` has already been initialized"
if self.initialized:
err = "AcceleratorState has already been initialized and cannot be changed, restart your runtime completely and pass `{flag}` to `Accelerator()`."
if cpu and self.device.type != "cpu":
raise ValueError(err.format(flag="cpu=True"))
if (
mixed_precision is not None
and mixed_precision != self._mixed_precision
and self.distributed_type != DistributedType.DEEPSPEED
):
raise ValueError(err.format(flag=f"mixed_precision='{mixed_precision}'"))
@property
def mixed_precision(self):
if self.distributed_type == DistributedType.DEEPSPEED:
config = self.deepspeed_plugin.deepspeed_config
if config.get("fp16", {}).get("enabled", False):
mixed_precision = "fp16"
elif config.get("bf16", {}).get("enabled", False):
mixed_precision = "bf16"
else:
mixed_precision = "no"
else:
mixed_precision = self._mixed_precision
return mixed_precision
@staticmethod
def _reset_state(reset_partial_state: bool = False):
"Resets `_shared_state`, is used internally and should not be called"
AcceleratorState._shared_state.clear()
if reset_partial_state:
PartialState._reset_state()
def destroy_process_group(self, group=None):
"""
Destroys the process group. If one is not specified, the default process group is destroyed.
If `self.fork_lauched` is `True` and `group` is `None`, nothing happens.
"""
PartialState().destroy_process_group(group)
@property
def fork_launched(self):
return PartialState().fork_launched
@property
def use_distributed(self):
"""
Whether the Accelerator is configured for distributed training
"""
return PartialState().use_distributed
@property
def is_last_process(self) -> bool:
"Returns whether the current process is the last one"
return PartialState().is_last_process
@property
def is_main_process(self) -> bool:
"Returns whether the current process is the main process"
return PartialState().is_main_process
@property
def is_local_main_process(self) -> bool:
"Returns whether the current process is the main process on the local node"
return PartialState().is_local_main_process
def wait_for_everyone(self):
PartialState().wait_for_everyone()
@contextmanager
def split_between_processes(self, inputs: list | tuple | dict | torch.Tensor, apply_padding: bool = False):
"""
Splits `input` between `self.num_processes` quickly and can be then used on that process. Useful when doing
distributed inference, such as with different prompts.
Note that when using a `dict`, all keys need to have the same number of elements.
Args:
inputs (`list`, `tuple`, `torch.Tensor`, or `dict` of `list`/`tuple`/`torch.Tensor`):
The input to split between processes.
apply_padding (`bool`, `optional`, defaults to `False`):
Whether to apply padding by repeating the last element of the input so that all processes have the same
number of elements. Useful when trying to perform actions such as `gather()` on the outputs or passing
in less inputs than there are processes. If so, just remember to drop the padded elements afterwards.
Example:
```python
# Assume there are two processes
from accelerate.state import AcceleratorState
state = AcceleratorState()
with state.split_between_processes(["A", "B", "C"]) as inputs:
print(inputs)
# Process 0
["A", "B"]
# Process 1
["C"]
with state.split_between_processes(["A", "B", "C"], apply_padding=True) as inputs:
print(inputs)
# Process 0
["A", "B"]
# Process 1
["C", "C"]
```
"""
with PartialState().split_between_processes(inputs, apply_padding=apply_padding) as inputs:
yield inputs
@contextmanager
def main_process_first(self):
"""
Lets the main process go first inside a with block.
The other processes will enter the with block after the main process exits.
"""
with PartialState().main_process_first():
yield
@contextmanager
def local_main_process_first(self):
"""
Lets the local main process go inside a with block.
The other processes will enter the with block after the main process exits.
"""
with PartialState().local_main_process_first():
yield
@property
def deepspeed_plugin(self):
"""
Returns the currently active DeepSpeedPlugin.
If not using deepspeed, returns `None`.
"""
# To maintain original behavior, return None if not using deepspeed.
if self.distributed_type != DistributedType.DEEPSPEED:
return None
from accelerate.utils.deepspeed import get_active_deepspeed_plugin
return get_active_deepspeed_plugin(self)
@deepspeed_required
def get_deepspeed_plugin(self, name: str):
"""
Returns the DeepSpeedPlugin with the given plugin_key.
"""
return self.deepspeed_plugins[name]
@deepspeed_required
def select_deepspeed_plugin(self, name: str = None):
"""
Activates the DeepSpeedPlugin with the given `name`, and will disable all other plugins.
"""
for key, plugin in self.deepspeed_plugins.items():
if key != name:
plugin._unselect()
self.deepspeed_plugins[name].select(_from_accelerator_state=True)
def print(self, *args, **kwargs):
PartialState().print(*args, **kwargs)
def __getattr__(self, name: str):
# By this point we know that no attributes of `self` contain `name`,
# so we just modify the error message
if name in self._known_attrs:
raise AttributeError(
f"`AcceleratorState` object has no attribute `{name}`. "
"This happens if `AcceleratorState._reset_state()` was called and "
"an `Accelerator` or `PartialState` was not reinitialized."
)
# Raise a typical AttributeError
raise AttributeError(f"'AcceleratorState' object has no attribute '{name}'")
class GradientState:
"""
Singleton class that has information related to gradient synchronization for gradient accumulation
**Available attributes:**
- **end_of_dataloader** (`bool`) -- Whether we have reached the end the current dataloader
- **remainder** (`int`) -- The number of extra samples that were added from padding the dataloader
- **sync_gradients** (`bool`) -- Whether the gradients should be synced across all devices
- **active_dataloader** (`Optional[DataLoader]`) -- The dataloader that is currently being iterated over
- **dataloader_references** (`List[Optional[DataLoader]]`) -- A list of references to the dataloaders that are
being iterated over
- **num_steps** (`int`) -- The number of steps to accumulate over
- **adjust_scheduler** (`bool`) -- Whether the scheduler should be adjusted to account for the gradient
accumulation
- **sync_with_dataloader** (`bool`) -- Whether the gradients should be synced at the end of the dataloader
iteration and the number of total steps reset
- **is_xla_gradients_synced** (`bool`) -- Whether the XLA gradients have been synchronized. It is initialized
as false. Once gradients have been reduced before the optimizer step, this flag is set to true. Subsequently,
after each step, the flag is reset to false. FSDP will always synchronize the gradients, hence
is_xla_gradients_synced is always true.
"""
_shared_state = SharedDict()
def __init__(self, gradient_accumulation_plugin: Optional[GradientAccumulationPlugin] = None):
self.__dict__ = self._shared_state
if not self.initialized:
self.sync_gradients = True
self.active_dataloader = None
self.dataloader_references = [None]
self.plugin_kwargs = (
gradient_accumulation_plugin.to_kwargs() if gradient_accumulation_plugin is not None else {}
)
self._is_xla_gradients_synced = False
# Plugin args are different and can be updated
if gradient_accumulation_plugin is not None and self.plugin_kwargs != gradient_accumulation_plugin.to_kwargs():
self.plugin_kwargs = gradient_accumulation_plugin.to_kwargs()
@property
def num_steps(self) -> int:
"Returns the number of steps to accumulate over"
return self.plugin_kwargs.get("num_steps", 1)
@property
def adjust_scheduler(self) -> bool:
"Returns whether the scheduler should be adjusted"
return self.plugin_kwargs.get("adjust_scheduler", False)
@property
def sync_with_dataloader(self) -> bool:
"Returns whether the gradients should be synced at the end of the dataloader iteration and the number of total steps reset"
return self.plugin_kwargs.get("sync_with_dataloader", True)
@property
def initialized(self) -> bool:
"Returns whether the `GradientState` has been initialized"
return GradientState._shared_state != {}
@property
def end_of_dataloader(self) -> bool:
"Returns whether we have reached the end of the current dataloader"
if not self.in_dataloader:
return False
return self.active_dataloader.end_of_dataloader
@property
def remainder(self) -> int:
"Returns the number of extra samples that were added from padding the dataloader"
if not self.in_dataloader:
return -1
return self.active_dataloader.remainder
def __repr__(self):
return (
f"Sync Gradients: {self.sync_gradients}\n"
f"At end of current dataloader: {self.end_of_dataloader}\n"
f"Extra samples added: {self.remainder}\n"
f"Gradient accumulation plugin: {self.plugin_kwargs}\n"
)
@property
def is_xla_gradients_synced(self):
"Returns the value of is_xla_gradients_synced. FSDP will always synchronize the gradients, hence is_xla_gradients_synced is always true."
if parse_flag_from_env("ACCELERATE_USE_FSDP", default=False):
return True
return self._is_xla_gradients_synced
@is_xla_gradients_synced.setter
def is_xla_gradients_synced(self, is_synced):
"Set the _is_xla_gradients_synced attribute."
self._is_xla_gradients_synced = is_synced
def _set_sync_gradients(self, sync_gradients):
"Private function that sets whether gradients should be synchronized. Users should not have to call this."
self.sync_gradients = sync_gradients
# Allow grad-sync to automatically work on TPUs
if (
self.sync_gradients
and is_torch_xla_available(check_is_tpu=True)
and PartialState().distributed_type == DistributedType.XLA
):
xm.mark_step()
def _add_dataloader(self, dataloader):
"Private function that adds a dataloader to `self.dataloader_references` and sets `in_dataloader` to `True`. Users should not have to call this."
self.active_dataloader = dataloader
self.dataloader_references.append(self.active_dataloader)
def _remove_dataloader(self, dataloader):
"Private function that removes a dataloader from `self.dataloader_references` and sets `in_dataloader` to `False` if there are no more dataloaders. Users should not have to call this."
self.dataloader_references.remove(dataloader)
self.active_dataloader = self.dataloader_references[-1]
@property
def in_dataloader(self) -> bool:
"Returns whether the current process is in a dataloader"
return self.active_dataloader is not None
@staticmethod
def _reset_state():
"Resets `_shared_state`, is used internally and should not be called"
GradientState._shared_state.clear()
| accelerate/src/accelerate/state.py/0 | {
"file_path": "accelerate/src/accelerate/state.py",
"repo_id": "accelerate",
"token_count": 23024
} |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Test file to ensure that in general certain situational setups for notebooks work.
"""
import os
import time
from multiprocessing import Queue
from pytest import mark, raises
from torch.distributed.elastic.multiprocessing.errors import ChildFailedError
from accelerate import PartialState, notebook_launcher
from accelerate.test_utils import require_bnb
from accelerate.utils import is_bnb_available
def basic_function():
# Just prints the PartialState
print(f"PartialState:\n{PartialState()}")
def tough_nut_function(queue: Queue):
if queue.empty():
return
trial = queue.get()
if trial > 0:
queue.put(trial - 1)
raise RuntimeError("The nut hasn't cracked yet! Try again.")
print(f"PartialState:\n{PartialState()}")
def bipolar_sleep_function(sleep_sec: int):
state = PartialState()
if state.process_index % 2 == 0:
raise RuntimeError("I'm an even process. I don't like to sleep.")
else:
time.sleep(sleep_sec)
NUM_PROCESSES = int(os.environ.get("ACCELERATE_NUM_PROCESSES", 1))
def test_can_initialize():
notebook_launcher(basic_function, (), num_processes=NUM_PROCESSES)
@mark.skipif(NUM_PROCESSES < 2, reason="Need at least 2 processes to test static rendezvous backends")
def test_static_rdzv_backend():
notebook_launcher(basic_function, (), num_processes=NUM_PROCESSES, rdzv_backend="static")
@mark.skipif(NUM_PROCESSES < 2, reason="Need at least 2 processes to test c10d rendezvous backends")
def test_c10d_rdzv_backend():
notebook_launcher(basic_function, (), num_processes=NUM_PROCESSES, rdzv_backend="c10d")
@mark.skipif(NUM_PROCESSES < 2, reason="Need at least 2 processes to test fault tolerance")
def test_fault_tolerant(max_restarts: int = 3):
queue = Queue()
queue.put(max_restarts)
notebook_launcher(tough_nut_function, (queue,), num_processes=NUM_PROCESSES, max_restarts=max_restarts)
@mark.skipif(NUM_PROCESSES < 2, reason="Need at least 2 processes to test monitoring")
def test_monitoring(monitor_interval: float = 0.01, sleep_sec: int = 100):
start_time = time.time()
with raises(ChildFailedError, match="I'm an even process. I don't like to sleep."):
notebook_launcher(
bipolar_sleep_function,
(sleep_sec,),
num_processes=NUM_PROCESSES,
monitor_interval=monitor_interval,
)
assert time.time() - start_time < sleep_sec, "Monitoring did not stop the process in time."
@require_bnb
def test_problematic_imports():
with raises(RuntimeError, match="Please keep these imports"):
import bitsandbytes as bnb # noqa: F401
notebook_launcher(basic_function, (), num_processes=NUM_PROCESSES)
def main():
print("Test basic notebook can be ran")
test_can_initialize()
print("Test static rendezvous backend")
test_static_rdzv_backend()
print("Test c10d rendezvous backend")
test_c10d_rdzv_backend()
print("Test fault tolerant")
test_fault_tolerant()
print("Test monitoring")
test_monitoring()
if is_bnb_available():
print("Test problematic imports (bnb)")
test_problematic_imports()
if NUM_PROCESSES > 1:
PartialState().destroy_process_group()
if __name__ == "__main__":
main()
| accelerate/src/accelerate/test_utils/scripts/test_notebook.py/0 | {
"file_path": "accelerate/src/accelerate/test_utils/scripts/test_notebook.py",
"repo_id": "accelerate",
"token_count": 1371
} |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import argparse
import math
import os
from abc import ABC
from functools import partial
import torch
import torch.nn.functional as F
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
from torch.nn.parallel.distributed import DistributedDataParallel as torchDDP
from ..optimizer import AcceleratedOptimizer
from ..scheduler import AcceleratedScheduler
from .imports import is_megatron_lm_available
from .operations import recursively_apply, send_to_device
if is_megatron_lm_available():
from megatron.core import mpu, tensor_parallel
from megatron.core.distributed import DistributedDataParallel as LocalDDP
from megatron.core.distributed import finalize_model_grads
from megatron.core.enums import ModelType
from megatron.core.num_microbatches_calculator import get_num_microbatches
from megatron.core.optimizer import get_megatron_optimizer
from megatron.core.parallel_state import get_tensor_model_parallel_group, get_tensor_model_parallel_src_rank
from megatron.core.pipeline_parallel import get_forward_backward_func
from megatron.core.utils import get_model_config
from megatron.inference.text_generation.communication import broadcast_int_list, broadcast_tensor
from megatron.inference.text_generation.generation import (
beam_search_and_return_on_first_stage,
generate_tokens_probs_and_return_on_first_stage,
)
from megatron.legacy.data.dataset_utils import build_train_valid_test_datasets
from megatron.legacy.model import BertModel, Float16Module, GPTModel, T5Model
from megatron.legacy.model.classification import Classification
from megatron.training import (
get_args,
get_tensorboard_writer,
get_tokenizer,
print_rank_last,
)
from megatron.training.arguments import (
_add_data_args,
_add_validation_args,
core_transformer_config_from_args,
parse_args,
validate_args,
)
from megatron.training.checkpointing import load_args_from_checkpoint, load_checkpoint, save_checkpoint
from megatron.training.global_vars import set_global_variables
from megatron.training.initialize import (
_compile_dependencies,
_init_autoresume,
_initialize_distributed,
_set_random_seed,
set_jit_fusion_options,
write_args_to_tensorboard,
)
from megatron.training.tokenizer.tokenizer import _vocab_size_with_padding
from megatron.training.training import (
build_train_valid_test_data_iterators,
get_optimizer_param_scheduler,
num_floating_point_operations,
setup_model_and_optimizer,
train_step,
training_log,
)
from megatron.training.utils import (
average_losses_across_data_parallel_group,
calc_params_l2_norm,
get_ltor_masks_and_position_ids,
unwrap_model,
)
# model utilities
def model_provider_func(pre_process=True, post_process=True, add_encoder=True, add_decoder=True):
"""Build the model."""
args = get_args()
mode = "pre-training" if args.pretraining_flag else "fine-tuning"
if args.rank == 0:
print(f"Building {args.model_type_name} model in the {mode} mode.")
print(
"The Megatron LM model weights are initialized at random in `accelerator.prepare`. "
"Please use `accelerator.load_checkpoint` to load a pre-trained checkpoint matching the distributed setup."
)
config = core_transformer_config_from_args(args)
if args.model_type_name == "bert":
if args.pretraining_flag:
num_tokentypes = 2 if args.bert_binary_head else 0
model = BertModel(
config=config,
num_tokentypes=num_tokentypes,
add_binary_head=args.bert_binary_head,
parallel_output=True,
pre_process=pre_process,
post_process=post_process,
)
else:
model = Classification(
config=config,
num_classes=args.num_labels,
num_tokentypes=2,
pre_process=pre_process,
post_process=post_process,
)
elif args.model_type_name == "gpt":
model = GPTModel(
config=config,
num_tokentypes=0,
parallel_output=True,
pre_process=pre_process,
post_process=post_process,
)
elif args.model_type_name == "t5":
model = T5Model(
config=config,
num_tokentypes=0,
parallel_output=True,
pre_process=pre_process,
post_process=post_process,
add_encoder=add_encoder,
add_decoder=add_decoder,
)
else:
raise ValueError(f"Unsupported model type: {args.model_type_name}")
return model
def prepare_model_optimizer_scheduler(accelerator):
accelerator.print("Preparing model optimizer scheduler")
args = get_args()
if accelerator.state.megatron_lm_plugin.custom_prepare_model_function is not None:
if accelerator.state.megatron_lm_plugin.custom_model_provider_function is None:
raise ValueError(
"You must provide a `custom_model_provider_function` when using a `custom_prepare_model_function`."
)
custom_model_provider_func = accelerator.state.megatron_lm_plugin.custom_model_provider_function
model = accelerator.state.megatron_lm_plugin.custom_prepare_model_function(custom_model_provider_func)
optimizer = prepare_optimizer(accelerator, model)
scheduler = prepare_scheduler(accelerator, optimizer, scheduler=None)
else:
model_type = ModelType.encoder_or_decoder
if args.model_type_name == "t5":
model_type = ModelType.encoder_and_decoder
model_provider_func_ = model_provider_func
if accelerator.state.megatron_lm_plugin.custom_model_provider_function is not None:
model_provider_func_ = accelerator.state.megatron_lm_plugin.custom_model_provider_function
(model, optimizer, scheduler) = setup_model_and_optimizer(
model_provider_func_,
model_type,
no_wd_decay_cond=args.no_wd_decay_cond,
scale_lr_cond=args.scale_lr_cond,
lr_mult=args.lr_mult,
)
args.model_len = len(model)
return model, optimizer, scheduler
# dataloader utilities
class MegatronLMDummyDataLoader:
"""
Dummy dataloader presents model parameters or param groups, this is primarily used to follow conventional training
Args:
**dataset_kwargs: Megatron data arguments.
"""
def __init__(self, **dataset_kwargs):
parser = argparse.ArgumentParser()
parser = _add_data_args(parser)
parser = _add_validation_args(parser)
data_args = parser.parse_known_args()
self.dataset_args = vars(data_args[0])
self.dataset_args.update(dataset_kwargs)
self.dataset_args["megatron_dataset_flag"] = True
def set_megatron_data_args(self):
args = get_args()
for key, value in self.dataset_args.items():
old_value = getattr(args, key, "")
if old_value != value:
print(
f"WARNING: MegatronLMDummyDataLoader overriding arguments for "
f"{key}:{old_value} with {key}:{value}"
)
setattr(args, key, value)
def get_train_valid_test_datasets_provider(self, accelerator):
def train_valid_test_datasets_provider(train_val_test_num_samples):
"""Build train, valid, and test datasets."""
args = get_args()
dataset_args = {
"data_prefix": args.data_path if isinstance(args.data_path, (list, tuple)) else [args.data_path],
"splits_string": args.split,
"train_valid_test_num_samples": train_val_test_num_samples,
"seed": args.seed,
}
if args.model_type_name == "bert":
dataset_args.update(
{
"max_seq_length": args.seq_length,
"binary_head": args.bert_binary_head,
}
)
elif args.model_type_name == "gpt":
dataset_args.update(
{
"max_seq_length": args.seq_length,
}
)
elif args.model_type_name == "t5":
dataset_args.update(
{
"max_seq_length": args.encoder_seq_length,
"max_seq_length_dec": args.decoder_seq_length,
"dataset_type": "t5",
}
)
else:
raise ValueError(f"Unsupported model type: {args.model_type_name}")
train_ds, valid_ds, test_ds = build_train_valid_test_datasets(**dataset_args)
return train_ds, valid_ds, test_ds
if accelerator.state.megatron_lm_plugin.custom_megatron_datasets_provider_function is not None:
return accelerator.state.megatron_lm_plugin.custom_megatron_datasets_provider_function
try:
args = get_args()
# Use '--no-use-pep517 -e' to pip install nvidia's megatron from source
if args.model_type_name == "bert":
from pretrain_bert import train_valid_test_datasets_provider
train_valid_test_datasets_provider.is_distributed = True
return train_valid_test_datasets_provider
elif args.model_type_name == "gpt":
from pretrain_gpt import train_valid_test_datasets_provider
train_valid_test_datasets_provider.is_distributed = True
return train_valid_test_datasets_provider
elif args.model_type_name == "t5":
from pretrain_t5 import train_valid_test_datasets_provider
train_valid_test_datasets_provider.is_distributed = True
return train_valid_test_datasets_provider
except ImportError:
pass
return train_valid_test_datasets_provider
def build_train_valid_test_data_iterators(self, accelerator):
args = get_args()
train_valid_test_dataset_provider = self.get_train_valid_test_datasets_provider(accelerator)
if args.virtual_pipeline_model_parallel_size is not None:
train_data_iterator = []
valid_data_iterator = []
test_data_iterator = []
for i in range(getattr(args, "model_len", 0)):
mpu.set_virtual_pipeline_model_parallel_rank(i)
iterators = build_train_valid_test_data_iterators(train_valid_test_dataset_provider)
train_data_iterator.append(iterators[0])
valid_data_iterator.append(iterators[1])
test_data_iterator.append(iterators[2])
else:
train_data_iterator, valid_data_iterator, test_data_iterator = build_train_valid_test_data_iterators(
train_valid_test_dataset_provider
)
return train_data_iterator, valid_data_iterator, test_data_iterator
def _handle_megatron_data_iterator(accelerator, data_iterator):
class DummyMegatronDataloader:
def __iter__(self):
return self
def __next__(self):
return {}
is_data_iterator_empty = data_iterator is None
is_src_data_iterator_empty = torch.tensor(is_data_iterator_empty, dtype=torch.bool, device=accelerator.device)
torch.distributed.broadcast(
is_src_data_iterator_empty, get_tensor_model_parallel_src_rank(), group=get_tensor_model_parallel_group()
)
if not is_src_data_iterator_empty and is_data_iterator_empty:
return DummyMegatronDataloader()
return data_iterator
def prepare_data_loader(accelerator, dataloader):
accelerator.print("Preparing dataloader")
args = get_args()
if not args.megatron_dataset_flag:
from ..data_loader import _PYTORCH_DATALOADER_KWARGS, prepare_data_loader
micro_batch_size = args.micro_batch_size * args.num_micro_batches
kwargs = {k: getattr(dataloader, k, _PYTORCH_DATALOADER_KWARGS[k]) for k in _PYTORCH_DATALOADER_KWARGS}
if kwargs["batch_size"] is None:
if isinstance(kwargs["sampler"], torch.utils.data.BatchSampler):
kwargs["sampler"].batch_size = micro_batch_size
else:
del kwargs["sampler"]
del kwargs["shuffle"]
del kwargs["batch_size"]
kwargs["batch_sampler"].batch_size = micro_batch_size
else:
del kwargs["batch_sampler"]
kwargs["batch_size"] = micro_batch_size
dataloader = torch.utils.data.DataLoader(dataloader.dataset, **kwargs)
# split_batches:
# Megatron only needs to fetch different data between different dp groups,
# and does not need to split the data within the dp group.
return prepare_data_loader(
dataloader,
accelerator.device,
num_processes=mpu.get_data_parallel_world_size(),
process_index=mpu.get_data_parallel_rank(),
split_batches=False,
put_on_device=True,
rng_types=accelerator.rng_types.copy(),
dispatch_batches=accelerator.dispatch_batches,
)
else:
if args.consumed_samples is not None:
(
args.consumed_train_samples,
args.consumed_valid_samples,
args.consumed_test_samples,
) = args.consumed_samples
else:
args.consumed_train_samples, args.consumed_valid_samples, args.consumed_test_samples = 0, 0, 0
args.micro_batch_size = args.micro_batch_size * args.num_micro_batches
# In order to be compatible with data in transform format,
# it needs to increase the size of mbs first,
# and then split the large batch data into some mbs.
(
train_data_iterator,
valid_data_iterator,
test_data_iterator,
) = dataloader.build_train_valid_test_data_iterators(accelerator)
args.micro_batch_size = args.micro_batch_size // args.num_micro_batches
train_data_iterator = _handle_megatron_data_iterator(
accelerator=accelerator, data_iterator=train_data_iterator
)
valid_data_iterator = _handle_megatron_data_iterator(
accelerator=accelerator, data_iterator=valid_data_iterator
)
test_data_iterator = _handle_megatron_data_iterator(accelerator=accelerator, data_iterator=test_data_iterator)
return train_data_iterator, valid_data_iterator, test_data_iterator
# optimizer utilities
class MegatronLMOptimizerWrapper(AcceleratedOptimizer):
def __init__(self, optimizer):
super().__init__(optimizer, device_placement=False, scaler=None)
def zero_grad(self, set_to_none=None):
pass # `model(**batch)` is doing that automatically. Therefore, it's implementation is not needed
def step(self):
pass # `model(**batch)` is doing that automatically. Therefore, it's implementation is not needed
@property
def step_was_skipped(self):
"""Whether or not the optimizer step was done, or skipped because of gradient overflow."""
return self.optimizer.skipped_iter
def prepare_optimizer(accelerator, model):
accelerator.print("Preparing optimizer")
args = get_args()
return get_megatron_optimizer(model, args.no_wd_decay_cond, args.scale_lr_cond, args.lr_mult)
# scheduler utilities
class MegatronLMDummyScheduler:
"""
Dummy scheduler presents model parameters or param groups, this is primarily used to follow conventional training
loop when scheduler config is specified in the deepspeed config file.
Args:
optimizer (`torch.optim.optimizer.Optimizer`):
The optimizer to wrap.
total_num_steps (int):
Total number of steps.
warmup_num_steps (int):
Number of steps for warmup.
**kwargs (additional keyword arguments, *optional*):
Other arguments.
"""
def __init__(self, optimizer, total_num_steps=None, warmup_num_steps=0, **kwargs):
self.optimizer = optimizer
self.total_num_steps = total_num_steps
self.warmup_num_steps = warmup_num_steps
self.kwargs = kwargs
class MegatronLMSchedulerWrapper(AcceleratedScheduler):
def __init__(self, scheduler, optimizers):
super().__init__(scheduler, optimizers)
def step(self, *args, **kwargs):
return # `model(**batch)` is doing that automatically. Therefore, it's implementation is not needed
def prepare_scheduler(accelerator, optimizer, scheduler):
accelerator.print("Preparing scheduler")
scheduler = get_optimizer_param_scheduler(optimizer)
return scheduler
class AbstractTrainStep(ABC):
"""Abstract class for batching, forward pass and loss handler."""
def __init__(self, name):
super().__init__()
self.name = name
def get_batch_func(self, accelerator, megatron_dataset_flag):
pass
def get_forward_step_func(self):
pass
def get_loss_func(self, accelerator):
pass
class BertTrainStep(AbstractTrainStep):
"""
Bert train step class.
Args:
args (`argparse.Namespace`): Megatron-LM arguments.
"""
def __init__(self, accelerator, args):
super().__init__("BertTrainStep")
self.get_batch = self.get_batch_func(accelerator, args.megatron_dataset_flag)
self.loss_func = self.get_loss_func(accelerator, args.pretraining_flag, args.num_labels)
self.forward_step = self.get_forward_step_func(args.pretraining_flag, args.bert_binary_head)
if not args.model_return_dict:
self.model_output_class = None
else:
from transformers.modeling_outputs import SequenceClassifierOutput
self.model_output_class = SequenceClassifierOutput
def get_batch_func(self, accelerator, megatron_dataset_flag):
def get_batch_megatron(data_iterator):
"""Build the batch."""
# Items and their type.
keys = ["text", "types", "labels", "is_random", "loss_mask", "padding_mask"]
datatype = torch.int64
# Broadcast data.
if data_iterator is not None:
data = next(data_iterator)
else:
data = None
data_b = tensor_parallel.broadcast_data(keys, data, datatype)
# Unpack.
tokens = data_b["text"].long()
types = data_b["types"].long()
sentence_order = data_b["is_random"].long()
loss_mask = data_b["loss_mask"].float()
lm_labels = data_b["labels"].long()
padding_mask = data_b["padding_mask"].long()
return tokens, types, sentence_order, loss_mask, lm_labels, padding_mask
def get_batch_transformer(data_iterator):
"""Build the batch."""
data = next(data_iterator)
data = send_to_device(data, torch.cuda.current_device())
# Unpack.
tokens = data["input_ids"].long()
padding_mask = data["attention_mask"].long()
if "token_type_ids" in data:
types = data["token_type_ids"].long()
else:
types = None
if "labels" in data:
lm_labels = data["labels"].long()
loss_mask = (data["labels"] != -100).to(torch.float)
else:
lm_labels = None
loss_mask = None
if "next_sentence_label" in data:
sentence_order = data["next_sentence_label"].long()
else:
sentence_order = None
return tokens, types, sentence_order, loss_mask, lm_labels, padding_mask
if accelerator.state.megatron_lm_plugin.custom_get_batch_function is not None:
return accelerator.state.megatron_lm_plugin.custom_get_batch_function
if megatron_dataset_flag:
try:
# Use '--no-use-pep517 -e' to pip install nvidia's megatron from source
from pretrain_bert import get_batch
return get_batch
except ImportError:
pass
return get_batch_megatron
else:
return get_batch_transformer
def get_loss_func(self, accelerator, pretraining_flag, num_labels):
def loss_func_pretrain(loss_mask, sentence_order, output_tensor):
lm_loss_, sop_logits = output_tensor
lm_loss_ = lm_loss_.float()
loss_mask = loss_mask.float()
lm_loss = torch.sum(lm_loss_.view(-1) * loss_mask.reshape(-1)) / loss_mask.sum()
if sop_logits is not None:
sop_loss = F.cross_entropy(sop_logits.view(-1, 2).float(), sentence_order.view(-1), ignore_index=-1)
sop_loss = sop_loss.float()
loss = lm_loss + sop_loss
averaged_losses = average_losses_across_data_parallel_group([lm_loss, sop_loss])
return loss, {"lm loss": averaged_losses[0], "sop loss": averaged_losses[1]}
else:
loss = lm_loss
averaged_losses = average_losses_across_data_parallel_group([lm_loss])
return loss, {"lm loss": averaged_losses[0]}
def loss_func_finetune(labels, logits):
if num_labels == 1:
# We are doing regression
loss_fct = MSELoss()
loss = loss_fct(logits.view(-1), labels.view(-1))
elif self.num_labels > 1 and (labels.dtype in (torch.long, torch.int)):
loss_fct = CrossEntropyLoss()
loss = loss_fct(logits.view(-1, num_labels), labels.view(-1))
else:
loss_fct = BCEWithLogitsLoss()
loss = loss_fct(logits, labels)
averaged_losses = average_losses_across_data_parallel_group([loss])
return loss, {"loss": averaged_losses[0]}
if accelerator.state.megatron_lm_plugin.custom_loss_function is not None:
return accelerator.state.megatron_lm_plugin.custom_loss_function
if pretraining_flag:
return loss_func_pretrain
else:
return loss_func_finetune
def get_forward_step_func(self, pretraining_flag, bert_binary_head):
def forward_step(data_iterator, model):
"""Forward step."""
tokens, types, sentence_order, loss_mask, labels, padding_mask = self.get_batch(data_iterator)
if not bert_binary_head:
types = None
# Forward pass through the model.
if pretraining_flag:
output_tensor = model(tokens, padding_mask, tokentype_ids=types, lm_labels=labels)
return output_tensor, partial(self.loss_func, loss_mask, sentence_order)
else:
logits = model(tokens, padding_mask, tokentype_ids=types)
return logits, partial(self.loss_func, labels)
return forward_step
class GPTTrainStep(AbstractTrainStep):
"""
GPT train step class.
Args:
args (`argparse.Namespace`): Megatron-LM arguments.
"""
def __init__(self, accelerator, args):
super().__init__("GPTTrainStep")
self.get_batch = self.get_batch_func(accelerator, args.megatron_dataset_flag)
self.loss_func = self.get_loss_func(accelerator)
self.forward_step = self.get_forward_step_func()
self.eod_token = args.padded_vocab_size - 1
if args.vocab_file is not None:
tokenizer = get_tokenizer()
self.eod_token = tokenizer.eod
self.reset_position_ids = args.reset_position_ids
self.reset_attention_mask = args.reset_attention_mask
self.eod_mask_loss = args.eod_mask_loss
if not args.model_return_dict:
self.model_output_class = None
else:
from transformers.modeling_outputs import CausalLMOutputWithCrossAttentions
self.model_output_class = CausalLMOutputWithCrossAttentions
def get_batch_func(self, accelerator, megatron_dataset_flag):
def get_batch_megatron(data_iterator):
"""Generate a batch"""
# Items and their type.
keys = ["text"]
datatype = torch.int64
# Broadcast data.
if data_iterator is not None:
data = next(data_iterator)
else:
data = None
data_b = tensor_parallel.broadcast_data(keys, data, datatype)
# Unpack.
tokens_ = data_b["text"].long()
labels = tokens_[:, 1:].contiguous()
tokens = tokens_[:, :-1].contiguous()
# Get the masks and postition ids.
attention_mask, loss_mask, position_ids = get_ltor_masks_and_position_ids(
tokens, self.eod_token, self.reset_position_ids, self.reset_attention_mask, self.eod_mask_loss
)
return tokens, labels, loss_mask, attention_mask, position_ids
def get_batch_transformer(data_iterator):
data = next(data_iterator)
data = {"input_ids": data["input_ids"]}
data = send_to_device(data, torch.cuda.current_device())
tokens_ = data["input_ids"].long()
padding = torch.zeros((tokens_.shape[0], 1), dtype=tokens_.dtype, device=tokens_.device) + self.eod_token
tokens_ = torch.concat([tokens_, padding], dim=1)
labels = tokens_[:, 1:].contiguous()
tokens = tokens_[:, :-1].contiguous()
# Get the masks and postition ids.
attention_mask, loss_mask, position_ids = get_ltor_masks_and_position_ids(
tokens, self.eod_token, self.reset_position_ids, self.reset_attention_mask, True
)
return tokens, labels, loss_mask, attention_mask, position_ids
if accelerator.state.megatron_lm_plugin.custom_get_batch_function is not None:
return accelerator.state.megatron_lm_plugin.custom_get_batch_function
if megatron_dataset_flag:
try:
# Use '--no-use-pep517 -e' to pip install nvidia's megatron from source
from pretrain_gpt import get_batch
return get_batch
except ImportError:
pass
return get_batch_megatron
else:
return get_batch_transformer
def get_loss_func(self, accelerator):
args = get_args()
def loss_func(loss_mask, output_tensor):
if args.return_logits:
losses, logits = output_tensor
else:
losses = output_tensor
losses = losses.float()
loss_mask = loss_mask.view(-1).float()
if args.context_parallel_size > 1:
loss = torch.cat([torch.sum(losses.view(-1) * loss_mask).view(1), loss_mask.sum().view(1)])
torch.distributed.all_reduce(loss, group=mpu.get_context_parallel_group())
loss = loss[0] / loss[1]
else:
loss = torch.sum(losses.view(-1) * loss_mask) / loss_mask.sum()
# Check individual rank losses are not NaN prior to DP all-reduce.
if args.check_for_nan_in_loss_and_grad:
global_rank = torch.distributed.get_rank()
assert not loss.isnan(), (
f"Rank {global_rank}: found NaN in local forward loss calculation. "
f"Device: {torch.cuda.current_device()}, node: {os.uname()[1]}"
)
# Reduce loss for logging.
averaged_loss = average_losses_across_data_parallel_group([loss])
output_dict = {"lm loss": averaged_loss[0]}
if args.return_logits:
output_dict.update({"logits": logits})
return loss, output_dict
if accelerator.state.megatron_lm_plugin.custom_loss_function is not None:
return accelerator.state.megatron_lm_plugin.custom_loss_function
return loss_func
def get_forward_step_func(self):
def forward_step(data_iterator, model):
"""Forward step."""
# Get the batch.
tokens, labels, loss_mask, attention_mask, position_ids = self.get_batch(data_iterator)
output_tensor = model(tokens, position_ids, attention_mask, labels=labels)
return output_tensor, partial(self.loss_func, loss_mask)
return forward_step
class T5TrainStep(AbstractTrainStep):
"""
T5 train step class.
Args:
args (`argparse.Namespace`): Megatron-LM arguments.
"""
def __init__(self, accelerator, args):
super().__init__("T5TrainStep")
self.get_batch = self.get_batch_func(accelerator, args.megatron_dataset_flag)
self.loss_func = self.get_loss_func(accelerator)
self.forward_step = self.get_forward_step_func()
if not args.model_return_dict:
self.model_output_class = None
else:
from transformers.modeling_outputs import Seq2SeqLMOutput
self.model_output_class = Seq2SeqLMOutput
@staticmethod
def attn_mask_postprocess(attention_mask):
# We create a 3D attention mask from a 2D tensor mask.
# [b, 1, s]
attention_mask_b1s = attention_mask.unsqueeze(1)
# [b, s, 1]
attention_mask_bs1 = attention_mask.unsqueeze(2)
# [b, s, s]
attention_mask_bss = attention_mask_b1s * attention_mask_bs1
# Convert attention mask to binary:
extended_attention_mask = attention_mask_bss < 0.5
return extended_attention_mask
@staticmethod
def get_decoder_mask(seq_length, device):
attention_mask = torch.tril(torch.ones((1, seq_length, seq_length), device=device))
attention_mask = attention_mask < 0.5
return attention_mask
@staticmethod
def get_enc_dec_mask(attention_mask, dec_seq_length, device):
batch_size, _ = attention_mask.shape
# We create a 3D attention mask from a 2D tensor mask.
# [b, 1, s]
attention_mask_b1s = attention_mask.unsqueeze(1)
# [b, s, 1]
attention_mask_bs1 = torch.ones((batch_size, dec_seq_length, 1), device=device)
attention_mask_bss = attention_mask_bs1 * attention_mask_b1s
extended_attention_mask = attention_mask_bss < 0.5
return extended_attention_mask
def get_batch_func(self, accelerator, megatron_dataset_flag):
def get_batch_megatron(data_iterator):
"""Build the batch."""
keys = ["text_enc", "text_dec", "labels", "loss_mask", "enc_mask", "dec_mask", "enc_dec_mask"]
datatype = torch.int64
# Broadcast data.
if data_iterator is not None:
data = next(data_iterator)
else:
data = None
data_b = tensor_parallel.broadcast_data(keys, data, datatype)
# Unpack.
tokens_enc = data_b["text_enc"].long()
tokens_dec = data_b["text_dec"].long()
labels = data_b["labels"].long()
loss_mask = data_b["loss_mask"].float()
enc_mask = data_b["enc_mask"] < 0.5
dec_mask = data_b["dec_mask"] < 0.5
enc_dec_mask = data_b["enc_dec_mask"] < 0.5
return tokens_enc, tokens_dec, loss_mask, labels, enc_mask, dec_mask, enc_dec_mask
def get_batch_transformer(data_iterator):
"""Build the batch."""
data = next(data_iterator)
data = send_to_device(data, torch.cuda.current_device())
tokens_enc = data["input_ids"].long()
labels = data["labels"].long()
loss_mask = (labels != -100).to(torch.float)
if "decoder_input_ids" in data:
tokens_dec = data["decoder_input_ids"].long()
else:
tokens_dec = labels.new_zeros(labels.shape, device=labels.device, dtype=torch.long)
tokens_dec[..., 1:] = labels[..., :-1].clone()
tokens_dec[..., 0] = 0
tokens_dec.masked_fill_(tokens_dec == -100, 0)
enc_mask = T5TrainStep.attn_mask_postprocess(data["attention_mask"].long())
dec_mask = T5TrainStep.get_decoder_mask(tokens_dec.shape[1], tokens_dec.device)
enc_dec_mask = T5TrainStep.get_enc_dec_mask(
data["attention_mask"].long(), tokens_dec.shape[1], tokens_dec.device
)
return tokens_enc, tokens_dec, loss_mask, labels, enc_mask, dec_mask, enc_dec_mask
if accelerator.state.megatron_lm_plugin.custom_get_batch_function is not None:
return accelerator.state.megatron_lm_plugin.custom_get_batch_function
if megatron_dataset_flag:
try:
# Use '--no-use-pep517 -e' to pip install nvidia's megatron from source
from pretrain_t5 import get_batch
return get_batch
except ImportError:
pass
return get_batch_megatron
else:
return get_batch_transformer
def get_loss_func(self, accelerator):
def loss_func(loss_mask, output_tensor):
lm_loss_ = output_tensor.float()
lm_loss = torch.sum(lm_loss_.view(-1) * loss_mask.reshape(-1)) / loss_mask.sum()
loss = lm_loss
averaged_losses = average_losses_across_data_parallel_group([lm_loss])
return loss, {"lm loss": averaged_losses[0]}
if accelerator.state.megatron_lm_plugin.custom_loss_function is not None:
return accelerator.state.megatron_lm_plugin.custom_loss_function
return loss_func
def get_forward_step_func(self):
def forward_step(data_iterator, model):
"""Forward step."""
# Get the batch.
tokens_enc, tokens_dec, loss_mask, lm_labels, enc_mask, dec_mask, enc_dec_mask = self.get_batch(
data_iterator
)
# Forward model lm_labels
output_tensor = model(
tokens_enc, tokens_dec, enc_mask, dec_mask, enc_dec_mask, tokentype_ids=None, lm_labels=lm_labels
)
return output_tensor, partial(self.loss_func, loss_mask)
return forward_step
def finish_mpu_init():
# torch.distributed initialization
args = get_args()
# Pytorch distributed.
_initialize_distributed()
# Random seeds for reproducibility.
if args.rank == 0:
print(f"> setting random seeds to {args.seed} ...")
_set_random_seed(args.seed, args.data_parallel_random_init)
# intialize megatron setup
def initialize(accelerator, extra_args_provider=None, args_defaults={}):
accelerator.print("Initializing Megatron-LM")
assert torch.cuda.is_available(), "Megatron requires CUDA."
# Parse arguments
args = parse_args(extra_args_provider, ignore_unknown_args=True)
# Set defaults
for key, value in args_defaults.items():
if getattr(args, key, None) is not None:
if args.rank == 0:
print(
f"WARNING: overriding default arguments for " f"{key}:{getattr(args, key)} with {key}:{value}",
flush=True,
)
setattr(args, key, value)
if args.use_checkpoint_args or args_defaults.get("use_checkpoint_args", False):
assert args.load is not None, "--use-checkpoints-args requires --load argument"
load_args_from_checkpoint(args)
validate_args(args)
# set global args, build tokenizer, and set adlr-autoresume,
# tensorboard-writer, and timers.
set_global_variables(args)
# Megatron's MPU is the master. Complete initialization right away.
finish_mpu_init()
# Autoresume.
_init_autoresume()
# Compile dependencies.
_compile_dependencies()
# Set pytorch JIT layer fusion options and warmup JIT functions.
set_jit_fusion_options()
args = get_args()
if getattr(args, "padded_vocab_size", None) is None:
args.padded_vocab_size = _vocab_size_with_padding(args.orig_vocab_size, args)
if args.model_type_name == "bert" and args.pretraining_flag and args.num_labels == 2:
args.bert_binary_head = True
else:
args.bert_binary_head = False
args.iteration = 0
class MegatronEngine(torch.nn.Module):
"""
Megatron-LM model wrapper
Args:
accelerator (:class:`~accelerate.Accelerator`): The accelerator object to use.
model: Megatron-LM model
optimizer: Megatron-LM optimizer
lr_scheduler: Megatron-LM lr scheduler
"""
def __init__(self, accelerator, model, optimizer, scheduler):
super().__init__()
self.module = model
self.base_model = model[0]
self.optimizer = optimizer
self.scheduler = scheduler
args = get_args()
if accelerator.state.megatron_lm_plugin.custom_train_step_class is not None:
self.train_step_handler = accelerator.state.megatron_lm_plugin.custom_train_step_class(
args, **accelerator.state.megatron_lm_plugin.custom_train_step_kwargs
)
elif args.model_type_name == "bert":
self.train_step_handler = BertTrainStep(accelerator, args)
elif args.model_type_name == "gpt":
self.train_step_handler = GPTTrainStep(accelerator, args)
elif args.model_type_name == "t5":
self.train_step_handler = T5TrainStep(accelerator, args)
else:
raise ValueError(f"Unsupported model type: {args.model_type_name}")
self.optimizer.skipped_iter = False
# Tracking loss.
self.total_loss_dict = {}
self.eval_total_loss_dict = {}
self.iteration = 0
self.report_memory_flag = True
self.num_floating_point_operations_so_far = 0
self.module_config = None
if args.tensorboard_dir is not None:
write_args_to_tensorboard()
def get_module_config(self):
args = get_args()
config = get_model_config(self.module[0])
# Setup some training config params
config.grad_scale_func = self.optimizer.scale_loss
if isinstance(self.module[0], LocalDDP) and args.overlap_grad_reduce:
assert config.no_sync_func is None, (
"When overlap_grad_reduce is True, config.no_sync_func must be None; "
"a custom no_sync_func is not supported when overlapping grad-reduce"
)
config.no_sync_func = [model_chunk.no_sync for model_chunk in self.module]
if len(self.module) == 1:
config.no_sync_func = config.no_sync_func[0]
if args.delay_grad_reduce:
config.grad_sync_func = [model_chunk.start_grad_sync for model_chunk in self.module]
if len(self.module) == 1:
config.grad_sync_func = config.grad_sync_func[0]
if args.overlap_param_gather and args.delay_param_gather:
config.param_sync_func = [
lambda x: self.optimizer.finish_param_sync(model_index, x) for model_index in range(len(self.module))
]
if len(self.module) == 1:
config.param_sync_func = config.param_sync_func[0]
config.finalize_model_grads_func = finalize_model_grads
return config
def train(self):
for model_module in self.module:
model_module.train()
if self.module_config is None:
self.module_config = self.get_module_config()
self.log_eval_results()
def eval(self):
for model_module in self.module:
model_module.eval()
if self.module_config is None:
self.module_config = self.get_module_config()
def get_batch_data_iterator(self, batch_data):
args = get_args()
data_chunks = []
if len(batch_data) > 0:
if args.num_micro_batches > 1:
for i in range(0, args.num_micro_batches):
data_chunks.append(
{
k: v[i * args.micro_batch_size : (i + 1) * args.micro_batch_size]
for k, v in batch_data.items()
}
)
else:
data_chunks = [batch_data]
if len(self.module) > 1:
batch_data_iterator = (
[iter(data_chunks) for _ in range(len(self.module))]
if len(batch_data) > 0
else [None] * len(self.module)
)
else:
batch_data_iterator = iter(data_chunks) if len(batch_data) > 0 else None
return batch_data_iterator
def train_step(self, **batch_data):
"""
Training step for Megatron-LM
Args:
batch_data (:obj:`dict`): The batch data to train on.
"""
batch_data_iterator = self.get_batch_data_iterator(batch_data)
loss_reduced, skipped_iter, grad_norm, num_zeros_in_grad = train_step(
forward_step_func=self.train_step_handler.forward_step,
data_iterator=batch_data_iterator,
model=self.module,
optimizer=self.optimizer,
opt_param_scheduler=self.scheduler,
config=self.module_config,
)
self.optimizer.skipped_iter = skipped_iter == 1
return loss_reduced, skipped_iter, grad_norm, num_zeros_in_grad
def eval_step(self, **batch_data):
"""
Evaluation step for Megatron-LM
Args:
batch_data (:obj:`dict`): The batch data to evaluate on.
"""
args = get_args()
batch_data_iterator = self.get_batch_data_iterator(batch_data)
forward_backward_func = get_forward_backward_func()
loss_dicts = forward_backward_func(
forward_step_func=self.train_step_handler.forward_step,
data_iterator=batch_data_iterator,
model=self.module,
num_microbatches=get_num_microbatches(),
seq_length=args.seq_length,
micro_batch_size=args.micro_batch_size,
forward_only=True,
)
# Empty unused memory
if args.empty_unused_memory_level >= 1:
torch.cuda.empty_cache()
args.consumed_valid_samples += (
mpu.get_data_parallel_world_size() * args.micro_batch_size * get_num_microbatches()
)
if mpu.is_pipeline_last_stage(ignore_virtual=True):
# Average loss across microbatches.
loss_reduced = {}
for key in loss_dicts[0]:
losses_reduced_for_key = [x[key] for x in loss_dicts]
if len(losses_reduced_for_key[0].shape) == 0:
loss_reduced[key] = sum(losses_reduced_for_key) / len(losses_reduced_for_key)
else:
loss_reduced[key] = torch.concat(losses_reduced_for_key)
return loss_reduced
return {}
def forward(self, **batch_data):
# During training, we use train_step()
# model(**batch_data) performs following operations by delegating it to `self.train_step`:
# 1. Prepare **batch_data for Tendor, Pipeline and Model Parallelism
# 2. Set grad to zero.
# 3. forward pass and backward pass using Pipeline Parallelism
# 4. Empty unused memory.
# 5. Reduce gradients.
# 6. Update parameters.
# 7. Gather params when using Distributed Optimizer (Data Parallelism).
# 8. Update learning rate if scheduler is specified.
# 9. Empty unused memory.
# 10. Average loss across microbatches and across DP ranks.
#
# During evaluation, we use eval_step()
args = get_args()
if self.module[0].training:
loss_dict, skipped_iter, grad_norm, num_zeros_in_grad = self.train_step(**batch_data)
self.iteration += 1
batch_size = mpu.get_data_parallel_world_size() * args.micro_batch_size * get_num_microbatches()
args.consumed_train_samples += batch_size
self.num_floating_point_operations_so_far += num_floating_point_operations(args, batch_size)
if args.tensorboard_dir is not None:
# Logging.
loss_scale = self.optimizer.get_loss_scale().item()
params_norm = None
if args.log_params_norm:
params_norm = calc_params_l2_norm(self.model)
self.report_memory_flag = training_log(
loss_dict,
self.total_loss_dict,
self.optimizer.param_groups[0]["lr"],
self.iteration,
loss_scale,
self.report_memory_flag,
skipped_iter,
grad_norm,
params_norm,
num_zeros_in_grad,
)
else:
loss_dict = self.eval_step(**batch_data)
if args.tensorboard_dir is not None:
for key in loss_dict:
self.eval_total_loss_dict[key] = (
self.eval_total_loss_dict.get(key, torch.cuda.FloatTensor([0.0])) + loss_dict[key]
)
self.eval_total_loss_dict[key + "_num_iters"] = self.eval_total_loss_dict.get(
key + "_num_iters", torch.cuda.FloatTensor([0.0])
) + torch.cuda.FloatTensor([1.0])
loss = torch.tensor(0.0, device=torch.cuda.current_device())
for key in loss_dict:
if len(loss_dict[key].shape) == 0:
loss += loss_dict[key]
logits = None
if "logits" in loss_dict:
logits = loss_dict["logits"]
if self.train_step_handler.model_output_class is not None:
return self.train_step_handler.model_output_class(loss=loss, logits=logits)
return loss
def log_eval_results(self):
args = get_args()
if args.tensorboard_dir is None or self.iteration == 0:
return
args = get_args()
writer = get_tensorboard_writer()
string = f"validation loss at iteration {self.iteration} | "
for key in self.eval_total_loss_dict:
if key.endswith("_num_iters"):
continue
value = self.eval_total_loss_dict[key] / self.eval_total_loss_dict[key + "_num_iters"]
string += f"{key} value: {value} | "
ppl = math.exp(min(20, value.item()))
if args.pretraining_flag:
string += f"{key} PPL: {ppl} | "
if writer:
writer.add_scalar(f"{key} validation", value.item(), self.iteration)
if args.pretraining_flag:
writer.add_scalar(f"{key} validation ppl", ppl, self.iteration)
length = len(string) + 1
print_rank_last("-" * length)
print_rank_last(string)
print_rank_last("-" * length)
self.eval_total_loss_dict = {}
def save_checkpoint(self, output_dir):
self.log_eval_results()
args = get_args()
args.save = output_dir
torch.distributed.barrier()
save_checkpoint(
self.iteration,
self.module,
self.optimizer,
self.scheduler,
num_floating_point_operations_so_far=self.num_floating_point_operations_so_far,
)
torch.distributed.barrier()
def load_checkpoint(self, input_dir):
args = get_args()
args.load = input_dir
args.consumed_train_samples = 0
args.consumed_valid_samples = 0
torch.distributed.barrier()
iteration, num_floating_point_operations_so_far = load_checkpoint(self.module, self.optimizer, self.scheduler)
torch.distributed.barrier()
self.iteration = iteration
self.num_floating_point_operations_so_far = num_floating_point_operations_so_far
if args.fp16 and self.iteration == 0:
self.optimizer.reload_model_params()
def megatron_generate(
self,
inputs,
attention_mask=None,
max_length=None,
max_new_tokens=None,
num_beams=None,
temperature=None,
top_k=None,
top_p=None,
length_penalty=None,
**kwargs,
):
"""
Generate method for GPT2 model. This method is used for inference. Supports both greedy and beam search along
with sampling. Refer the Megatron-LM repo for more details
Args:
inputs (torch.Tensor): input ids
attention_mask (torch.Tensor, optional): attention mask. Defaults to None.
max_length (int, optional): max length of the generated sequence. Defaults to None.
Either this or max_new_tokens should be provided.
max_new_tokens (int, optional): max number of tokens to be generated. Defaults to None.
Either this or max_length should be provided.
num_beams (int, optional): number of beams to use for beam search. Defaults to None.
temperature (float, optional): temperature for sampling. Defaults to 1.0.
top_k (int, optional): top k tokens to consider for sampling. Defaults to 0.0.
top_p (float, optional): tokens in top p probability are considered for sampling. Defaults to 0.0.
length_penalty (float, optional): length penalty for beam search. Defaults to None.
kwargs: additional key-value arguments
"""
# checking if required arguments are passed
args = get_args()
if args.model_type_name != "gpt":
raise NotImplementedError("Generate method is not implemented for this model")
if args.data_parallel_size > 1:
raise ValueError("Generate method requires data parallelism to be 1")
if args.sequence_parallel:
raise ValueError("Generate method requires sequence parallelism to be False")
if args.recompute_granularity is not None:
raise ValueError("Checkpoint activations cannot be set for inference")
if args.vocab_file is None:
raise ValueError("Vocab file is required for inference")
# Prepare inputs
if max_length is None and max_new_tokens is None:
raise ValueError("`max_length` or `max_new_tokens` are required for inference")
if temperature is None:
temperature = 1.0
elif not (0.0 < temperature <= 100.0):
raise ValueError("temperature must be a positive number less than or equal to 100.0")
if top_k is None:
top_k = 0
elif not (0 <= top_k <= 1000):
raise ValueError("top_k must be a positive number less than or equal to 1000")
if top_p is None:
top_p = 0.0
elif top_p > 0.0 and top_k > 0.0:
raise ValueError("top_p and top_k sampling cannot be set together")
else:
if not (0.0 <= top_p <= 1.0):
raise ValueError("top_p must be less than or equal to 1.0")
top_p_decay = kwargs.get("top_p_decay", 0.0)
if not (0.0 <= top_p_decay <= 1.0):
raise ValueError("top_p_decay must be less than or equal to 1.0")
top_p_bound = kwargs.get("top_p_bound", 0.0)
if not (0.0 <= top_p_bound <= 1.0):
raise ValueError("top_p_bound must be less than or equal to 1.0")
add_BOS = kwargs.get("add_BOS", False)
if not (isinstance(add_BOS, bool)):
raise ValueError("add_BOS must be a boolean")
beam_width = num_beams
if beam_width is not None:
if not isinstance(beam_width, int):
raise ValueError("beam_width must be an integer")
if beam_width < 1:
raise ValueError("beam_width must be greater than 0")
if inputs.shape[0] > 1:
return "When doing beam_search, batch size must be 1"
tokenizer = get_tokenizer()
stop_token = kwargs.get("stop_token", tokenizer.eod)
if stop_token is not None:
if not isinstance(stop_token, int):
raise ValueError("stop_token must be an integer")
if length_penalty is None:
length_penalty = 1.0
sizes_list = None
prompts_tokens_tensor = None
prompts_length_tensor = None
if torch.distributed.get_rank() == 0:
# Get the prompts length.
if attention_mask is None:
prompts_length_tensor = torch.cuda.LongTensor([inputs.shape[1]] * inputs.shape[0])
else:
prompts_length_tensor = attention_mask.sum(axis=-1).cuda()
if max_new_tokens is None:
max_new_tokens = max_length - inputs.shape[1]
if max_new_tokens <= 0:
raise ValueError("max_new_tokens must be greater than 0")
if add_BOS:
max_length = max_new_tokens + inputs.shape[1] + 1
# making sure that `max_length` is a multiple of 4 to leverage fused kernels
max_length = 4 * math.ceil(max_length / 4)
max_new_tokens = max_length - (inputs.shape[1] + 1)
padding = torch.cuda.LongTensor([[tokenizer.eod] * max_new_tokens] * inputs.shape[0])
prompts_tokens_tensor = torch.concat(
[torch.unsqueeze(padding[:, 0], axis=-1), inputs.cuda(), padding], axis=-1
)
else:
# making sure that `max_length` is a multiple of 4 to leverage fused kernels
max_length = max_new_tokens + inputs.shape[1]
max_length = 4 * math.ceil(max_length / 4)
max_new_tokens = max_length - inputs.shape[1]
padding = torch.cuda.LongTensor([[tokenizer.eod] * max_new_tokens] * inputs.shape[0])
prompts_tokens_tensor = torch.concat([inputs.cuda(), padding], axis=-1)
# We need the sizes of these tensors for the boradcast
sizes_list = [
prompts_tokens_tensor.size(0), # Batch size
prompts_tokens_tensor.size(1),
] # Sequence lenght
# First, broadcast the sizes.
sizes_tensor = broadcast_int_list(2, int_list=sizes_list, rank=0)
# Now that we have the sizes, we can boradcast the tokens
# and length tensors.
sizes = sizes_tensor.tolist()
context_tokens_tensor = broadcast_tensor(sizes, torch.int64, tensor=prompts_tokens_tensor, rank=0)
context_length_tensor = broadcast_tensor(sizes[0], torch.int64, tensor=prompts_length_tensor, rank=0)
# Run the inference
random_seed = kwargs.get("random_seed", 0)
torch.random.manual_seed(random_seed)
unwrapped_model = unwrap_model(self.base_model, (torchDDP, LocalDDP, Float16Module))
if beam_width is not None:
tokens, _ = beam_search_and_return_on_first_stage(
unwrapped_model,
context_tokens_tensor,
context_length_tensor,
beam_width,
stop_token=stop_token,
num_return_gen=1,
length_penalty=length_penalty,
)
else:
tokens, _, _ = generate_tokens_probs_and_return_on_first_stage(
unwrapped_model,
context_tokens_tensor,
context_length_tensor,
return_output_log_probs=False,
top_k=top_k,
top_p=top_p,
top_p_decay=top_p_decay,
top_p_bound=top_p_bound,
temperature=temperature,
use_eod_token_for_early_termination=True,
)
return tokens
# other utilities
def avg_losses_across_data_parallel_group(losses):
"""
Average losses across data parallel group.
Args:
losses (List[Tensor]): List of losses to average across data parallel group.
"""
return average_losses_across_data_parallel_group(losses)
def gather_across_data_parallel_groups(tensor):
"""
Recursively gather tensor in a nested list/tuple/dictionary of tensors from data parallel ranks.
Args:
tensor (nested list/tuple/dictionary of `torch.Tensor`):
The data to gather across data parallel ranks.
"""
def _gpu_gather_one(tensor):
if tensor.ndim == 0:
tensor = tensor.clone()[None]
output_tensors = [
torch.empty_like(tensor)
for _ in range(torch.distributed.get_world_size(group=mpu.get_data_parallel_group()))
]
torch.distributed.all_gather(output_tensors, tensor, group=mpu.get_data_parallel_group())
return torch.cat(output_tensors, dim=0)
return recursively_apply(_gpu_gather_one, tensor, error_on_other_type=True)
| accelerate/src/accelerate/utils/megatron_lm.py/0 | {
"file_path": "accelerate/src/accelerate/utils/megatron_lm.py",
"repo_id": "accelerate",
"token_count": 26942
} |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import inspect
import itertools
import json
import os
import tempfile
from copy import deepcopy
from pathlib import Path
import torch
from parameterized import parameterized
from torch.utils.data import BatchSampler, DataLoader, RandomSampler, SequentialSampler
from transformers import AutoConfig, AutoModel, AutoModelForCausalLM, get_scheduler
from transformers.testing_utils import mockenv_context
from transformers.trainer_utils import set_seed
from transformers.utils import is_torch_bf16_available
from accelerate.accelerator import Accelerator
from accelerate.scheduler import AcceleratedScheduler
from accelerate.state import AcceleratorState
from accelerate.test_utils.testing import (
AccelerateTestCase,
TempDirTestCase,
execute_subprocess_async,
path_in_accelerate_package,
require_deepspeed,
require_huggingface_suite,
require_multi_device,
require_non_cpu,
slow,
)
from accelerate.test_utils.training import RegressionDataset, RegressionModel
from accelerate.utils import patch_environment
from accelerate.utils.dataclasses import DeepSpeedPlugin
from accelerate.utils.deepspeed import (
DeepSpeedEngineWrapper,
DeepSpeedOptimizerWrapper,
DeepSpeedSchedulerWrapper,
DummyOptim,
DummyScheduler,
)
from accelerate.utils.versions import compare_versions
set_seed(42)
GPT2_TINY = "sshleifer/tiny-gpt2"
MOBILEVIT = "apple/mobilevit-xx-small"
QWEN_MOE = "peft-internal-testing/tiny-random-qwen-1.5-MoE"
ZERO2 = "zero2"
ZERO3 = "zero3"
FP16 = "fp16"
BF16 = "bf16"
CUSTOM_OPTIMIZER = "custom_optimizer"
CUSTOM_SCHEDULER = "custom_scheduler"
DS_OPTIMIZER = "deepspeed_optimizer"
DS_SCHEDULER = "deepspeed_scheduler"
NO_CONFIG = "no_config"
CONFIG_WITH_NO_HIDDEN_SIZE = "config_with_no_hidden_size"
CONFIG_WITH_HIDDEN_SIZE = "config_with_hidden_size"
CONFIG_WITH_HIDDEN_SIZES = "config_with_hidden_sizes"
stages = [ZERO2, ZERO3]
optims = [CUSTOM_OPTIMIZER, DS_OPTIMIZER]
schedulers = [CUSTOM_SCHEDULER, DS_SCHEDULER]
model_types = [NO_CONFIG, CONFIG_WITH_NO_HIDDEN_SIZE, CONFIG_WITH_HIDDEN_SIZE, CONFIG_WITH_HIDDEN_SIZES]
if is_torch_bf16_available():
dtypes = [FP16, BF16]
else:
dtypes = [FP16]
def parameterized_custom_name_func(func, param_num, param):
# customize the test name generator function as we want both params to appear in the sub-test
# name, as by default it shows only the first param
param_based_name = parameterized.to_safe_name("_".join(str(x) for x in param.args))
return f"{func.__name__}_{param_based_name}"
# Cartesian-product of zero stages with models to test
params = list(itertools.product(stages, dtypes))
optim_scheduler_params = list(itertools.product(optims, schedulers))
class DummyConfig:
def __init__(self):
self._name_or_path = "dummy"
@require_deepspeed
@require_non_cpu
class DeepSpeedConfigIntegration(AccelerateTestCase):
def setUp(self):
super().setUp()
self._test_file_path = inspect.getfile(self.__class__)
path = Path(self._test_file_path).resolve()
self.test_file_dir_str = str(path.parents[0])
self.ds_config_file = dict(
zero2=f"{self.test_file_dir_str}/ds_config_zero2.json",
zero3=f"{self.test_file_dir_str}/ds_config_zero3.json",
)
# use self.get_config_dict(stage) to use these to ensure the original is not modified
with open(self.ds_config_file[ZERO2], encoding="utf-8") as f:
config_zero2 = json.load(f)
with open(self.ds_config_file[ZERO3], encoding="utf-8") as f:
config_zero3 = json.load(f)
# The following setting slows things down, so don't enable it by default unless needed by a test.
# It's in the file as a demo for users since we want everything to work out of the box even if slower.
config_zero3["zero_optimization"]["stage3_gather_16bit_weights_on_model_save"] = False
self.ds_config_dict = dict(zero2=config_zero2, zero3=config_zero3)
self.dist_env = dict(
ACCELERATE_USE_DEEPSPEED="true",
MASTER_ADDR="localhost",
MASTER_PORT="10999",
RANK="0",
LOCAL_RANK="0",
WORLD_SIZE="1",
)
def get_config_dict(self, stage):
# As some tests modify the dict, always make a copy
return deepcopy(self.ds_config_dict[stage])
@parameterized.expand(stages, name_func=parameterized_custom_name_func)
def test_deepspeed_plugin(self, stage):
# Test zero3_init_flag will be set to False when ZeRO stage != 3
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
zero3_init_flag=True,
)
assert not deepspeed_plugin.zero3_init_flag
deepspeed_plugin.deepspeed_config = None
# Test zero3_init_flag will be set to True only when ZeRO stage == 3
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=3,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
zero3_init_flag=True,
)
assert deepspeed_plugin.zero3_init_flag
deepspeed_plugin.deepspeed_config = None
# Test config files are loaded correctly
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=self.ds_config_file[stage], zero3_init_flag=True)
if stage == ZERO2:
assert not deepspeed_plugin.zero3_init_flag
elif stage == ZERO3:
assert deepspeed_plugin.zero3_init_flag
# Test `gradient_accumulation_steps` is set to 1 if unavailable in config file
with tempfile.TemporaryDirectory() as dirpath:
ds_config = self.get_config_dict(stage)
del ds_config["gradient_accumulation_steps"]
with open(os.path.join(dirpath, "ds_config.json"), "w") as out_file:
json.dump(ds_config, out_file)
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=os.path.join(dirpath, "ds_config.json"))
assert deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"] == 1
deepspeed_plugin.deepspeed_config = None
# Test `ValueError` is raised if `zero_optimization` is unavailable in config file
with tempfile.TemporaryDirectory() as dirpath:
ds_config = self.get_config_dict(stage)
del ds_config["zero_optimization"]
with open(os.path.join(dirpath, "ds_config.json"), "w") as out_file:
json.dump(ds_config, out_file)
with self.assertRaises(ValueError) as cm:
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=os.path.join(dirpath, "ds_config.json"))
assert "Please specify the ZeRO optimization config in the DeepSpeed config." in str(cm.exception)
deepspeed_plugin.deepspeed_config = None
# Test `deepspeed_config_process`
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=self.ds_config_file[stage])
kwargs = {
"fp16.enabled": True,
"bf16.enabled": False,
"optimizer.params.lr": 5e-5,
"optimizer.params.weight_decay": 0.0,
"scheduler.params.warmup_min_lr": 0.0,
"scheduler.params.warmup_max_lr": 5e-5,
"scheduler.params.warmup_num_steps": 0,
"train_micro_batch_size_per_gpu": 16,
"gradient_clipping": 1.0,
"train_batch_size": 16,
"zero_optimization.reduce_bucket_size": 5e5,
"zero_optimization.stage3_prefetch_bucket_size": 5e5,
"zero_optimization.stage3_param_persistence_threshold": 5e5,
"zero_optimization.stage3_gather_16bit_weights_on_model_save": False,
}
deepspeed_plugin.deepspeed_config_process(**kwargs)
for ds_key_long, value in kwargs.items():
config, ds_key = deepspeed_plugin.hf_ds_config.find_config_node(ds_key_long)
if config.get(ds_key) is not None:
assert config.get(ds_key) == value
# Test mismatches
mismatches = {
"optimizer.params.lr": 1e-5,
"optimizer.params.weight_decay": 1e-5,
"gradient_accumulation_steps": 2,
}
with self.assertRaises(ValueError) as cm:
new_kwargs = deepcopy(kwargs)
new_kwargs.update(mismatches)
deepspeed_plugin.deepspeed_config_process(**new_kwargs)
for key in mismatches.keys():
assert key in str(cm.exception), f"{key} is not in the exception message: {cm.exception}"
# Test `ValueError` is raised if some config file fields with `auto` value is missing in `kwargs`
deepspeed_plugin.deepspeed_config["optimizer"]["params"]["lr"] = "auto"
with self.assertRaises(ValueError) as cm:
del kwargs["optimizer.params.lr"]
deepspeed_plugin.deepspeed_config_process(**kwargs)
assert "`optimizer.params.lr` not found in kwargs." in str(cm.exception)
@parameterized.expand([FP16, BF16], name_func=parameterized_custom_name_func)
def test_accelerate_state_deepspeed(self, dtype):
AcceleratorState._reset_state(True)
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=ZERO2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
zero3_init_flag=True,
)
with mockenv_context(**self.dist_env):
state = Accelerator(mixed_precision=dtype, deepspeed_plugin=deepspeed_plugin).state
assert state.deepspeed_plugin.deepspeed_config[dtype]["enabled"]
def test_init_zero3(self):
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=3,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
zero3_init_flag=True,
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin) # noqa: F841
from transformers.integrations import is_deepspeed_zero3_enabled
assert is_deepspeed_zero3_enabled()
@parameterized.expand(optim_scheduler_params, name_func=parameterized_custom_name_func)
def test_prepare_deepspeed(self, optim_type, scheduler_type):
# 1. Testing with one of the ZeRO Stages is enough to test the `_prepare_deepspeed` function.
# Here we test using ZeRO Stage 2 with FP16 enabled.
from deepspeed.runtime.engine import DeepSpeedEngine
kwargs = {
"optimizer.params.lr": 5e-5,
"optimizer.params.weight_decay": 0.0,
"scheduler.params.warmup_min_lr": 0.0,
"scheduler.params.warmup_max_lr": 5e-5,
"scheduler.params.warmup_num_steps": 0,
"train_micro_batch_size_per_gpu": 16,
"gradient_clipping": 1.0,
"train_batch_size": 16,
"zero_optimization.reduce_bucket_size": 5e5,
"zero_optimization.stage3_prefetch_bucket_size": 5e5,
"zero_optimization.stage3_param_persistence_threshold": 5e5,
"zero_optimization.stage3_gather_16bit_weights_on_model_save": False,
}
if optim_type == CUSTOM_OPTIMIZER and scheduler_type == CUSTOM_SCHEDULER:
# Test custom optimizer + custom scheduler
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=False,
zero3_init_flag=False,
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(mixed_precision="fp16", deepspeed_plugin=deepspeed_plugin)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=16, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=32, shuffle=False)
model = AutoModel.from_pretrained(GPT2_TINY)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
dummy_optimizer = DummyOptim(params=model.parameters())
dummy_lr_scheduler = DummyScheduler(dummy_optimizer)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert "You cannot create a `DummyOptim` without specifying an optimizer in the config file." in str(
cm.exception
)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
assert (
"Either specify a scheduler in the config file or "
"pass in the `lr_scheduler_callable` parameter when using `accelerate.utils.DummyScheduler`."
in str(cm.exception)
)
with self.assertRaises(ValueError) as cm:
model, optimizer, lr_scheduler = accelerator.prepare(model, optimizer, lr_scheduler)
assert (
"When using DeepSpeed, `accelerate.prepare()` requires you to pass at least one of training or evaluation dataloaders "
"with `batch_size` attribute returning an integer value "
"or alternatively set an integer value in `train_micro_batch_size_per_gpu` in the deepspeed config file "
"or assign integer value to `AcceleratorState().deepspeed_plugin.deepspeed_config['train_micro_batch_size_per_gpu']`."
in str(cm.exception)
)
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert accelerator.deepspeed_config["zero_allow_untested_optimizer"]
assert accelerator.deepspeed_config["train_batch_size"], 16
assert type(model) is DeepSpeedEngine
assert type(optimizer) is DeepSpeedOptimizerWrapper
assert type(lr_scheduler) is AcceleratedScheduler
assert type(accelerator.deepspeed_engine_wrapped) is DeepSpeedEngineWrapper
elif optim_type == DS_OPTIMIZER and scheduler_type == DS_SCHEDULER:
# Test DeepSpeed optimizer + DeepSpeed scheduler
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=self.ds_config_file[ZERO2])
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision="fp16")
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=10, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=5, shuffle=False)
model = AutoModel.from_pretrained(GPT2_TINY)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
dummy_optimizer = DummyOptim(params=model.parameters())
dummy_lr_scheduler = DummyScheduler(dummy_optimizer)
kwargs["train_batch_size"] = (
kwargs["train_micro_batch_size_per_gpu"]
* deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"]
* accelerator.num_processes
)
accelerator.state.deepspeed_plugin.deepspeed_config_process(**kwargs)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
assert "You cannot specify an optimizer in the config file and in the code at the same time" in str(
cm.exception
)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert "You cannot specify a scheduler in the config file and in the code at the same time" in str(
cm.exception
)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert "You cannot specify a scheduler in the config file and in the code at the same time" in str(
cm.exception
)
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
assert type(model) is DeepSpeedEngine
assert type(optimizer) is DeepSpeedOptimizerWrapper
assert type(lr_scheduler) is DeepSpeedSchedulerWrapper
assert type(accelerator.deepspeed_engine_wrapped) is DeepSpeedEngineWrapper
elif optim_type == CUSTOM_OPTIMIZER and scheduler_type == DS_SCHEDULER:
# Test custom optimizer + DeepSpeed scheduler
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=self.ds_config_file[ZERO2])
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision="fp16")
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=10, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=5, shuffle=False)
model = AutoModel.from_pretrained(GPT2_TINY)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
dummy_optimizer = DummyOptim(params=model.parameters())
dummy_lr_scheduler = DummyScheduler(dummy_optimizer)
kwargs["train_batch_size"] = (
kwargs["train_micro_batch_size_per_gpu"]
* deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"]
* accelerator.num_processes
)
accelerator.state.deepspeed_plugin.deepspeed_config_process(**kwargs)
del accelerator.state.deepspeed_plugin.deepspeed_config["optimizer"]
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
assert type(model) is DeepSpeedEngine
assert type(optimizer) is DeepSpeedOptimizerWrapper
assert type(lr_scheduler) is DeepSpeedSchedulerWrapper
assert type(accelerator.deepspeed_engine_wrapped) is DeepSpeedEngineWrapper
elif optim_type == DS_OPTIMIZER and scheduler_type is CUSTOM_SCHEDULER:
# Test deepspeed optimizer + custom scheduler
deepspeed_plugin = DeepSpeedPlugin(hf_ds_config=self.ds_config_file[ZERO2])
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision="fp16")
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=10, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=5, shuffle=False)
model = AutoModel.from_pretrained(GPT2_TINY)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
dummy_optimizer = DummyOptim(params=model.parameters())
dummy_lr_scheduler = DummyScheduler(dummy_optimizer)
kwargs["train_batch_size"] = (
kwargs["train_micro_batch_size_per_gpu"]
* deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"]
* accelerator.num_processes
)
accelerator.state.deepspeed_plugin.deepspeed_config_process(**kwargs)
del accelerator.state.deepspeed_plugin.deepspeed_config["scheduler"]
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert (
"You can only specify `accelerate.utils.DummyScheduler` in the code when using `accelerate.utils.DummyOptim`."
in str(cm.exception)
)
# passing `DummyScheduler` without `lr_scheduler_callable` should fail
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
assert (
"Either specify a scheduler in the config file or "
"pass in the `lr_scheduler_callable` parameter when using `accelerate.utils.DummyScheduler`."
in str(cm.exception)
)
# passing `lr_scheduler_callable` to DummyScheduler should enable DS Optim + Custom Scheduler
def _lr_scheduler_callable(optimizer):
return get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
dummy_lr_scheduler = DummyScheduler(dummy_optimizer, lr_scheduler_callable=_lr_scheduler_callable)
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
def test_dataloader_with_batch_sampler(self):
deepspeed_plugin = DeepSpeedPlugin(
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=False,
zero3_init_flag=False,
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(mixed_precision="fp16", deepspeed_plugin=deepspeed_plugin)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(
train_set, batch_sampler=BatchSampler(RandomSampler(train_set), batch_size=10, drop_last=False)
)
eval_dataloader = DataLoader(
eval_set, batch_sampler=BatchSampler(SequentialSampler(eval_set), batch_size=10, drop_last=False)
)
model = AutoModel.from_pretrained(GPT2_TINY)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
assert (
"At least one of the dataloaders passed to `accelerate.prepare()` has `None` as batch size. "
"Please set an integer value in `train_micro_batch_size_per_gpu` in the deepspeed config file "
"or assign integer value to `AcceleratorState().deepspeed_plugin.deepspeed_config['train_micro_batch_size_per_gpu']`."
in str(cm.exception)
)
def test_save_checkpoints(self):
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=self.ds_config_file[ZERO3],
zero3_init_flag=True,
)
del deepspeed_plugin.deepspeed_config["bf16"]
kwargs = {
"optimizer.params.lr": 5e-5,
"optimizer.params.weight_decay": 0.0,
"scheduler.params.warmup_min_lr": 0.0,
"scheduler.params.warmup_max_lr": 5e-5,
"scheduler.params.warmup_num_steps": 0,
"train_micro_batch_size_per_gpu": 16,
"gradient_clipping": 1.0,
"train_batch_size": 16,
"zero_optimization.reduce_bucket_size": 5e5,
"zero_optimization.stage3_prefetch_bucket_size": 5e5,
"zero_optimization.stage3_param_persistence_threshold": 5e5,
"zero_optimization.stage3_gather_16bit_weights_on_model_save": False,
}
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision="fp16")
kwargs["train_batch_size"] = (
kwargs["train_micro_batch_size_per_gpu"]
* deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"]
* accelerator.num_processes
)
accelerator.state.deepspeed_plugin.deepspeed_config_process(**kwargs)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=16, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=32, shuffle=False)
model = AutoModelForCausalLM.from_pretrained("gpt2")
dummy_optimizer = DummyOptim(params=model.parameters())
dummy_lr_scheduler = DummyScheduler(dummy_optimizer)
model, _, train_dataloader, eval_dataloader, _ = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
with self.assertRaises(ValueError) as cm:
accelerator.get_state_dict(model)
msg = (
"Cannot get 16bit model weights because `stage3_gather_16bit_weights_on_model_save` in DeepSpeed config is False. "
"To save the model weights in 16bit, set `stage3_gather_16bit_weights_on_model_save` to True in DeepSpeed config file or "
"set `zero3_save_16bit_model` to True when using `accelerate config`. "
"To save the full checkpoint, run `model.save_checkpoint(save_dir)` and use `zero_to_fp32.py` to recover weights."
)
assert msg in str(cm.exception)
def test_autofill_dsconfig(self):
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=self.ds_config_file[ZERO3],
zero3_init_flag=True,
)
del deepspeed_plugin.deepspeed_config["bf16"]
del deepspeed_plugin.deepspeed_config["fp16"]
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=16, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=32, shuffle=False)
model = AutoModelForCausalLM.from_pretrained("gpt2")
dummy_optimizer = DummyOptim(params=model.parameters(), lr=5e-5, weight_decay=1e-4)
dummy_lr_scheduler = DummyScheduler(dummy_optimizer, warmup_num_steps=10, total_num_steps=1000)
hidden_size = model.config.hidden_size
model, _, train_dataloader, eval_dataloader, _ = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
config = accelerator.deepspeed_config
assert config["train_micro_batch_size_per_gpu"] == 16
assert config["train_batch_size"] == 16
assert config["optimizer"]["params"]["lr"] == 5e-05
assert config["optimizer"]["params"]["weight_decay"] == 1e-4
assert config["scheduler"]["params"]["warmup_min_lr"] == 0.0
assert config["scheduler"]["params"]["warmup_max_lr"] == 5e-05
assert config["scheduler"]["params"]["warmup_num_steps"] == 10
assert config["gradient_clipping"] == 1.0
assert config["zero_optimization"]["reduce_bucket_size"] == (hidden_size * hidden_size)
assert config["zero_optimization"]["stage3_prefetch_bucket_size"] == int((0.9 * hidden_size) * hidden_size)
assert config["zero_optimization"]["stage3_param_persistence_threshold"] == (10 * hidden_size)
assert not config["zero_optimization"]["stage3_gather_16bit_weights_on_model_save"]
@parameterized.expand(model_types, name_func=parameterized_custom_name_func)
def test_autofill_comm_buffers_dsconfig(self, model_type):
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=self.ds_config_file[ZERO3],
zero3_init_flag=True,
)
del deepspeed_plugin.deepspeed_config["bf16"]
del deepspeed_plugin.deepspeed_config["fp16"]
del deepspeed_plugin.deepspeed_config["optimizer"]
del deepspeed_plugin.deepspeed_config["scheduler"]
with mockenv_context(**self.dist_env):
accelerator = Accelerator(mixed_precision="fp16", deepspeed_plugin=deepspeed_plugin)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=16, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=32, shuffle=False)
model = RegressionModel()
if model_type == CONFIG_WITH_NO_HIDDEN_SIZE:
model.config = DummyConfig()
elif model_type == CONFIG_WITH_HIDDEN_SIZE:
model.config = AutoConfig.from_pretrained(GPT2_TINY)
hidden_size = model.config.hidden_size
elif model_type == CONFIG_WITH_HIDDEN_SIZES:
model.config = AutoConfig.from_pretrained(MOBILEVIT)
hidden_size = max(model.config.hidden_sizes)
optimizer = torch.optim.AdamW(model.parameters(), lr=5e-5)
lr_scheduler = get_scheduler(
name="linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=1000,
)
if model_type == NO_CONFIG:
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
msg = "Can't find `model.config` entry"
assert msg in str(cm.exception)
elif model_type == CONFIG_WITH_NO_HIDDEN_SIZE:
with self.assertRaises(ValueError) as cm:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
msg = "Can find neither `model.config.hidden_size` nor `model.config.hidden_sizes`"
assert msg in str(cm.exception)
else:
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, eval_dataloader, lr_scheduler
)
zero_opt = accelerator.deepspeed_config["zero_optimization"]
assert zero_opt["reduce_bucket_size"] == (hidden_size * hidden_size)
assert zero_opt["stage3_prefetch_bucket_size"] == int((0.9 * hidden_size) * hidden_size)
assert zero_opt["stage3_param_persistence_threshold"] == (10 * hidden_size)
@parameterized.expand([FP16, BF16], name_func=parameterized_custom_name_func)
def test_autofill_dsconfig_from_ds_plugin(self, dtype):
ds_config = self.ds_config_dict["zero3"]
if dtype == BF16:
del ds_config["fp16"]
else:
del ds_config["bf16"]
ds_config[dtype]["enabled"] = "auto"
ds_config["zero_optimization"]["stage"] = "auto"
ds_config["zero_optimization"]["stage3_gather_16bit_weights_on_model_save"] = "auto"
ds_config["zero_optimization"]["offload_optimizer"]["device"] = "auto"
ds_config["zero_optimization"]["offload_param"]["device"] = "auto"
ds_config["gradient_accumulation_steps"] = "auto"
ds_config["gradient_clipping"] = "auto"
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=ds_config,
zero3_init_flag=True,
gradient_accumulation_steps=2,
gradient_clipping=1.0,
zero_stage=2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision=dtype)
config = accelerator.state.deepspeed_plugin.deepspeed_config
assert config["gradient_clipping"] == 1.0
assert config["gradient_accumulation_steps"] == 2
assert config["zero_optimization"]["stage"] == 2
assert config["zero_optimization"]["offload_optimizer"]["device"] == "cpu"
assert config["zero_optimization"]["offload_param"]["device"] == "cpu"
assert config["zero_optimization"]["stage3_gather_16bit_weights_on_model_save"]
assert config[dtype]["enabled"]
AcceleratorState._reset_state(True)
diff_dtype = "bf16" if dtype == "fp16" else "fp16"
with mockenv_context(**self.dist_env):
with self.assertRaises(ValueError) as cm:
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision=diff_dtype)
assert (
f"`--mixed_precision` arg cannot be set to `{diff_dtype}` when `{dtype}` is set in the DeepSpeed config file."
in str(cm.exception)
)
# base case of passing in `gradient_accumulation_steps` to `DeepSpeedPlugin`
AcceleratorState._reset_state(True)
deepspeed_plugin = DeepSpeedPlugin(zero_stage=2, gradient_accumulation_steps=4)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision=dtype)
deepspeed_plugin = accelerator.state.deepspeed_plugin
assert deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"] == 4
# filling the `auto` gradient_accumulation_steps via Accelerator's value
AcceleratorState._reset_state(True)
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=ds_config,
zero3_init_flag=True,
gradient_clipping=1.0,
zero_stage=2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(
deepspeed_plugin=deepspeed_plugin, mixed_precision=dtype, gradient_accumulation_steps=8
)
train_set = RegressionDataset(length=80)
eval_set = RegressionDataset(length=20)
train_dataloader = DataLoader(train_set, batch_size=16, shuffle=True)
eval_dataloader = DataLoader(eval_set, batch_size=32, shuffle=False)
model = AutoModelForCausalLM.from_pretrained("gpt2")
dummy_optimizer = DummyOptim(params=model.parameters(), lr=5e-5, weight_decay=1e-4)
dummy_lr_scheduler = DummyScheduler(dummy_optimizer, warmup_num_steps=10, total_num_steps=1000)
model, _, train_dataloader, eval_dataloader, _ = accelerator.prepare(
model, dummy_optimizer, train_dataloader, eval_dataloader, dummy_lr_scheduler
)
deepspeed_plugin = accelerator.state.deepspeed_plugin
assert deepspeed_plugin.deepspeed_config["gradient_accumulation_steps"] == 8
def test_ds_config_assertions(self):
ambiguous_env = self.dist_env.copy()
ambiguous_env["ACCELERATE_CONFIG_DS_FIELDS"] = (
"gradient_accumulation_steps,gradient_clipping,zero_stage,offload_optimizer_device,offload_param_device,zero3_save_16bit_model,mixed_precision"
)
with mockenv_context(**ambiguous_env):
with self.assertRaises(ValueError) as cm:
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=self.ds_config_file[ZERO3],
zero3_init_flag=True,
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=ZERO2,
offload_optimizer_device="cpu",
offload_param_device="cpu",
zero3_save_16bit_model=True,
)
_ = Accelerator(deepspeed_plugin=deepspeed_plugin, mixed_precision=FP16)
assert (
"If you are using an accelerate config file, remove others config variables mentioned in the above specified list."
in str(cm.exception)
)
def test_ds_zero3_no_init_autofill(self):
ds_config = {
"bf16": {"enabled": True},
"zero_optimization": {
"stage": 3,
"allgather_partitions": True,
"allgather_bucket_size": 5e8,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": "auto",
"contiguous_gradients": True,
"stage3_gather_16bit_weights_on_model_save": False,
"offload_optimizer": {"device": "none"},
"offload_param": {"device": "none"},
},
"gradient_clipping": 1.0,
"gradient_accumulation_steps": 1,
"train_batch_size": "auto",
"train_micro_batch_size_per_gpu": "auto",
"steps_per_print": 2000000,
}
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=ds_config,
zero3_init_flag=False,
)
with mockenv_context(**self.dist_env):
_ = Accelerator(deepspeed_plugin=deepspeed_plugin)
_ = AutoModelForCausalLM.from_pretrained("gpt2")
@parameterized.expand(stages, name_func=parameterized_custom_name_func)
def test_ds_config(self, stage):
deepspeed_plugin = DeepSpeedPlugin(
hf_ds_config=self.ds_config_file[stage],
zero3_init_flag=True,
)
assert deepspeed_plugin.zero_stage == int(stage.replace("zero", ""))
def test_prepare_deepspeed_prepare_moe(self):
if compare_versions("transformers", "<", "4.40") and compare_versions("deepspeed", "<", "0.14"):
return
deepspeed_plugin = DeepSpeedPlugin(
zero3_init_flag=True,
gradient_accumulation_steps=1,
gradient_clipping=1.0,
zero_stage=3,
offload_optimizer_device="none",
offload_param_device="none",
zero3_save_16bit_model=True,
transformer_moe_cls_names="Qwen2MoeSparseMoeBlock",
)
with mockenv_context(**self.dist_env):
accelerator = Accelerator(mixed_precision="fp16", deepspeed_plugin=deepspeed_plugin)
accelerator.state.deepspeed_plugin.deepspeed_config["train_micro_batch_size_per_gpu"] = 1
model = AutoModelForCausalLM.from_pretrained(QWEN_MOE)
model = accelerator.prepare(model)
from transformers.models.qwen2_moe.modeling_qwen2_moe import Qwen2MoeSparseMoeBlock
for module in model.modules():
if isinstance(module, Qwen2MoeSparseMoeBlock):
assert hasattr(module, "_z3_leaf") and module._z3_leaf
def test_basic_run(self):
test_file_path = path_in_accelerate_package("test_utils", "scripts", "external_deps", "test_performance.py")
with tempfile.TemporaryDirectory() as dirpath:
cmd = [
"accelerate",
"launch",
"--num_processes=1",
"--num_machines=1",
"--machine_rank=0",
"--mixed_precision=fp16",
"--use_deepspeed",
"--gradient_accumulation_steps=1",
"--zero_stage=2",
"--offload_optimizer_device=none",
"--offload_param_device=none",
test_file_path,
"--model_name_or_path=distilbert-base-uncased",
"--num_epochs=1",
f"--output_dir={dirpath}",
]
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd)
@require_deepspeed
@require_multi_device
@slow
class DeepSpeedIntegrationTest(TempDirTestCase):
test_scripts_folder = path_in_accelerate_package("test_utils", "scripts", "external_deps")
def setUp(self):
super().setUp()
self._test_file_path = inspect.getfile(self.__class__)
path = Path(self._test_file_path).resolve()
self.test_file_dir_str = str(path.parents[0])
self.ds_config_file = dict(
zero2=f"{self.test_file_dir_str}/ds_config_zero2.json",
zero3=f"{self.test_file_dir_str}/ds_config_zero3.json",
)
self.stages = [1, 2, 3]
self.zero3_offload_config = False
self.performance_lower_bound = 0.82
self.peak_memory_usage_upper_bound = {
"multi_gpu_fp16": 3200,
"deepspeed_stage_1_fp16": 1600,
"deepspeed_stage_2_fp16": 2500,
"deepspeed_stage_3_zero_init_fp16": 2800,
# Disabling below test as it overwhelms the RAM memory usage
# on CI self-hosted runner leading to tests getting killed.
# "deepspeed_stage_3_cpu_offload_fp16": 1900,
}
self.n_train = 160
self.n_val = 160
def test_performance(self):
self.test_file_path = self.test_scripts_folder / "test_performance.py"
cmd = [
"accelerate",
"launch",
"--num_processes=2",
"--num_machines=1",
"--machine_rank=0",
"--mixed_precision=fp16",
"--use_deepspeed",
"--gradient_accumulation_steps=1",
"--gradient_clipping=1",
"--zero3_init_flag=True",
"--zero3_save_16bit_model=True",
]
for stage in self.stages:
if stage == 1:
continue
cmd_stage = cmd.copy()
cmd_stage.extend([f"--zero_stage={stage}"])
cmd_stage.extend(["--offload_optimizer_device=none", "--offload_param_device=none"])
if self.zero3_offload_config:
with open(self.ds_config_file[ZERO3], encoding="utf-8") as f:
ds_config = json.load(f)
del ds_config["bf16"]
del ds_config["optimizer"]["params"]["torch_adam"]
del ds_config["optimizer"]["params"]["adam_w_mode"]
ds_config["fp16"]["enabled"] = True
ds_config_path = os.path.join(self.tmpdir, "ds_config.json")
with open(ds_config_path, "w") as out_file:
json.dump(ds_config, out_file)
cmd_stage.extend([f"--deepspeed_config_file={ds_config_path}"])
cmd_stage.extend(
[
self.test_file_path,
f"--output_dir={self.tmpdir}",
f"--performance_lower_bound={self.performance_lower_bound}",
]
)
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd_stage)
def test_checkpointing(self):
self.test_file_path = self.test_scripts_folder / "test_checkpointing.py"
cmd = [
"accelerate",
"launch",
"--num_processes=2",
"--num_machines=1",
"--machine_rank=0",
"--mixed_precision=fp16",
"--use_deepspeed",
"--gradient_accumulation_steps=1",
"--gradient_clipping=1",
"--zero3_init_flag=True",
"--zero3_save_16bit_model=True",
]
for stage in self.stages:
if stage == 1:
continue
cmd_stage = cmd.copy()
cmd_stage.extend([f"--zero_stage={stage}"])
cmd_stage.extend(["--offload_optimizer_device=none", "--offload_param_device=none"])
if self.zero3_offload_config:
with open(self.ds_config_file[ZERO3], encoding="utf-8") as f:
ds_config = json.load(f)
del ds_config["bf16"]
del ds_config["optimizer"]["params"]["torch_adam"]
del ds_config["optimizer"]["params"]["adam_w_mode"]
ds_config["fp16"]["enabled"] = True
ds_config_path = os.path.join(self.tmpdir, "ds_config.json")
with open(ds_config_path, "w") as out_file:
json.dump(ds_config, out_file)
cmd_stage.extend([f"--deepspeed_config_file={ds_config_path}"])
cmd_stage.extend(
[
self.test_file_path,
f"--output_dir={self.tmpdir}",
"--partial_train_epoch=1",
]
)
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd_stage)
cmd_stage = cmd_stage[:-1]
resume_from_checkpoint = os.path.join(self.tmpdir, "epoch_0")
cmd_stage.extend(
[
f"--resume_from_checkpoint={resume_from_checkpoint}",
]
)
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd_stage)
def test_peak_memory_usage(self):
if compare_versions("deepspeed", ">", "0.12.6"):
self.skipTest(
"The test fails when deepspeed>0.12.6. This is something that needs to be fixed on deepspeed library"
)
self.test_file_path = self.test_scripts_folder / "test_peak_memory_usage.py"
cmd = [
"accelerate",
"launch",
"--num_processes=2",
"--num_machines=1",
"--machine_rank=0",
]
for spec, peak_mem_upper_bound in self.peak_memory_usage_upper_bound.items():
cmd_stage = cmd.copy()
if "fp16" in spec:
cmd_stage.extend(["--mixed_precision=fp16"])
if "multi_gpu" in spec:
continue
else:
cmd_stage.extend(
[
"--use_deepspeed",
"--gradient_accumulation_steps=1",
"--gradient_clipping=1",
"--zero3_init_flag=True",
"--zero3_save_16bit_model=True",
]
)
for i in range(3):
if f"stage_{i + 1}" in spec:
cmd_stage.extend([f"--zero_stage={i + 1}"])
break
cmd_stage.extend(
[
"--offload_optimizer_device=none",
"--offload_param_device=none",
"--offload_optimizer_nvme_path=none",
"--offload_param_nvme_path=none",
]
)
if "cpu_offload" in spec:
with open(self.ds_config_file[ZERO3], encoding="utf-8") as f:
ds_config = json.load(f)
del ds_config["bf16"]
del ds_config["fp16"]
del ds_config["optimizer"]["params"]["torch_adam"]
del ds_config["optimizer"]["params"]["adam_w_mode"]
ds_config_path = os.path.join(self.tmpdir, "ds_config.json")
with open(ds_config_path, "w") as out_file:
json.dump(ds_config, out_file)
cmd_stage.extend([f"--deepspeed_config_file={ds_config_path}"])
cmd_stage.extend(
[
self.test_file_path,
f"--output_dir={self.tmpdir}",
f"--peak_memory_upper_bound={peak_mem_upper_bound}",
f"--n_train={self.n_train}",
f"--n_val={self.n_val}",
]
)
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd_stage)
def test_lr_scheduler(self):
self.test_file_path = self.test_scripts_folder / "test_performance.py"
cmd = [
"accelerate",
"launch",
"--num_processes=2",
"--num_machines=1",
"--machine_rank=0",
"--mixed_precision=no",
"--use_deepspeed",
"--gradient_accumulation_steps=1",
"--gradient_clipping=1",
"--zero3_init_flag=True",
"--zero3_save_16bit_model=True",
"--zero_stage=3",
"--offload_optimizer_device=none",
"--offload_param_device=none",
self.test_file_path,
f"--output_dir={self.tmpdir}",
f"--performance_lower_bound={self.performance_lower_bound}",
]
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd)
@require_huggingface_suite
def test_zero3_integration(self):
self.test_file_path = self.test_scripts_folder / "test_zero3_integration.py"
cmd = ["accelerate", "launch", "--num_processes=2", "--num_machines=1", self.test_file_path]
with patch_environment(omp_num_threads=1):
execute_subprocess_async(cmd)
| accelerate/tests/deepspeed/test_deepspeed.py/0 | {
"file_path": "accelerate/tests/deepspeed/test_deepspeed.py",
"repo_id": "accelerate",
"token_count": 26455
} |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import ast
import os
import re
import shutil
import tempfile
import unittest
from pathlib import Path
from unittest import mock, skip
import torch
from accelerate.test_utils.examples import compare_against_test
from accelerate.test_utils.testing import (
TempDirTestCase,
get_launch_command,
require_huggingface_suite,
require_multi_device,
require_multi_gpu,
require_non_xpu,
require_pippy,
require_schedulefree,
require_trackers,
run_command,
slow,
)
from accelerate.utils import write_basic_config
# DataLoaders built from `test_samples/MRPC` for quick testing
# Should mock `{script_name}.get_dataloaders` via:
# @mock.patch("{script_name}.get_dataloaders", mocked_dataloaders)
EXCLUDE_EXAMPLES = [
"cross_validation.py",
"checkpointing.py",
"gradient_accumulation.py",
"local_sgd.py",
"multi_process_metrics.py",
"memory.py",
"schedule_free.py",
"tracking.py",
"automatic_gradient_accumulation.py",
"gradient_accumulation_for_autoregressive_models.py",
"fsdp_with_peak_mem_tracking.py",
"deepspeed_with_config_support.py",
"megatron_lm_gpt_pretraining.py",
"early_stopping.py",
"ddp_comm_hook.py",
"profiler.py",
]
class ExampleDifferenceTests(unittest.TestCase):
"""
This TestCase checks that all of the `complete_*` scripts contain all of the
information found in the `by_feature` scripts, line for line. If one fails,
then a complete example does not contain all of the features in the features
scripts, and should be updated.
Each example script should be a single test (such as `test_nlp_example`),
and should run `one_complete_example` twice: once with `parser_only=True`,
and the other with `parser_only=False`. This is so that when the test
failures are returned to the user, they understand if the discrepancy lies in
the `main` function, or the `training_loop` function. Otherwise it will be
unclear.
Also, if there are any expected differences between the base script used and
`complete_nlp_example.py` (the canonical base script), these should be included in
`special_strings`. These would be differences in how something is logged, print statements,
etc (such as calls to `Accelerate.log()`)
"""
by_feature_path = Path("examples", "by_feature").resolve()
examples_path = Path("examples").resolve()
def one_complete_example(
self, complete_file_name: str, parser_only: bool, secondary_filename: str = None, special_strings: list = None
):
"""
Tests a single `complete` example against all of the implemented `by_feature` scripts
Args:
complete_file_name (`str`):
The filename of a complete example
parser_only (`bool`):
Whether to look at the main training function, or the argument parser
secondary_filename (`str`, *optional*):
A potential secondary base file to strip all script information not relevant for checking,
such as "cv_example.py" when testing "complete_cv_example.py"
special_strings (`list`, *optional*):
A list of strings to potentially remove before checking no differences are left. These should be
diffs that are file specific, such as different logging variations between files.
"""
self.maxDiff = None
for item in os.listdir(self.by_feature_path):
if item not in EXCLUDE_EXAMPLES:
item_path = self.by_feature_path / item
if item_path.is_file() and item_path.suffix == ".py":
with self.subTest(
tested_script=complete_file_name,
feature_script=item,
tested_section="main()" if parser_only else "training_function()",
):
diff = compare_against_test(
self.examples_path / complete_file_name, item_path, parser_only, secondary_filename
)
diff = "\n".join(diff)
if special_strings is not None:
for string in special_strings:
diff = diff.replace(string, "")
assert diff == ""
def test_nlp_examples(self):
self.one_complete_example("complete_nlp_example.py", True)
self.one_complete_example("complete_nlp_example.py", False)
def test_cv_examples(self):
cv_path = (self.examples_path / "cv_example.py").resolve()
special_strings = [
" " * 16 + "{\n\n",
" " * 20 + '"accuracy": eval_metric["accuracy"],\n\n',
" " * 20 + '"f1": eval_metric["f1"],\n\n',
" " * 20 + '"train_loss": total_loss.item() / len(train_dataloader),\n\n',
" " * 20 + '"epoch": epoch,\n\n',
" " * 16 + "},\n\n",
" " * 16 + "step=epoch,\n",
" " * 12,
" " * 8 + "for step, batch in enumerate(active_dataloader):\n",
]
self.one_complete_example("complete_cv_example.py", True, cv_path, special_strings)
self.one_complete_example("complete_cv_example.py", False, cv_path, special_strings)
@mock.patch.dict(os.environ, {"TESTING_MOCKED_DATALOADERS": "1"})
@require_huggingface_suite
class FeatureExamplesTests(TempDirTestCase):
clear_on_setup = False
@classmethod
def setUpClass(cls):
super().setUpClass()
cls._tmpdir = tempfile.mkdtemp()
cls.config_file = Path(cls._tmpdir) / "default_config.yml"
write_basic_config(save_location=cls.config_file)
cls.launch_args = get_launch_command(config_file=cls.config_file)
@classmethod
def tearDownClass(cls):
super().tearDownClass()
shutil.rmtree(cls._tmpdir)
def test_checkpointing_by_epoch(self):
testargs = f"""
examples/by_feature/checkpointing.py
--checkpointing_steps epoch
--output_dir {self.tmpdir}
""".split()
run_command(self.launch_args + testargs)
assert (self.tmpdir / "epoch_0").exists()
def test_checkpointing_by_steps(self):
testargs = f"""
examples/by_feature/checkpointing.py
--checkpointing_steps 1
--output_dir {self.tmpdir}
""".split()
_ = run_command(self.launch_args + testargs)
assert (self.tmpdir / "step_2").exists()
def test_load_states_by_epoch(self):
testargs = f"""
examples/by_feature/checkpointing.py
--resume_from_checkpoint {self.tmpdir / "epoch_0"}
""".split()
output = run_command(self.launch_args + testargs, return_stdout=True)
assert "epoch 0:" not in output
assert "epoch 1:" in output
def test_load_states_by_steps(self):
testargs = f"""
examples/by_feature/checkpointing.py
--resume_from_checkpoint {self.tmpdir / "step_2"}
""".split()
output = run_command(self.launch_args + testargs, return_stdout=True)
if torch.cuda.is_available():
num_processes = torch.cuda.device_count()
else:
num_processes = 1
if num_processes > 1:
assert "epoch 0:" not in output
assert "epoch 1:" in output
else:
assert "epoch 0:" in output
assert "epoch 1:" in output
@slow
def test_cross_validation(self):
testargs = """
examples/by_feature/cross_validation.py
--num_folds 2
""".split()
with mock.patch.dict(os.environ, {"TESTING_MOCKED_DATALOADERS": "0"}):
output = run_command(self.launch_args + testargs, return_stdout=True)
results = re.findall("({.+})", output)
results = [r for r in results if "accuracy" in r][-1]
results = ast.literal_eval(results)
assert results["accuracy"] >= 0.75
def test_multi_process_metrics(self):
testargs = ["examples/by_feature/multi_process_metrics.py"]
run_command(self.launch_args + testargs)
@require_schedulefree
def test_schedulefree(self):
testargs = ["examples/by_feature/schedule_free.py"]
run_command(self.launch_args + testargs)
@require_trackers
@mock.patch.dict(os.environ, {"WANDB_MODE": "offline", "DVCLIVE_TEST": "true"})
def test_tracking(self):
with tempfile.TemporaryDirectory() as tmpdir:
testargs = f"""
examples/by_feature/tracking.py
--with_tracking
--project_dir {tmpdir}
""".split()
run_command(self.launch_args + testargs)
assert os.path.exists(os.path.join(tmpdir, "tracking"))
def test_gradient_accumulation(self):
testargs = ["examples/by_feature/gradient_accumulation.py"]
run_command(self.launch_args + testargs)
def test_gradient_accumulation_for_autoregressive_models(self):
testargs = [
"examples/by_feature/gradient_accumulation_for_autoregressive_models.py",
"--gradient_accumulation_steps",
"2",
]
run_command(self.launch_args + testargs)
def test_local_sgd(self):
testargs = ["examples/by_feature/local_sgd.py"]
run_command(self.launch_args + testargs)
def test_early_stopping(self):
testargs = ["examples/by_feature/early_stopping.py"]
run_command(self.launch_args + testargs)
def test_profiler(self):
testargs = ["examples/by_feature/profiler.py"]
run_command(self.launch_args + testargs)
@require_multi_device
def test_ddp_comm_hook(self):
testargs = ["examples/by_feature/ddp_comm_hook.py", "--ddp_comm_hook", "fp16"]
run_command(self.launch_args + testargs)
@skip(
reason="stable-diffusion-v1-5 is no longer available. Potentially `Comfy-Org/stable-diffusion-v1-5-archive` once diffusers support is added."
)
@require_multi_device
def test_distributed_inference_examples_stable_diffusion(self):
testargs = ["examples/inference/distributed/stable_diffusion.py"]
run_command(self.launch_args + testargs)
@require_multi_device
def test_distributed_inference_examples_phi2(self):
testargs = ["examples/inference/distributed/phi2.py"]
run_command(self.launch_args + testargs)
@require_non_xpu
@require_pippy
@require_multi_gpu
def test_pippy_examples_bert(self):
testargs = ["examples/inference/pippy/bert.py"]
run_command(self.launch_args + testargs)
@require_non_xpu
@require_pippy
@require_multi_gpu
def test_pippy_examples_gpt2(self):
testargs = ["examples/inference/pippy/gpt2.py"]
run_command(self.launch_args + testargs)
| accelerate/tests/test_examples.py/0 | {
"file_path": "accelerate/tests/test_examples.py",
"repo_id": "accelerate",
"token_count": 4893
} |
repos:
- repo: https://github.com/Narsil/pre-commit-rust
rev: 2eed6366172ef2a5186e8785ec0e67243d7d73d0
hooks:
- id: fmt
name: "Rust (fmt)"
- id: clippy
name: "Rust (clippy)"
args:
[
"--tests",
"--examples",
"--",
"-Dwarnings",
]
| candle/.pre-commit-config.yaml/0 | {
"file_path": "candle/.pre-commit-config.yaml",
"repo_id": "candle",
"token_count": 210
} |
//! #A simplified example in Rust of training a neural network and then using it based on the Candle Framework by Hugging Face.
//! Author: Evgeny Igumnov 2023 [email protected]
//! This program implements a neural network to predict the winner of the second round of elections based on the results of the first round.
//!
//! ##Basic moments:
//!
//! A multilayer perceptron with two hidden layers is used. The first hidden layer has 4 neurons, the second has 2 neurons.
//! The input is a vector of 2 numbers - the percentage of votes for the first and second candidates in the first stage.
//! The output is the number 0 or 1, where 1 means that the first candidate will win in the second stage, 0 means that he will lose.
//! For training, samples with real data on the results of the first and second stages of different elections are used.
//! The model is trained by backpropagation using gradient descent and the cross-entropy loss function.
//! Model parameters (weights of neurons) are initialized randomly, then optimized during training.
//! After training, the model is tested on a deferred sample to evaluate the accuracy.
//! If the accuracy on the test set is below 100%, the model is considered underfit and the learning process is repeated.
//! Thus, this neural network learns to find hidden relationships between the results of the first and second rounds of voting in order to make predictions for new data.
#[rustfmt::skip]
mod tests {
use candle::{DType, Result, Tensor, D, Device};
use candle_nn::{loss, ops, Linear, Module, VarBuilder, VarMap, Optimizer};
// ANCHOR: book_training_simplified1
const VOTE_DIM: usize = 2;
const RESULTS: usize = 1;
const EPOCHS: usize = 10;
const LAYER1_OUT_SIZE: usize = 4;
const LAYER2_OUT_SIZE: usize = 2;
const LEARNING_RATE: f64 = 0.05;
#[derive(Clone)]
pub struct Dataset {
pub train_votes: Tensor,
pub train_results: Tensor,
pub test_votes: Tensor,
pub test_results: Tensor,
}
struct MultiLevelPerceptron {
ln1: Linear,
ln2: Linear,
ln3: Linear,
}
impl MultiLevelPerceptron {
fn new(vs: VarBuilder) -> Result<Self> {
let ln1 = candle_nn::linear(VOTE_DIM, LAYER1_OUT_SIZE, vs.pp("ln1"))?;
let ln2 = candle_nn::linear(LAYER1_OUT_SIZE, LAYER2_OUT_SIZE, vs.pp("ln2"))?;
let ln3 = candle_nn::linear(LAYER2_OUT_SIZE, RESULTS + 1, vs.pp("ln3"))?;
Ok(Self { ln1, ln2, ln3 })
}
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
let xs = self.ln1.forward(xs)?;
let xs = xs.relu()?;
let xs = self.ln2.forward(&xs)?;
let xs = xs.relu()?;
self.ln3.forward(&xs)
}
}
// ANCHOR_END: book_training_simplified1
// ANCHOR: book_training_simplified3
#[tokio::test]
async fn simplified() -> anyhow::Result<()> {
let dev = Device::cuda_if_available(0)?;
let train_votes_vec: Vec<u32> = vec![
15, 10,
10, 15,
5, 12,
30, 20,
16, 12,
13, 25,
6, 14,
31, 21,
];
let train_votes_tensor = Tensor::from_vec(train_votes_vec.clone(), (train_votes_vec.len() / VOTE_DIM, VOTE_DIM), &dev)?.to_dtype(DType::F32)?;
let train_results_vec: Vec<u32> = vec![
1,
0,
0,
1,
1,
0,
0,
1,
];
let train_results_tensor = Tensor::from_vec(train_results_vec, train_votes_vec.len() / VOTE_DIM, &dev)?;
let test_votes_vec: Vec<u32> = vec![
13, 9,
8, 14,
3, 10,
];
let test_votes_tensor = Tensor::from_vec(test_votes_vec.clone(), (test_votes_vec.len() / VOTE_DIM, VOTE_DIM), &dev)?.to_dtype(DType::F32)?;
let test_results_vec: Vec<u32> = vec![
1,
0,
0,
];
let test_results_tensor = Tensor::from_vec(test_results_vec.clone(), test_results_vec.len(), &dev)?;
let m = Dataset {
train_votes: train_votes_tensor,
train_results: train_results_tensor,
test_votes: test_votes_tensor,
test_results: test_results_tensor,
};
let trained_model: MultiLevelPerceptron;
loop {
println!("Trying to train neural network.");
match train(m.clone(), &dev) {
Ok(model) => {
trained_model = model;
break;
},
Err(e) => {
println!("Error: {}", e);
continue;
}
}
}
let real_world_votes: Vec<u32> = vec![
13, 22,
];
let tensor_test_votes = Tensor::from_vec(real_world_votes.clone(), (1, VOTE_DIM), &dev)?.to_dtype(DType::F32)?;
let final_result = trained_model.forward(&tensor_test_votes)?;
let result = final_result
.argmax(D::Minus1)?
.to_dtype(DType::F32)?
.get(0).map(|x| x.to_scalar::<f32>())??;
println!("real_life_votes: {:?}", real_world_votes);
println!("neural_network_prediction_result: {:?}", result);
Ok(())
}
// ANCHOR_END: book_training_simplified3
// ANCHOR: book_training_simplified2
fn train(m: Dataset, dev: &Device) -> anyhow::Result<MultiLevelPerceptron> {
let train_results = m.train_results.to_device(dev)?;
let train_votes = m.train_votes.to_device(dev)?;
let varmap = VarMap::new();
let vs = VarBuilder::from_varmap(&varmap, DType::F32, dev);
let model = MultiLevelPerceptron::new(vs.clone())?;
let mut sgd = candle_nn::SGD::new(varmap.all_vars(), LEARNING_RATE)?;
let test_votes = m.test_votes.to_device(dev)?;
let test_results = m.test_results.to_device(dev)?;
let mut final_accuracy: f32 = 0.0;
for epoch in 1..EPOCHS + 1 {
let logits = model.forward(&train_votes)?;
let log_sm = ops::log_softmax(&logits, D::Minus1)?;
let loss = loss::nll(&log_sm, &train_results)?;
sgd.backward_step(&loss)?;
let test_logits = model.forward(&test_votes)?;
let sum_ok = test_logits
.argmax(D::Minus1)?
.eq(&test_results)?
.to_dtype(DType::F32)?
.sum_all()?
.to_scalar::<f32>()?;
let test_accuracy = sum_ok / test_results.dims1()? as f32;
final_accuracy = 100. * test_accuracy;
println!("Epoch: {epoch:3} Train loss: {:8.5} Test accuracy: {:5.2}%",
loss.to_scalar::<f32>()?,
final_accuracy
);
if final_accuracy == 100.0 {
break;
}
}
if final_accuracy < 100.0 {
Err(anyhow::Error::msg("The model is not trained well enough."))
} else {
Ok(model)
}
}
// ANCHOR_END: book_training_simplified2
}
| candle/candle-book/src/simplified.rs/0 | {
"file_path": "candle/candle-book/src/simplified.rs",
"repo_id": "candle",
"token_count": 2903
} |
use crate::benchmarks::{BenchDevice, BenchDeviceHandler};
use candle_core::{DType, Device, Tensor};
use criterion::{black_box, criterion_group, Criterion, Throughput};
use half::{bf16, f16};
use std::time::Instant;
fn run_sum(a: &Tensor) {
a.sum_keepdim(2).unwrap();
}
fn run_arg_min(a: &Tensor) {
a.argmin_keepdim(2).unwrap();
}
fn criterion_benchmark(c: &mut Criterion) {
let handler = BenchDeviceHandler::new().unwrap();
let (lo, up) = (-1000.0f32, 1000.0f32);
for device in handler.devices {
run_reduce(c, &device, (lo, up), false);
run_reduce(c, &device, (f16::from_f32(lo), f16::from_f32(up)), false);
run_reduce(c, &device, (bf16::from_f32(lo), bf16::from_f32(up)), false);
run_arg_reduce(c, &device, (lo, up), false);
run_arg_reduce(c, &device, (f16::from_f32(lo), f16::from_f32(up)), false);
run_arg_reduce(c, &device, (bf16::from_f32(lo), bf16::from_f32(up)), false);
run_reduce(c, &device, (lo, up), true);
run_reduce(c, &device, (f16::from_f32(lo), f16::from_f32(up)), true);
run_reduce(c, &device, (bf16::from_f32(lo), bf16::from_f32(up)), true);
run_arg_reduce(c, &device, (lo, up), true);
run_arg_reduce(c, &device, (f16::from_f32(lo), f16::from_f32(up)), true);
run_arg_reduce(c, &device, (bf16::from_f32(lo), bf16::from_f32(up)), true);
}
}
fn run_reduce<T: candle_core::FloatDType>(
c: &mut Criterion,
device: &Device,
(lo, up): (T, T),
strided: bool,
) {
let b = 1;
let m = 1024;
let k = 1024;
let a = if strided {
Tensor::rand(lo, up, (b, m, k), &device)
.unwrap()
.transpose(0, 2)
.unwrap()
} else {
Tensor::rand(lo, up, (b, m, k), &device).unwrap()
};
let flops = b * m * k * T::DTYPE.size_in_bytes();
let name = match T::DTYPE {
DType::F32 => {
if strided {
"reduce_f32_strided"
} else {
"reduce_f32"
}
}
DType::F16 => {
if strided {
"reduce_f16_strided"
} else {
"reduce_f16"
}
}
DType::BF16 => {
if strided {
"reduce_bf16_strided"
} else {
"reduce_bf16"
}
}
_ => "unknown",
};
let mut group = c.benchmark_group(device.bench_name(name));
group.throughput(Throughput::Bytes(flops as u64));
group.bench_function("iter", move |b| {
b.iter_custom(|iters| {
let start = Instant::now();
for _i in 0..iters {
run_sum(black_box(&a));
}
device.sync().unwrap();
start.elapsed()
})
});
group.finish();
}
fn run_arg_reduce<T: candle_core::FloatDType>(
c: &mut Criterion,
device: &Device,
(lo, up): (T, T),
strided: bool,
) {
let b = 1;
let m = 1024;
let k = 1024;
let a = if strided {
Tensor::rand(lo, up, (b, m, k), &device)
.unwrap()
.transpose(0, 2)
.unwrap()
} else {
Tensor::rand(lo, up, (b, m, k), &device).unwrap()
};
let flops = b * m * k * T::DTYPE.size_in_bytes();
let name = match T::DTYPE {
DType::F32 => {
if strided {
"arg_reduce_f32_strided"
} else {
"arg_reduce_f32"
}
}
DType::F16 => {
if strided {
"arg_reduce_f16_strided"
} else {
"arg_reduce_f16"
}
}
DType::BF16 => {
if strided {
"arg_reduce_bf16_strided"
} else {
"arg_reduce_bf16"
}
}
_ => "unknown",
};
let mut group = c.benchmark_group(device.bench_name(name));
group.throughput(Throughput::Bytes(flops as u64));
group.bench_function("iter", move |b| {
b.iter_custom(|iters| {
let start = Instant::now();
for _i in 0..iters {
run_arg_min(black_box(&a));
}
device.sync().unwrap();
start.elapsed()
})
});
group.finish();
}
criterion_group!(benches, criterion_benchmark);
| candle/candle-core/benches/benchmarks/reduce.rs/0 | {
"file_path": "candle/candle-core/benches/benchmarks/reduce.rs",
"repo_id": "candle",
"token_count": 2382
} |
use super::Cpu;
#[cfg(target_arch = "arm")]
use core::arch::arm::*;
#[cfg(target_arch = "aarch64")]
use core::arch::aarch64::*;
pub struct CurrentCpu {}
const STEP: usize = 16;
const EPR: usize = 4;
const ARR: usize = STEP / EPR;
impl CurrentCpu {
#[cfg(target_arch = "aarch64")]
unsafe fn reduce_one(x: float32x4_t) -> f32 {
vaddvq_f32(x)
}
#[cfg(target_arch = "arm")]
unsafe fn reduce_one(x: float32x4_t) -> f32 {
vgetq_lane_f32(x, 0) + vgetq_lane_f32(x, 1) + vgetq_lane_f32(x, 2) + vgetq_lane_f32(x, 3)
}
}
impl Cpu<ARR> for CurrentCpu {
type Unit = float32x4_t;
type Array = [float32x4_t; ARR];
const STEP: usize = STEP;
const EPR: usize = EPR;
fn n() -> usize {
ARR
}
unsafe fn zero() -> Self::Unit {
vdupq_n_f32(0.0)
}
unsafe fn from_f32(x: f32) -> Self::Unit {
vdupq_n_f32(x)
}
unsafe fn zero_array() -> Self::Array {
[Self::zero(); ARR]
}
unsafe fn load(mem_addr: *const f32) -> Self::Unit {
vld1q_f32(mem_addr)
}
unsafe fn vec_add(a: Self::Unit, b: Self::Unit) -> Self::Unit {
vaddq_f32(a, b)
}
unsafe fn vec_fma(a: Self::Unit, b: Self::Unit, c: Self::Unit) -> Self::Unit {
vfmaq_f32(a, b, c)
}
unsafe fn vec_store(mem_addr: *mut f32, a: Self::Unit) {
vst1q_f32(mem_addr, a);
}
unsafe fn vec_reduce(mut x: Self::Array, y: *mut f32) {
for i in 0..ARR / 2 {
x[2 * i] = vaddq_f32(x[2 * i], x[2 * i + 1]);
}
for i in 0..ARR / 4 {
x[4 * i] = vaddq_f32(x[4 * i], x[4 * i + 2]);
}
*y = Self::reduce_one(x[0]);
}
}
| candle/candle-core/src/cpu/neon.rs/0 | {
"file_path": "candle/candle-core/src/cpu/neon.rs",
"repo_id": "candle",
"token_count": 897
} |
use crate::{Error, Tensor};
use std::ops::{
Bound, Range, RangeBounds, RangeFrom, RangeFull, RangeInclusive, RangeTo, RangeToInclusive,
};
impl Tensor {
/// Intended to be use by the trait `.i()`
///
/// ```
/// # use candle_core::{Tensor, DType, Device, IndexOp};
/// let a = Tensor::zeros((2, 3), DType::F32, &Device::Cpu)?;
///
/// let c = a.i(0..1)?;
/// assert_eq!(c.shape().dims(), &[1, 3]);
///
/// let c = a.i(0)?;
/// assert_eq!(c.shape().dims(), &[3]);
///
/// let c = a.i((.., ..2) )?;
/// assert_eq!(c.shape().dims(), &[2, 2]);
///
/// let c = a.i((.., ..=2))?;
/// assert_eq!(c.shape().dims(), &[2, 3]);
///
/// # Ok::<(), candle_core::Error>(())
/// ```
fn index(&self, indexers: &[TensorIndexer]) -> Result<Self, Error> {
let mut x = self.clone();
let dims = self.shape().dims();
let mut current_dim = 0;
for (i, indexer) in indexers.iter().enumerate() {
x = match indexer {
TensorIndexer::Select(n) => x.narrow(current_dim, *n, 1)?.squeeze(current_dim)?,
TensorIndexer::Narrow(left_bound, right_bound) => {
let start = match left_bound {
Bound::Included(n) => *n,
Bound::Excluded(n) => *n + 1,
Bound::Unbounded => 0,
};
let stop = match right_bound {
Bound::Included(n) => *n + 1,
Bound::Excluded(n) => *n,
Bound::Unbounded => dims[i],
};
let out = x.narrow(current_dim, start, stop.saturating_sub(start))?;
current_dim += 1;
out
}
TensorIndexer::IndexSelect(indexes) => {
if indexes.rank() != 1 {
crate::bail!("multi-dimensional tensor indexing is not supported")
}
let out = x.index_select(&indexes.to_device(x.device())?, current_dim)?;
current_dim += 1;
out
}
TensorIndexer::Err(e) => crate::bail!("indexing error {e:?}"),
};
}
Ok(x)
}
}
#[derive(Debug)]
/// Generic structure used to index a slice of the tensor
pub enum TensorIndexer {
/// This selects the elements for which an index has some specific value.
Select(usize),
/// This is a regular slice, purely indexing a chunk of the tensor
Narrow(Bound<usize>, Bound<usize>),
/// Indexing via a 1d tensor
IndexSelect(Tensor),
Err(Error),
}
impl From<usize> for TensorIndexer {
fn from(index: usize) -> Self {
TensorIndexer::Select(index)
}
}
impl From<&[u32]> for TensorIndexer {
fn from(index: &[u32]) -> Self {
match Tensor::new(index, &crate::Device::Cpu) {
Ok(tensor) => TensorIndexer::IndexSelect(tensor),
Err(e) => TensorIndexer::Err(e),
}
}
}
impl From<Vec<u32>> for TensorIndexer {
fn from(index: Vec<u32>) -> Self {
let len = index.len();
match Tensor::from_vec(index, len, &crate::Device::Cpu) {
Ok(tensor) => TensorIndexer::IndexSelect(tensor),
Err(e) => TensorIndexer::Err(e),
}
}
}
impl From<&Tensor> for TensorIndexer {
fn from(tensor: &Tensor) -> Self {
TensorIndexer::IndexSelect(tensor.clone())
}
}
trait RB: RangeBounds<usize> {}
impl RB for Range<usize> {}
impl RB for RangeFrom<usize> {}
impl RB for RangeFull {}
impl RB for RangeInclusive<usize> {}
impl RB for RangeTo<usize> {}
impl RB for RangeToInclusive<usize> {}
impl<T: RB> From<T> for TensorIndexer {
fn from(range: T) -> Self {
use std::ops::Bound::*;
let start = match range.start_bound() {
Included(idx) => Included(*idx),
Excluded(idx) => Excluded(*idx),
Unbounded => Unbounded,
};
let end = match range.end_bound() {
Included(idx) => Included(*idx),
Excluded(idx) => Excluded(*idx),
Unbounded => Unbounded,
};
TensorIndexer::Narrow(start, end)
}
}
/// Trait used to implement multiple signatures for ease of use of the slicing
/// of a tensor
pub trait IndexOp<T> {
/// Returns a slicing iterator which are the chunks of data necessary to
/// reconstruct the desired tensor.
fn i(&self, index: T) -> Result<Tensor, Error>;
}
impl<T> IndexOp<T> for Tensor
where
T: Into<TensorIndexer>,
{
///```rust
/// use candle_core::{Tensor, DType, Device, IndexOp};
/// let a = Tensor::new(&[
/// [0., 1.],
/// [2., 3.],
/// [4., 5.]
/// ], &Device::Cpu)?;
///
/// let b = a.i(0)?;
/// assert_eq!(b.shape().dims(), &[2]);
/// assert_eq!(b.to_vec1::<f64>()?, &[0., 1.]);
///
/// let c = a.i(..2)?;
/// assert_eq!(c.shape().dims(), &[2, 2]);
/// assert_eq!(c.to_vec2::<f64>()?, &[
/// [0., 1.],
/// [2., 3.]
/// ]);
///
/// let d = a.i(1..)?;
/// assert_eq!(d.shape().dims(), &[2, 2]);
/// assert_eq!(d.to_vec2::<f64>()?, &[
/// [2., 3.],
/// [4., 5.]
/// ]);
/// # Ok::<(), candle_core::Error>(())
/// ```
fn i(&self, index: T) -> Result<Tensor, Error> {
self.index(&[index.into()])
}
}
impl<A> IndexOp<(A,)> for Tensor
where
A: Into<TensorIndexer>,
{
///```rust
/// use candle_core::{Tensor, DType, Device, IndexOp};
/// let a = Tensor::new(&[
/// [0f32, 1.],
/// [2. , 3.],
/// [4. , 5.]
/// ], &Device::Cpu)?;
///
/// let b = a.i((0,))?;
/// assert_eq!(b.shape().dims(), &[2]);
/// assert_eq!(b.to_vec1::<f32>()?, &[0., 1.]);
///
/// let c = a.i((..2,))?;
/// assert_eq!(c.shape().dims(), &[2, 2]);
/// assert_eq!(c.to_vec2::<f32>()?, &[
/// [0., 1.],
/// [2., 3.]
/// ]);
///
/// let d = a.i((1..,))?;
/// assert_eq!(d.shape().dims(), &[2, 2]);
/// assert_eq!(d.to_vec2::<f32>()?, &[
/// [2., 3.],
/// [4., 5.]
/// ]);
/// # Ok::<(), candle_core::Error>(())
/// ```
fn i(&self, (a,): (A,)) -> Result<Tensor, Error> {
self.index(&[a.into()])
}
}
#[allow(non_snake_case)]
impl<A, B> IndexOp<(A, B)> for Tensor
where
A: Into<TensorIndexer>,
B: Into<TensorIndexer>,
{
///```rust
/// use candle_core::{Tensor, DType, Device, IndexOp};
/// let a = Tensor::new(&[[0f32, 1., 2.], [3., 4., 5.], [6., 7., 8.]], &Device::Cpu)?;
///
/// let b = a.i((1, 0))?;
/// assert_eq!(b.to_vec0::<f32>()?, 3.);
///
/// let c = a.i((..2, 1))?;
/// assert_eq!(c.shape().dims(), &[2]);
/// assert_eq!(c.to_vec1::<f32>()?, &[1., 4.]);
///
/// let d = a.i((2.., ..))?;
/// assert_eq!(c.shape().dims(), &[2]);
/// assert_eq!(c.to_vec1::<f32>()?, &[1., 4.]);
/// # Ok::<(), candle_core::Error>(())
/// ```
fn i(&self, (a, b): (A, B)) -> Result<Tensor, Error> {
self.index(&[a.into(), b.into()])
}
}
macro_rules! index_op_tuple {
($doc:tt, $($t:ident),+) => {
#[allow(non_snake_case)]
impl<$($t),*> IndexOp<($($t,)*)> for Tensor
where
$($t: Into<TensorIndexer>,)*
{
#[doc=$doc]
fn i(&self, ($($t,)*): ($($t,)*)) -> Result<Tensor, Error> {
self.index(&[$($t.into(),)*])
}
}
};
}
index_op_tuple!("see [TensorIndex#method.i]", A, B, C);
index_op_tuple!("see [TensorIndex#method.i]", A, B, C, D);
index_op_tuple!("see [TensorIndex#method.i]", A, B, C, D, E);
index_op_tuple!("see [TensorIndex#method.i]", A, B, C, D, E, F);
index_op_tuple!("see [TensorIndex#method.i]", A, B, C, D, E, F, G);
| candle/candle-core/src/indexer.rs/0 | {
"file_path": "candle/candle-core/src/indexer.rs",
"repo_id": "candle",
"token_count": 4032
} |
use super::{GgmlDType, QStorage};
use crate::backend::BackendStorage;
use crate::{DType, MetalDevice, MetalStorage, Result, Shape};
use metal::Buffer;
use std::sync::Arc;
pub struct QMetalStorage {
dtype: GgmlDType,
device: MetalDevice,
buffer: Arc<Buffer>,
}
impl QMetalStorage {
pub fn zeros(device: &MetalDevice, elem_count: usize, dtype: GgmlDType) -> Result<Self> {
let size = elem_count * dtype.type_size() / dtype.block_size();
let buffer = device.allocate_zeros(size)?;
Ok(Self {
buffer,
device: device.clone(),
dtype,
})
}
pub fn dtype(&self) -> GgmlDType {
self.dtype
}
pub fn device(&self) -> &MetalDevice {
&self.device
}
pub fn buffer(&self) -> &Buffer {
&self.buffer
}
pub fn dequantize(&self, elem_count: usize) -> Result<MetalStorage> {
use crate::quantized::k_quants::GgmlType;
let buffer = self.device.new_buffer_managed(self.buffer.length())?;
let command_buffer = self.device.command_buffer()?;
command_buffer.set_label("to_cpu");
let blit = command_buffer.new_blit_command_encoder();
blit.set_label("blit_to_cpu");
blit.copy_from_buffer(&self.buffer, 0, &buffer, 0, self.buffer.length());
blit.end_encoding();
self.device.wait_until_completed()?;
let mut out = vec![0.0; elem_count];
let block_len = elem_count / self.dtype.block_size();
match self.dtype {
GgmlDType::F32 => {
let vec: Vec<f32> = read_to_vec(&buffer, block_len);
f32::to_float(&vec, &mut out)?;
}
GgmlDType::F16 => {
let vec: Vec<half::f16> = read_to_vec(&buffer, block_len);
half::f16::to_float(&vec, &mut out)?;
}
GgmlDType::Q4_0 => {
let vec: Vec<crate::quantized::BlockQ4_0> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ4_0::to_float(&vec, &mut out)?;
}
GgmlDType::Q4_1 => {
let vec: Vec<crate::quantized::BlockQ4_1> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ4_1::to_float(&vec, &mut out)?;
}
GgmlDType::Q5_0 => {
let vec: Vec<crate::quantized::BlockQ5_0> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ5_0::to_float(&vec, &mut out)?;
}
GgmlDType::Q5_1 => {
let vec: Vec<crate::quantized::BlockQ5_1> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ5_1::to_float(&vec, &mut out)?;
}
GgmlDType::Q8_0 => {
let vec: Vec<crate::quantized::BlockQ8_0> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ8_0::to_float(&vec, &mut out)?;
}
GgmlDType::Q8_1 => {
let vec: Vec<crate::quantized::BlockQ8_1> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ8_1::to_float(&vec, &mut out)?;
}
GgmlDType::Q2K => {
let vec: Vec<crate::quantized::BlockQ2K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ2K::to_float(&vec, &mut out)?;
}
GgmlDType::Q3K => {
let vec: Vec<crate::quantized::BlockQ3K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ3K::to_float(&vec, &mut out)?;
}
GgmlDType::Q4K => {
let vec: Vec<crate::quantized::BlockQ4K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ4K::to_float(&vec, &mut out)?;
}
GgmlDType::Q5K => {
let vec: Vec<crate::quantized::BlockQ5K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ5K::to_float(&vec, &mut out)?;
}
GgmlDType::Q6K => {
let vec: Vec<crate::quantized::BlockQ6K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ6K::to_float(&vec, &mut out)?;
}
GgmlDType::Q8K => {
let vec: Vec<crate::quantized::BlockQ8K> = read_to_vec(&buffer, block_len);
crate::quantized::BlockQ8K::to_float(&vec, &mut out)?;
}
}
let buffer = self.device.new_buffer_with_data(&out)?;
Ok(MetalStorage::new(
buffer,
self.device.clone(),
elem_count,
DType::F32,
))
}
pub fn quantize(&mut self, src: &MetalStorage) -> Result<()> {
// Quantization only happens on CPU for now.
let src = src.to_cpu::<f32>()?;
let elem_count = src.len();
let src = crate::Storage::Cpu(crate::CpuStorage::F32(src));
let mut qcpu_storage = crate::Device::Cpu.qzeros(elem_count, self.dtype)?;
qcpu_storage.quantize(&src)?;
let buffer = self.device.new_buffer_with_data(&qcpu_storage.data()?)?;
self.buffer = buffer;
Ok(())
}
pub fn storage_size_in_bytes(&self) -> usize {
self.buffer.length() as usize
}
pub fn fwd(
&self,
self_shape: &Shape,
storage: &MetalStorage,
layout: &crate::Layout,
) -> Result<(MetalStorage, Shape)> {
use crate::MetalError;
if !layout.is_contiguous() {
crate::bail!("input tensor is not contiguous {layout:?}")
}
let src_shape = layout.shape();
// self is transposed so n is first then k.
if src_shape.rank() < 2 {
crate::bail!("input tensor has only one dimension {layout:?}")
}
let (n, k) = self_shape.dims2()?;
let mut dst_shape = src_shape.dims().to_vec();
// We always use a single batch dimension and stack all the tensors in the batch on the
// second dimension as the implementation in candle-metal-kernels doesn't handle batch
// properly.
let m = match dst_shape.len() {
3 => dst_shape[0] * dst_shape[1],
2 => dst_shape[0],
n => crate::bail!("Invalid rank {n} for quantized matmul metal"),
};
let last_k = dst_shape.pop().unwrap();
if last_k != k {
crate::bail!("input tensor {layout:?} incompatible with {:?}", self_shape)
}
dst_shape.push(n);
let dst_shape = Shape::from(dst_shape);
let device = storage.device().clone();
let dst = device.new_buffer(dst_shape.elem_count(), DType::F32, "qmatmul")?;
let command_buffer = device.command_buffer()?;
// In some cases it would be better to use the mm variant, though it has its drawbacks
// around memory alignemnt.
for batch_id in 0..m {
candle_metal_kernels::call_quantized_matmul_mv_t(
device.device(),
&command_buffer,
device.kernels(),
self.dtype.into(),
(1, 1, n, k),
storage.buffer(),
(layout.start_offset() + batch_id * k) * storage.dtype().size_in_bytes(),
&self.buffer,
batch_id * n * DType::F32.size_in_bytes(),
&dst,
)
.map_err(MetalError::from)?;
}
let dst_storage = crate::MetalStorage::new(dst, device, dst_shape.elem_count(), DType::F32);
Ok((dst_storage, dst_shape))
}
}
pub fn load_quantized<T: super::GgmlType + Send + Sync + 'static>(
device: &MetalDevice,
data: &[T],
) -> Result<QStorage> {
let buffer = device.new_buffer_with_data(data)?;
let device = device.clone();
Ok(QStorage::Metal(QMetalStorage {
dtype: T::DTYPE,
device,
buffer,
}))
}
fn read_to_vec<T: Clone>(buffer: &Buffer, n: usize) -> Vec<T> {
let ptr = buffer.contents() as *const T;
assert!(!ptr.is_null());
let slice = unsafe { std::slice::from_raw_parts(ptr, n) };
slice.to_vec()
}
impl From<GgmlDType> for candle_metal_kernels::GgmlDType {
fn from(value: GgmlDType) -> Self {
match value {
GgmlDType::Q4_0 => candle_metal_kernels::GgmlDType::Q4_0,
GgmlDType::Q4_1 => candle_metal_kernels::GgmlDType::Q4_1,
GgmlDType::Q5_0 => candle_metal_kernels::GgmlDType::Q5_0,
GgmlDType::Q5_1 => candle_metal_kernels::GgmlDType::Q5_1,
GgmlDType::Q8_0 => candle_metal_kernels::GgmlDType::Q8_0,
GgmlDType::Q8_1 => candle_metal_kernels::GgmlDType::Q8_1,
GgmlDType::Q2K => candle_metal_kernels::GgmlDType::Q2K,
GgmlDType::Q3K => candle_metal_kernels::GgmlDType::Q3K,
GgmlDType::Q4K => candle_metal_kernels::GgmlDType::Q4K,
GgmlDType::Q5K => candle_metal_kernels::GgmlDType::Q5K,
GgmlDType::Q6K => candle_metal_kernels::GgmlDType::Q6K,
GgmlDType::Q8K => candle_metal_kernels::GgmlDType::Q8K,
GgmlDType::F16 => candle_metal_kernels::GgmlDType::F16,
GgmlDType::F32 => candle_metal_kernels::GgmlDType::F32,
}
}
}
| candle/candle-core/src/quantized/metal.rs/0 | {
"file_path": "candle/candle-core/src/quantized/metal.rs",
"repo_id": "candle",
"token_count": 4704
} |
// Variables are wrappers around tensors that can be modified, they are typically used for holding
// weights and being modified by gradient descent.
// We do not expose a public way to create variables as this would break the invariant that the
// tensor within a variable is actually with `is_variable` set to `true`.
use crate::{DType, Device, Error, Result, Shape, Tensor};
/// A variable is a wrapper around a tensor, however variables can have their content modified
/// whereas tensors are immutable.
#[derive(Clone, Debug)]
pub struct Var(Tensor);
impl std::fmt::Display for Var {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
std::fmt::Display::fmt(&self.0, f)
}
}
impl std::ops::Deref for Var {
type Target = Tensor;
fn deref(&self) -> &Self::Target {
self.0.as_ref()
}
}
impl Var {
pub fn zeros<S: Into<Shape>>(shape: S, dtype: DType, device: &Device) -> Result<Self> {
let inner = Tensor::zeros_impl(shape, dtype, device, true)?;
Ok(Self(inner))
}
pub fn ones<S: Into<Shape>>(shape: S, dtype: DType, device: &Device) -> Result<Self> {
let inner = Tensor::ones_impl(shape, dtype, device, true)?;
Ok(Self(inner))
}
// Convert a tensor to a variable, if the tensor is already a variable then it is returned as is.
pub fn from_tensor(t: &Tensor) -> Result<Self> {
if t.is_variable() {
Ok(Self(t.clone()))
} else {
let inner = t.make_var()?;
Ok(Self(inner))
}
}
pub fn rand_f64<S: Into<Shape>>(
lo: f64,
up: f64,
s: S,
dtype: DType,
device: &Device,
) -> Result<Self> {
let inner = Tensor::rand_f64_impl(lo, up, s, dtype, device, true)?;
Ok(Self(inner))
}
pub fn randn_f64<S: Into<Shape>>(
mean: f64,
std: f64,
s: S,
dtype: DType,
device: &Device,
) -> Result<Self> {
let inner = Tensor::randn_f64_impl(mean, std, s, dtype, device, true)?;
Ok(Self(inner))
}
pub fn rand<S: Into<Shape>, T: crate::FloatDType>(
lo: T,
up: T,
s: S,
device: &Device,
) -> Result<Self> {
let inner = Tensor::rand_impl(lo, up, s, device, true)?;
Ok(Self(inner))
}
pub fn randn<S: Into<Shape>, T: crate::FloatDType>(
mean: T,
std: T,
s: S,
device: &Device,
) -> Result<Self> {
let inner = Tensor::randn_impl(mean, std, s, device, true)?;
Ok(Self(inner))
}
/// Creates a new tensor on the specified device using the content and shape of the input.
/// This is similar to `new` but the resulting tensor is a variable.
pub fn new<A: crate::device::NdArray>(array: A, device: &Device) -> Result<Self> {
let shape = array.shape()?;
let inner = Tensor::new_impl(array, shape, device, true)?;
Ok(Self(inner))
}
pub fn from_vec<S: Into<Shape>, D: crate::WithDType>(
data: Vec<D>,
shape: S,
device: &Device,
) -> Result<Self> {
let inner = Tensor::from_vec_impl(data, shape, device, true)?;
Ok(Self(inner))
}
pub fn from_slice<S: Into<Shape>, D: crate::WithDType>(
array: &[D],
shape: S,
device: &Device,
) -> Result<Self> {
let inner = Tensor::new_impl(array, shape.into(), device, true)?;
Ok(Self(inner))
}
pub fn as_detached_tensor(&self) -> Tensor {
self.0.detach()
}
pub fn as_tensor(&self) -> &Tensor {
&self.0
}
/// Consumes this `Var` and return the underlying tensor.
pub fn into_inner(self) -> Tensor {
self.0
}
/// Sets the content of the inner tensor, this does not require a mutable reference as inner
/// mutability is used.
pub fn set(&self, src: &Tensor) -> Result<()> {
if self.same_storage(src) {
let msg = "cannot set a variable to a tensor that is derived from its value";
Err(Error::CannotSetVar { msg }.bt())?
}
let (mut dst, layout) = self.storage_mut_and_layout();
if !layout.is_contiguous() {
let msg = "cannot set a non-contiguous variable";
Err(Error::CannotSetVar { msg }.bt())?
}
let (src, src_l) = src.storage_and_layout();
if layout.shape() != src_l.shape() {
Err(Error::ShapeMismatchBinaryOp {
lhs: layout.shape().clone(),
rhs: src_l.shape().clone(),
op: "set",
}
.bt())?
}
src.copy_strided_src(&mut dst, layout.start_offset(), src_l)?;
Ok(())
}
}
| candle/candle-core/src/variable.rs/0 | {
"file_path": "candle/candle-core/src/variable.rs",
"repo_id": "candle",
"token_count": 2150
} |
#![allow(unused)]
use anyhow::{Context, Result};
use std::io::Write;
use std::path::PathBuf;
struct KernelDirectories {
kernel_glob: &'static str,
rust_target: &'static str,
include_dirs: &'static [&'static str],
}
const KERNEL_DIRS: [KernelDirectories; 1] = [KernelDirectories {
kernel_glob: "examples/custom-ops/kernels/*.cu",
rust_target: "examples/custom-ops/cuda_kernels.rs",
include_dirs: &[],
}];
fn main() -> Result<()> {
println!("cargo:rerun-if-changed=build.rs");
#[cfg(feature = "cuda")]
{
for kdir in KERNEL_DIRS.iter() {
let builder = bindgen_cuda::Builder::default().kernel_paths_glob(kdir.kernel_glob);
println!("cargo:info={builder:?}");
let bindings = builder.build_ptx().unwrap();
bindings.write(kdir.rust_target).unwrap()
}
}
Ok(())
}
| candle/candle-examples/build.rs/0 | {
"file_path": "candle/candle-examples/build.rs",
"repo_id": "candle",
"token_count": 391
} |
use candle::{DType, Device, Tensor};
use candle_nn::VarBuilder;
use candle_transformers::generation::LogitsProcessor;
use candle_transformers::models::codegeex4_9b::*;
use clap::Parser;
use hf_hub::{Repo, RepoType};
use tokenizers::Tokenizer;
struct TextGeneration {
model: Model,
device: Device,
tokenizer: Tokenizer,
logits_processor: LogitsProcessor,
repeat_penalty: f32,
repeat_last_n: usize,
verbose: bool,
dtype: DType,
}
impl TextGeneration {
#[allow(clippy::too_many_arguments)]
fn new(
model: Model,
tokenizer: Tokenizer,
seed: u64,
temp: f64,
top_p: f64,
repeat_penalty: f32,
repeat_last_n: usize,
verbose: bool,
device: &Device,
dtype: DType,
) -> Self {
let logits_processor = LogitsProcessor::new(seed, Some(temp), Some(top_p));
Self {
model,
tokenizer,
logits_processor,
repeat_penalty,
repeat_last_n,
verbose,
device: device.clone(),
dtype,
}
}
fn run(&mut self, prompt: &str, sample_len: usize) -> anyhow::Result<()> {
use std::io::Write;
println!("starting the inference loop");
let tokens = self.tokenizer.encode(prompt, true).expect("tokens error");
if tokens.is_empty() {
panic!("Empty prompts are not supported in the chatglm model.")
}
if self.verbose {
for (token, id) in tokens.get_tokens().iter().zip(tokens.get_ids().iter()) {
let token = token.replace('▁', " ").replace("<0x0A>", "\n");
println!("{id:7} -> '{token}'");
}
}
let eos_token = match self.tokenizer.get_vocab(true).get("<|endoftext|>") {
Some(token) => *token,
None => panic!("cannot find the endoftext token"),
};
let mut tokens = tokens.get_ids().to_vec();
let mut generated_tokens = 0usize;
print!("{prompt}");
std::io::stdout().flush().expect("output flush error");
let start_gen = std::time::Instant::now();
println!("\n start_gen");
println!("samplelen {}", sample_len);
let mut count = 0;
let mut result = vec![];
for index in 0..sample_len {
count += 1;
let context_size = if index > 0 { 1 } else { tokens.len() };
let ctxt = &tokens[tokens.len().saturating_sub(context_size)..];
let input = Tensor::new(ctxt, &self.device)?.unsqueeze(0)?;
let logits = self.model.forward(&input)?;
let logits = logits.squeeze(0)?.to_dtype(self.dtype)?;
let logits = if self.repeat_penalty == 1. {
logits
} else {
let start_at = tokens.len().saturating_sub(self.repeat_last_n);
candle_transformers::utils::apply_repeat_penalty(
&logits,
self.repeat_penalty,
&tokens[start_at..],
)?
};
let next_token = self.logits_processor.sample(&logits)?;
tokens.push(next_token);
generated_tokens += 1;
if next_token == eos_token {
break;
}
let token = self
.tokenizer
.decode(&[next_token], true)
.expect("Token error");
if self.verbose {
println!(
"[Count: {}] [Raw Token: {}] [Decode Token: {}]",
count, next_token, token
);
}
result.push(token);
std::io::stdout().flush()?;
}
let dt = start_gen.elapsed();
println!(
"\n{generated_tokens} tokens generated ({:.2} token/s)",
generated_tokens as f64 / dt.as_secs_f64(),
);
println!("Result:");
for tokens in result {
print!("{tokens}");
}
Ok(())
}
}
#[derive(Parser, Debug)]
#[command(author, version, about, long_about = None)]
struct Args {
#[arg(name = "cache", short)]
cache_path: Option<String>,
/// Run on CPU rather than on GPU.
#[arg(long)]
cpu: bool,
/// Display the token for the specified prompt.
#[arg(long)]
prompt: String,
/// Display the tokens for the specified prompt and outputs.
#[arg(long)]
verbose: bool,
/// The temperature used to generate samples.
#[arg(long, default_value_t = 0.95)]
temperature: f64,
/// Nucleus sampling probability cutoff.
#[arg(long, default_value_t = 0.8)]
top_p: f64,
/// The seed to use when generating random samples.
#[arg(long, default_value_t = 299792458)]
seed: u64,
/// The length of the sample to generate (in tokens).
#[arg(long, short = 'n', default_value_t = 8192)]
sample_len: usize,
#[arg(long)]
model_id: Option<String>,
#[arg(long)]
revision: Option<String>,
#[arg(long)]
weight_path: Option<String>,
#[arg(long)]
tokenizer: Option<String>,
/// Penalty to be applied for repeating tokens, 1. means no penalty.
#[arg(long, default_value_t = 1.2)]
repeat_penalty: f32,
/// The context size to consider for the repeat penalty.
#[arg(long, default_value_t = 64)]
repeat_last_n: usize,
}
fn main() -> anyhow::Result<()> {
let args = Args::parse();
println!(
"avx: {}, neon: {}, simd128: {}, f16c: {}",
candle::utils::with_avx(),
candle::utils::with_neon(),
candle::utils::with_simd128(),
candle::utils::with_f16c()
);
println!(
"temp: {:.2} repeat-penalty: {:.2} repeat-last-n: {}",
args.temperature, args.repeat_penalty, args.repeat_last_n
);
let start = std::time::Instant::now();
let api = match args.cache_path.as_ref() {
None => hf_hub::api::sync::Api::new()?,
Some(path) => {
hf_hub::api::sync::ApiBuilder::from_cache(hf_hub::Cache::new(path.to_string().into()))
.build()
.map_err(anyhow::Error::msg)?
}
};
let model_id = match args.model_id {
Some(model_id) => model_id.to_string(),
None => "THUDM/codegeex4-all-9b".to_string(),
};
let revision = match args.revision {
Some(rev) => rev.to_string(),
None => "main".to_string(),
};
let repo = api.repo(Repo::with_revision(model_id, RepoType::Model, revision));
let tokenizer_filename = match args.tokenizer {
Some(file) => std::path::PathBuf::from(file),
None => api
.model("THUDM/codegeex4-all-9b".to_string())
.get("tokenizer.json")
.map_err(anyhow::Error::msg)?,
};
let config_filename = match &args.weight_path {
Some(path) => std::path::Path::new(path).join("config.json"),
None => repo.get("config.json")?,
};
let filenames = match &args.weight_path {
Some(path) => {
candle_examples::hub_load_local_safetensors(path, "model.safetensors.index.json")?
}
_ => candle_examples::hub_load_safetensors(&repo, "model.safetensors.index.json")?,
};
println!("retrieved the files in {:?}", start.elapsed());
let tokenizer = Tokenizer::from_file(tokenizer_filename).expect("Tokenizer Error");
let start = std::time::Instant::now();
let config: Config = serde_json::from_slice(&std::fs::read(config_filename)?)?;
let device = candle_examples::device(args.cpu)?;
let dtype = if device.is_cuda() {
DType::BF16
} else {
DType::F32
};
let vb = unsafe { VarBuilder::from_mmaped_safetensors(&filenames, dtype, &device)? };
let model = Model::new(&config, vb)?;
println!("loaded the model in {:?}", start.elapsed());
let mut pipeline = TextGeneration::new(
model,
tokenizer,
args.seed,
args.temperature,
args.top_p,
args.repeat_penalty,
args.repeat_last_n,
args.verbose,
&device,
dtype,
);
pipeline.run(&args.prompt, args.sample_len)?;
Ok(())
}
| candle/candle-examples/examples/codegeex4-9b/main.rs/0 | {
"file_path": "candle/candle-examples/examples/codegeex4-9b/main.rs",
"repo_id": "candle",
"token_count": 3921
} |
//! DINOv2: Learning Robust Visual Features without Supervision
//! https://github.com/facebookresearch/dinov2
#[cfg(feature = "mkl")]
extern crate intel_mkl_src;
#[cfg(feature = "accelerate")]
extern crate accelerate_src;
use clap::Parser;
use candle::{DType, IndexOp, D};
use candle_nn::{Module, VarBuilder};
use candle_transformers::models::dinov2;
#[derive(Parser)]
struct Args {
#[arg(long)]
model: Option<String>,
#[arg(long)]
image: String,
/// Run on CPU rather than on GPU.
#[arg(long)]
cpu: bool,
}
pub fn main() -> anyhow::Result<()> {
let args = Args::parse();
let device = candle_examples::device(args.cpu)?;
let image = candle_examples::imagenet::load_image224(args.image)?.to_device(&device)?;
println!("loaded image {image:?}");
let model_file = match args.model {
None => {
let api = hf_hub::api::sync::Api::new()?;
let api = api.model("lmz/candle-dino-v2".into());
api.get("dinov2_vits14.safetensors")?
}
Some(model) => model.into(),
};
let vb = unsafe { VarBuilder::from_mmaped_safetensors(&[model_file], DType::F32, &device)? };
let model = dinov2::vit_small(vb)?;
println!("model built");
let logits = model.forward(&image.unsqueeze(0)?)?;
let prs = candle_nn::ops::softmax(&logits, D::Minus1)?
.i(0)?
.to_vec1::<f32>()?;
let mut prs = prs.iter().enumerate().collect::<Vec<_>>();
prs.sort_by(|(_, p1), (_, p2)| p2.total_cmp(p1));
for &(category_idx, pr) in prs.iter().take(5) {
println!(
"{:24}: {:.2}%",
candle_examples::imagenet::CLASSES[category_idx],
100. * pr
);
}
Ok(())
}
| candle/candle-examples/examples/dinov2/main.rs/0 | {
"file_path": "candle/candle-examples/examples/dinov2/main.rs",
"repo_id": "candle",
"token_count": 791
} |
# candle-fastvit
[FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization](https://arxiv.org/abs/2303.14189).
This candle implementation uses a pre-trained FastViT network for inference. The
classification head has been trained on the ImageNet dataset and returns the
probabilities for the top-5 classes.
## Running an example
```
$ cargo run --example fastvit --release -- --image candle-examples/examples/yolo-v8/assets/bike.jpg --which sa12
loaded image Tensor[dims 3, 256, 256; f32]
model built
mountain bike, all-terrain bike, off-roader: 52.67%
bicycle-built-for-two, tandem bicycle, tandem: 7.93%
unicycle, monocycle : 3.46%
maillot : 1.32%
crash helmet : 1.28%
```
| candle/candle-examples/examples/fastvit/README.md/0 | {
"file_path": "candle/candle-examples/examples/fastvit/README.md",
"repo_id": "candle",
"token_count": 258
} |
# hiera
[Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles](https://arxiv.org/abs/2306.00989)
This candle implementation uses pre-trained Hiera models from timm for inference.
The classification head has been trained on the ImageNet dataset and returns the probabilities for the top-5 classes.
## Running an example
```
$ cargo run --example hiera --release -- --image candle-examples/examples/yolo-v8/assets/bike.jpg --which tiny
loaded image Tensor[dims 3, 224, 224; f32]
model built
mountain bike, all-terrain bike, off-roader: 71.15%
unicycle, monocycle : 7.11%
knee pad : 4.26%
crash helmet : 1.48%
moped : 1.07%
```
| candle/candle-examples/examples/hiera/README.md/0 | {
"file_path": "candle/candle-examples/examples/hiera/README.md",
"repo_id": "candle",
"token_count": 260
} |
/// This follows the lines of:
/// https://github.com/johnma2006/mamba-minimal/blob/master/model.py
/// Simple, minimal implementation of Mamba in one file of PyTorch.
use candle::{IndexOp, Module, Result, Tensor, D};
use candle_nn::{RmsNorm, VarBuilder};
use candle_transformers::models::with_tracing::{linear, linear_no_bias, Linear};
#[derive(Debug, Clone, serde::Deserialize)]
pub struct Config {
d_model: usize,
n_layer: usize,
vocab_size: usize,
pad_vocab_size_multiple: usize,
}
impl Config {
fn vocab_size(&self) -> usize {
let pad = self.pad_vocab_size_multiple;
self.vocab_size.div_ceil(pad) * pad
}
fn dt_rank(&self) -> usize {
(self.d_model + 15) / 16
}
fn d_conv(&self) -> usize {
4
}
fn d_state(&self) -> usize {
16
}
fn d_inner(&self) -> usize {
self.d_model * 2
}
}
// https://github.com/johnma2006/mamba-minimal/blob/61f01953ca153f8c4a850d7111beecbf4be9cee1/model.py#L177
#[derive(Clone, Debug)]
pub struct MambaBlock {
in_proj: Linear,
conv1d: candle_nn::Conv1d,
x_proj: Linear,
dt_proj: Linear,
a_log: Tensor,
d: Tensor,
out_proj: Linear,
dt_rank: usize,
}
impl MambaBlock {
pub fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let d_inner = cfg.d_inner();
let d_conv = cfg.d_conv();
let d_state = cfg.d_state();
let dt_rank = cfg.dt_rank();
let in_proj = linear_no_bias(cfg.d_model, d_inner * 2, vb.pp("in_proj"))?;
let conv_cfg = candle_nn::Conv1dConfig {
groups: d_inner,
padding: d_conv - 1,
..Default::default()
};
let conv1d = candle_nn::conv1d(d_inner, d_inner, d_conv, conv_cfg, vb.pp("conv1d"))?;
let x_proj = linear_no_bias(d_inner, dt_rank + d_state * 2, vb.pp("x_proj"))?;
let dt_proj = linear(dt_rank, d_inner, vb.pp("dt_proj"))?;
let a_log = vb.get((d_inner, d_state), "A_log")?;
let d = vb.get(d_inner, "D")?;
let out_proj = linear_no_bias(d_inner, cfg.d_model, vb.pp("out_proj"))?;
Ok(Self {
in_proj,
conv1d,
x_proj,
dt_proj,
a_log,
d,
out_proj,
dt_rank,
})
}
fn ssm(&self, xs: &Tensor) -> Result<Tensor> {
let (_d_in, n) = self.a_log.dims2()?;
let a = self.a_log.to_dtype(candle::DType::F32)?.exp()?.neg()?;
let d = self.d.to_dtype(candle::DType::F32)?;
let x_dbl = xs.apply(&self.x_proj)?;
let delta = x_dbl.narrow(D::Minus1, 0, self.dt_rank)?;
let b = x_dbl.narrow(D::Minus1, self.dt_rank, n)?;
let c = x_dbl.narrow(D::Minus1, self.dt_rank + n, n)?;
let delta = delta.contiguous()?.apply(&self.dt_proj)?;
// softplus without threshold
let delta = (delta.exp()? + 1.)?.log()?;
let ss = selective_scan(xs, &delta, &a, &b, &c, &d)?;
Ok(ss)
}
}
// https://github.com/johnma2006/mamba-minimal/blob/61f01953ca153f8c4a850d7111beecbf4be9cee1/model.py#L275
fn selective_scan(
u: &Tensor,
delta: &Tensor,
a: &Tensor,
b: &Tensor,
c: &Tensor,
d: &Tensor,
) -> Result<Tensor> {
let (b_sz, l, d_in) = u.dims3()?;
let n = a.dim(1)?;
let delta = delta.t()?.reshape((b_sz, d_in, l, 1))?; // b d_in l 1
let delta_a = delta.broadcast_mul(&a.reshape((1, d_in, 1, n))?)?.exp()?;
let delta_b_u = delta
.broadcast_mul(&b.reshape((b_sz, 1, l, n))?)?
.broadcast_mul(&u.t()?.reshape((b_sz, d_in, l, 1))?)?;
let mut xs = Tensor::zeros((b_sz, d_in, n), delta_a.dtype(), delta_a.device())?;
let mut ys = Vec::with_capacity(l);
for i in 0..l {
xs = ((delta_a.i((.., .., i))? * xs)? + delta_b_u.i((.., .., i))?)?;
let y = xs.matmul(&c.i((.., i, ..))?.unsqueeze(2)?)?.squeeze(2)?;
ys.push(y)
}
let ys = Tensor::stack(ys.as_slice(), 1)?;
ys + u.broadcast_mul(d)
}
impl Module for MambaBlock {
// https://github.com/johnma2006/mamba-minimal/blob/61f01953ca153f8c4a850d7111beecbf4be9cee1/model.py#L206
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
let (_b_sz, seq_len, _dim) = xs.dims3()?;
let xs_and_res = xs.apply(&self.in_proj)?.chunk(2, D::Minus1)?;
let (xs, res) = (&xs_and_res[0], &xs_and_res[1]);
let xs = xs
.t()?
.apply(&self.conv1d)?
.narrow(D::Minus1, 0, seq_len)?
.t()?;
let xs = candle_nn::ops::silu(&xs)?;
let ys = (self.ssm(&xs)? * candle_nn::ops::silu(res))?;
ys.apply(&self.out_proj)
}
}
// https://github.com/johnma2006/mamba-minimal/blob/61f01953ca153f8c4a850d7111beecbf4be9cee1/model.py#L143
#[derive(Clone, Debug)]
pub struct ResidualBlock {
mixer: MambaBlock,
norm: RmsNorm,
}
impl ResidualBlock {
pub fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let norm = candle_nn::rms_norm(cfg.d_model, 1e-5, vb.pp("norm"))?;
let mixer = MambaBlock::new(cfg, vb.pp("mixer"))?;
Ok(Self { mixer, norm })
}
}
impl Module for ResidualBlock {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
xs.apply(&self.norm)?.apply(&self.mixer)? + xs
}
}
// https://github.com/johnma2006/mamba-minimal/blob/61f01953ca153f8c4a850d7111beecbf4be9cee1/model.py#L56
#[derive(Clone, Debug)]
pub struct Model {
embedding: candle_nn::Embedding,
layers: Vec<ResidualBlock>,
norm_f: RmsNorm,
lm_head: Linear,
}
impl Model {
pub fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let embedding = candle_nn::embedding(cfg.vocab_size(), cfg.d_model, vb.pp("embedding"))?;
let mut layers = Vec::with_capacity(cfg.n_layer);
let vb_l = vb.pp("layers");
for layer_idx in 0..cfg.n_layer {
let layer = ResidualBlock::new(cfg, vb_l.pp(layer_idx))?;
layers.push(layer)
}
let norm_f = candle_nn::rms_norm(cfg.d_model, 1e-5, vb.pp("norm_f"))?;
let lm_head = Linear::from_weights(embedding.embeddings().clone(), None);
Ok(Self {
embedding,
layers,
norm_f,
lm_head,
})
}
}
impl Module for Model {
fn forward(&self, input_ids: &Tensor) -> Result<Tensor> {
let (_b_size, seq_len) = input_ids.dims2()?;
let mut xs = self.embedding.forward(input_ids)?;
for layer in self.layers.iter() {
xs = layer.forward(&xs)?
}
xs.narrow(1, seq_len - 1, 1)?
.apply(&self.norm_f)?
.apply(&self.lm_head)
}
}
| candle/candle-examples/examples/mamba-minimal/model.rs/0 | {
"file_path": "candle/candle-examples/examples/mamba-minimal/model.rs",
"repo_id": "candle",
"token_count": 3488
} |
# candle-mobileclip
MobileCLIP is family of efficient CLIP-like models using FastViT-based image encoders.
See [MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training](https://arxiv.org/abs/2311.17049)
## Running on an example on cpu
```
$ cargo run --example mobileclip --release -- --images "candle-examples/examples/stable-diffusion/assets/stable-diffusion-xl.jpg","candle-examples/examples/yolo-v8/assets/bike.jpg" --cpu --sequences "a cycling race","a photo of two cats","a robot holding a candle"
softmax_image_vec: [2.4819004e-5, 3.81081e-6, 0.9999714, 0.9999738, 2.382714e-5, 2.3317718e-6]
Results for image: candle-examples/examples/stable-diffusion/assets/stable-diffusion-xl.jpg
Probability: 0.0025% Text: a cycling race
Probability: 0.0004% Text: a photo of two cats
Probability: 99.9971% Text: a robot holding a candle
Results for image: candle-examples/examples/yolo-v8/assets/bike.jpg
Probability: 99.9974% Text: a cycling race
Probability: 0.0024% Text: a photo of two cats
Probability: 0.0002% Text: a robot holding a candle
```
| candle/candle-examples/examples/mobileclip/README.md/0 | {
"file_path": "candle/candle-examples/examples/mobileclip/README.md",
"repo_id": "candle",
"token_count": 379
} |
## Using ONNX models in Candle
This example demonstrates how to run [ONNX](https://github.com/onnx/onnx) based models in Candle.
It contains small variants of two models, [SqueezeNet](https://arxiv.org/pdf/1602.07360.pdf) (default) and [EfficientNet](https://arxiv.org/pdf/1905.11946.pdf).
You can run the examples with following commands:
```bash
cargo run --example onnx --features=onnx --release -- --image candle-examples/examples/yolo-v8/assets/bike.jpg
```
Use the `--which` flag to specify explicitly which network to use, i.e.
```bash
$ cargo run --example onnx --features=onnx --release -- --which squeeze-net --image candle-examples/examples/yolo-v8/assets/bike.jpg
Finished release [optimized] target(s) in 0.21s
Running `target/release/examples/onnx --which squeeze-net --image candle-examples/examples/yolo-v8/assets/bike.jpg`
loaded image Tensor[dims 3, 224, 224; f32]
unicycle, monocycle : 83.23%
ballplayer, baseball player : 3.68%
bearskin, busby, shako : 1.54%
military uniform : 0.78%
cowboy hat, ten-gallon hat : 0.76%
```
```bash
$ cargo run --example onnx --features=onnx --release -- --which efficient-net --image candle-examples/examples/yolo-v8/assets/bike.jpg
Finished release [optimized] target(s) in 0.20s
Running `target/release/examples/onnx --which efficient-net --image candle-examples/examples/yolo-v8/assets/bike.jpg`
loaded image Tensor[dims 224, 224, 3; f32]
bicycle-built-for-two, tandem bicycle, tandem : 99.16%
mountain bike, all-terrain bike, off-roader : 0.60%
unicycle, monocycle : 0.17%
crash helmet : 0.02%
alp : 0.02%
```
| candle/candle-examples/examples/onnx/README.md/0 | {
"file_path": "candle/candle-examples/examples/onnx/README.md",
"repo_id": "candle",
"token_count": 832
} |
#[cfg(feature = "mkl")]
extern crate intel_mkl_src;
#[cfg(feature = "accelerate")]
extern crate accelerate_src;
use std::io::Write;
use std::path::PathBuf;
use candle_transformers::models::quantized_t5 as t5;
use anyhow::{Error as E, Result};
use candle::{Device, Tensor};
use candle_transformers::generation::LogitsProcessor;
use clap::{Parser, ValueEnum};
use hf_hub::{api::sync::Api, api::sync::ApiRepo, Repo, RepoType};
use tokenizers::Tokenizer;
#[derive(Clone, Debug, Copy, ValueEnum)]
enum Which {
T5Small,
FlanT5Small,
FlanT5Base,
FlanT5Large,
FlanT5Xl,
FlanT5Xxl,
}
#[derive(Parser, Debug, Clone)]
#[command(author, version, about, long_about = None)]
struct Args {
/// Enable tracing (generates a trace-timestamp.json file).
#[arg(long)]
tracing: bool,
/// The model repository to use on the HuggingFace hub.
#[arg(long)]
model_id: Option<String>,
#[arg(long)]
revision: Option<String>,
#[arg(long)]
weight_file: Option<String>,
#[arg(long)]
config_file: Option<String>,
// Enable/disable decoding.
#[arg(long, default_value = "false")]
disable_cache: bool,
/// Use this prompt, otherwise compute sentence similarities.
#[arg(long)]
prompt: String,
/// The temperature used to generate samples.
#[arg(long, default_value_t = 0.8)]
temperature: f64,
/// Nucleus sampling probability cutoff.
#[arg(long)]
top_p: Option<f64>,
/// Penalty to be applied for repeating tokens, 1. means no penalty.
#[arg(long, default_value_t = 1.1)]
repeat_penalty: f32,
/// The context size to consider for the repeat penalty.
#[arg(long, default_value_t = 64)]
repeat_last_n: usize,
/// The model size to use.
#[arg(long, default_value = "t5-small")]
which: Which,
}
struct T5ModelBuilder {
device: Device,
config: t5::Config,
weights_filename: PathBuf,
}
impl T5ModelBuilder {
pub fn load(args: &Args) -> Result<(Self, Tokenizer)> {
let device = Device::Cpu;
let default_model = "lmz/candle-quantized-t5".to_string();
let (model_id, revision) = match (args.model_id.to_owned(), args.revision.to_owned()) {
(Some(model_id), Some(revision)) => (model_id, revision),
(Some(model_id), None) => (model_id, "main".to_string()),
(None, Some(revision)) => (default_model, revision),
(None, None) => (default_model, "main".to_string()),
};
let repo = Repo::with_revision(model_id, RepoType::Model, revision);
let api = Api::new()?;
let api = api.repo(repo);
let config_filename = match &args.config_file {
Some(filename) => Self::get_local_or_remote_file(filename, &api)?,
None => match args.which {
Which::T5Small => api.get("config.json")?,
Which::FlanT5Small => api.get("config-flan-t5-small.json")?,
Which::FlanT5Base => api.get("config-flan-t5-base.json")?,
Which::FlanT5Large => api.get("config-flan-t5-large.json")?,
Which::FlanT5Xl => api.get("config-flan-t5-xl.json")?,
Which::FlanT5Xxl => api.get("config-flan-t5-xxl.json")?,
},
};
let tokenizer_filename = api.get("tokenizer.json")?;
let weights_filename = match &args.weight_file {
Some(filename) => Self::get_local_or_remote_file(filename, &api)?,
None => match args.which {
Which::T5Small => api.get("model.gguf")?,
Which::FlanT5Small => api.get("model-flan-t5-small.gguf")?,
Which::FlanT5Base => api.get("model-flan-t5-base.gguf")?,
Which::FlanT5Large => api.get("model-flan-t5-large.gguf")?,
Which::FlanT5Xl => api.get("model-flan-t5-xl.gguf")?,
Which::FlanT5Xxl => api.get("model-flan-t5-xxl.gguf")?,
},
};
let config = std::fs::read_to_string(config_filename)?;
let mut config: t5::Config = serde_json::from_str(&config)?;
config.use_cache = !args.disable_cache;
let tokenizer = Tokenizer::from_file(tokenizer_filename).map_err(E::msg)?;
Ok((
Self {
device,
config,
weights_filename,
},
tokenizer,
))
}
pub fn build_model(&self) -> Result<t5::T5ForConditionalGeneration> {
let device = Device::Cpu;
let vb = t5::VarBuilder::from_gguf(&self.weights_filename, &device)?;
Ok(t5::T5ForConditionalGeneration::load(vb, &self.config)?)
}
fn get_local_or_remote_file(filename: &str, api: &ApiRepo) -> Result<PathBuf> {
let local_filename = std::path::PathBuf::from(filename);
if local_filename.exists() {
Ok(local_filename)
} else {
Ok(api.get(filename)?)
}
}
}
fn main() -> Result<()> {
use tracing_chrome::ChromeLayerBuilder;
use tracing_subscriber::prelude::*;
let args = Args::parse();
let _guard = if args.tracing {
let (chrome_layer, guard) = ChromeLayerBuilder::new().build();
tracing_subscriber::registry().with(chrome_layer).init();
Some(guard)
} else {
None
};
let (builder, mut tokenizer) = T5ModelBuilder::load(&args)?;
let device = &builder.device;
let tokenizer = tokenizer
.with_padding(None)
.with_truncation(None)
.map_err(E::msg)?;
let tokens = tokenizer
.encode(args.prompt, true)
.map_err(E::msg)?
.get_ids()
.to_vec();
let input_token_ids = Tensor::new(&tokens[..], device)?.unsqueeze(0)?;
let mut model = builder.build_model()?;
let mut output_token_ids = [builder
.config
.decoder_start_token_id
.unwrap_or(builder.config.pad_token_id) as u32]
.to_vec();
let temperature = if args.temperature <= 0. {
None
} else {
Some(args.temperature)
};
let mut logits_processor = LogitsProcessor::new(299792458, temperature, args.top_p);
let encoder_output = model.encode(&input_token_ids)?;
let start = std::time::Instant::now();
for index in 0.. {
if output_token_ids.len() > 512 {
break;
}
let decoder_token_ids = if index == 0 || !builder.config.use_cache {
Tensor::new(output_token_ids.as_slice(), device)?.unsqueeze(0)?
} else {
let last_token = *output_token_ids.last().unwrap();
Tensor::new(&[last_token], device)?.unsqueeze(0)?
};
let logits = model
.decode(&decoder_token_ids, &encoder_output)?
.squeeze(0)?;
let logits = if args.repeat_penalty == 1. {
logits
} else {
let start_at = output_token_ids.len().saturating_sub(args.repeat_last_n);
candle_transformers::utils::apply_repeat_penalty(
&logits,
args.repeat_penalty,
&output_token_ids[start_at..],
)?
};
let next_token_id = logits_processor.sample(&logits)?;
if next_token_id as usize == builder.config.eos_token_id {
break;
}
output_token_ids.push(next_token_id);
if let Some(text) = tokenizer.id_to_token(next_token_id) {
let text = text.replace('▁', " ").replace("<0x0A>", "\n");
print!("{text}");
std::io::stdout().flush()?;
}
}
let dt = start.elapsed();
println!(
"\n{} tokens generated ({:.2} token/s)\n",
output_token_ids.len(),
output_token_ids.len() as f64 / dt.as_secs_f64(),
);
Ok(())
}
| candle/candle-examples/examples/quantized-t5/main.rs/0 | {
"file_path": "candle/candle-examples/examples/quantized-t5/main.rs",
"repo_id": "candle",
"token_count": 3631
} |
# candle-replit-code: code completion specialized model.
[replit-code-v1_5-3b](https://huggingface.co/replit/replit-code-v1_5-3b) is a
language model specialized for code completion. This model uses 3.3B parameters
in `bfloat16` (so the GPU version will only work on recent nvidia cards).
## Running some example
```bash
cargo run --example replit-code --release -- --prompt 'def fibonacci(n): '
```
This produces the following output.
```
def fibonacci(n): # write Fibonacci series up to n
"""Print a Fibonacci series up to n."""
a, b = 0, 1
while a < n:
print(a, end=' ')
a, b = b, a+b
print()
def fibonacci_loop(n): # write Fibonacci series up to n
"""Print a Fibonacci series up to n."""
result = []
a, b = 0, 1
while a < n:
result.append(a)
a, b = b, a+b
return result
def fibonacci_generator(n): # write Fibonacci series up to n
"""Print a Fibonacci series up to n."""
a, b = 0, 1
while a < n:
yield a
a, b = b, a+b
```
| candle/candle-examples/examples/replit-code/README.md/0 | {
"file_path": "candle/candle-examples/examples/replit-code/README.md",
"repo_id": "candle",
"token_count": 426
} |
//! SAM: Segment Anything Model
//! https://github.com/facebookresearch/segment-anything
#[cfg(feature = "mkl")]
extern crate intel_mkl_src;
#[cfg(feature = "accelerate")]
extern crate accelerate_src;
use candle::DType;
use candle_nn::VarBuilder;
use candle_transformers::models::segment_anything::sam;
use clap::Parser;
#[derive(Parser)]
struct Args {
#[arg(long)]
model: Option<String>,
#[arg(long)]
image: String,
/// Run on CPU rather than on GPU.
#[arg(long)]
cpu: bool,
#[arg(long)]
generate_masks: bool,
/// List of x,y coordinates, between 0 and 1 (0.5 is at the middle of the image). These points
/// should be part of the generated mask.
#[arg(long)]
point: Vec<String>,
/// List of x,y coordinates, between 0 and 1 (0.5 is at the middle of the image). These points
/// should not be part of the generated mask and should be part of the background instead.
#[arg(long)]
neg_point: Vec<String>,
/// The detection threshold for the mask, 0 is the default value, negative values mean a larger
/// mask, positive makes the mask more selective.
#[arg(long, allow_hyphen_values = true, default_value_t = 0.)]
threshold: f32,
/// Enable tracing (generates a trace-timestamp.json file).
#[arg(long)]
tracing: bool,
/// Use the TinyViT based models from MobileSAM
#[arg(long)]
use_tiny: bool,
}
pub fn main() -> anyhow::Result<()> {
use tracing_chrome::ChromeLayerBuilder;
use tracing_subscriber::prelude::*;
let args = Args::parse();
let _guard = if args.tracing {
let (chrome_layer, guard) = ChromeLayerBuilder::new().build();
tracing_subscriber::registry().with(chrome_layer).init();
Some(guard)
} else {
None
};
let device = candle_examples::device(args.cpu)?;
let (image, initial_h, initial_w) =
candle_examples::load_image(&args.image, Some(sam::IMAGE_SIZE))?;
let image = image.to_device(&device)?;
println!("loaded image {image:?}");
let model = match args.model {
Some(model) => std::path::PathBuf::from(model),
None => {
let api = hf_hub::api::sync::Api::new()?;
let api = api.model("lmz/candle-sam".to_string());
let filename = if args.use_tiny {
"mobile_sam-tiny-vitt.safetensors"
} else {
"sam_vit_b_01ec64.safetensors"
};
api.get(filename)?
}
};
let vb = unsafe { VarBuilder::from_mmaped_safetensors(&[model], DType::F32, &device)? };
let sam = if args.use_tiny {
sam::Sam::new_tiny(vb)? // tiny vit_t
} else {
sam::Sam::new(768, 12, 12, &[2, 5, 8, 11], vb)? // sam_vit_b
};
if args.generate_masks {
// Default options similar to the Python version.
let bboxes = sam.generate_masks(
&image,
/* points_per_side */ 32,
/* crop_n_layer */ 0,
/* crop_overlap_ratio */ 512. / 1500.,
/* crop_n_points_downscale_factor */ 1,
)?;
for (idx, bbox) in bboxes.iter().enumerate() {
println!("{idx} {bbox:?}");
let mask = (&bbox.data.to_dtype(DType::U8)? * 255.)?;
let (h, w) = mask.dims2()?;
let mask = mask.broadcast_as((3, h, w))?;
candle_examples::save_image_resize(
&mask,
format!("sam_mask{idx}.png"),
initial_h,
initial_w,
)?;
}
} else {
let iter_points = args.point.iter().map(|p| (p, true));
let iter_neg_points = args.neg_point.iter().map(|p| (p, false));
let points = iter_points
.chain(iter_neg_points)
.map(|(point, b)| {
use std::str::FromStr;
let xy = point.split(',').collect::<Vec<_>>();
if xy.len() != 2 {
anyhow::bail!("expected format for points is 0.4,0.2")
}
Ok((f64::from_str(xy[0])?, f64::from_str(xy[1])?, b))
})
.collect::<anyhow::Result<Vec<_>>>()?;
let start_time = std::time::Instant::now();
let (mask, iou_predictions) = sam.forward(&image, &points, false)?;
println!(
"mask generated in {:.2}s",
start_time.elapsed().as_secs_f32()
);
println!("mask:\n{mask}");
println!("iou_predictions: {iou_predictions}");
let mask = (mask.ge(args.threshold)? * 255.)?;
let (_one, h, w) = mask.dims3()?;
let mask = mask.expand((3, h, w))?;
let mut img = image::ImageReader::open(&args.image)?
.decode()
.map_err(candle::Error::wrap)?;
let mask_pixels = mask.permute((1, 2, 0))?.flatten_all()?.to_vec1::<u8>()?;
let mask_img: image::ImageBuffer<image::Rgb<u8>, Vec<u8>> =
match image::ImageBuffer::from_raw(w as u32, h as u32, mask_pixels) {
Some(image) => image,
None => anyhow::bail!("error saving merged image"),
};
let mask_img = image::DynamicImage::from(mask_img).resize_to_fill(
img.width(),
img.height(),
image::imageops::FilterType::CatmullRom,
);
for x in 0..img.width() {
for y in 0..img.height() {
let mask_p = imageproc::drawing::Canvas::get_pixel(&mask_img, x, y);
if mask_p.0[0] > 100 {
let mut img_p = imageproc::drawing::Canvas::get_pixel(&img, x, y);
img_p.0[2] = 255 - (255 - img_p.0[2]) / 2;
img_p.0[1] /= 2;
img_p.0[0] /= 2;
imageproc::drawing::Canvas::draw_pixel(&mut img, x, y, img_p)
}
}
}
for (x, y, b) in points {
let x = (x * img.width() as f64) as i32;
let y = (y * img.height() as f64) as i32;
let color = if b {
image::Rgba([255, 0, 0, 200])
} else {
image::Rgba([0, 255, 0, 200])
};
imageproc::drawing::draw_filled_circle_mut(&mut img, (x, y), 3, color);
}
img.save("sam_merged.jpg")?
}
Ok(())
}
| candle/candle-examples/examples/segment-anything/main.rs/0 | {
"file_path": "candle/candle-examples/examples/segment-anything/main.rs",
"repo_id": "candle",
"token_count": 3137
} |
# candle-stable-lm
StableLM-3B-4E1T is a 3 billion parameter decoder-only language model
pre-trained on 1 trillion tokens of diverse English and code datasets for 4
epochs. See the [HuggingFace Hub Model
Card](https://huggingface.co/stabilityai/stablelm-3b-4e1t).
Note that this model is gated so you will have to request access on the Hub in
order to be able to use it.
Other available models are Stable-Code-3B, StableLM-2 and Zephyr variants.
## Running some example
```bash
$ cargo run --example stable-lm --release --features cuda -- --prompt 'What is the most efficient programming language in use?' --sample-len 150
avx: true, neon: false, simd128: false, f16c: true
temp: 0.00 repeat-penalty: 1.10 repeat-last-n: 64
retrieved the files in 126.593µs
loaded the model in 3.474148965s
What is the most efficient programming language in use?
The answer to this question depends on what you mean by "efficient". If you're talking about speed, then C++ and Java are probably your best bets. But if you're talking about ease of development, then Python is probably the way to go.
Python is a high-level, interpreted language that is easy to learn and use. It has a large community of developers who are always working on new features and improvements.
C++ is a low-level, compiled language that can be used for both desktop applications and web development. It's more difficult to learn than Python but offers greater control over the code.
Java is another high-level language that is popular with programmers because it runs on many different platforms (including Android phones
150 tokens generated (37.61 token/s)
```
| candle/candle-examples/examples/stable-lm/README.md/0 | {
"file_path": "candle/candle-examples/examples/stable-lm/README.md",
"repo_id": "candle",
"token_count": 432
} |
#[cfg(feature = "accelerate")]
extern crate accelerate_src;
#[cfg(feature = "mkl")]
extern crate intel_mkl_src;
use anyhow::{Error as E, Result};
use candle::{Device, IndexOp, Tensor};
use candle_nn::{ops::softmax, VarBuilder};
use clap::{Parser, ValueEnum};
use hf_hub::{api::sync::Api, Repo, RepoType};
use rand::{distributions::Distribution, SeedableRng};
use tokenizers::Tokenizer;
mod multilingual;
use candle_transformers::models::whisper::{self as m, audio, Config};
use cpal::traits::{DeviceTrait, HostTrait, StreamTrait};
pub enum Model {
Normal(m::model::Whisper),
Quantized(m::quantized_model::Whisper),
}
// Maybe we should use some traits rather than doing the dispatch for all these.
impl Model {
pub fn config(&self) -> &Config {
match self {
Self::Normal(m) => &m.config,
Self::Quantized(m) => &m.config,
}
}
pub fn encoder_forward(&mut self, x: &Tensor, flush: bool) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.encoder.forward(x, flush),
Self::Quantized(m) => m.encoder.forward(x, flush),
}
}
pub fn decoder_forward(
&mut self,
x: &Tensor,
xa: &Tensor,
flush: bool,
) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.decoder.forward(x, xa, flush),
Self::Quantized(m) => m.decoder.forward(x, xa, flush),
}
}
pub fn decoder_final_linear(&self, x: &Tensor) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.decoder.final_linear(x),
Self::Quantized(m) => m.decoder.final_linear(x),
}
}
}
#[allow(dead_code)]
#[derive(Debug, Clone)]
struct DecodingResult {
tokens: Vec<u32>,
text: String,
avg_logprob: f64,
no_speech_prob: f64,
temperature: f64,
compression_ratio: f64,
}
#[allow(dead_code)]
#[derive(Debug, Clone)]
struct Segment {
start: f64,
duration: f64,
dr: DecodingResult,
}
struct Decoder {
model: Model,
rng: rand::rngs::StdRng,
task: Option<Task>,
timestamps: bool,
verbose: bool,
tokenizer: Tokenizer,
suppress_tokens: Tensor,
sot_token: u32,
transcribe_token: u32,
translate_token: u32,
eot_token: u32,
no_speech_token: u32,
no_timestamps_token: u32,
language_token: Option<u32>,
}
impl Decoder {
#[allow(clippy::too_many_arguments)]
fn new(
model: Model,
tokenizer: Tokenizer,
seed: u64,
device: &Device,
language_token: Option<u32>,
task: Option<Task>,
timestamps: bool,
verbose: bool,
) -> Result<Self> {
let no_timestamps_token = token_id(&tokenizer, m::NO_TIMESTAMPS_TOKEN)?;
// Suppress the notimestamps token when in timestamps mode.
// https://github.com/openai/whisper/blob/e8622f9afc4eba139bf796c210f5c01081000472/whisper/decoding.py#L452
let suppress_tokens: Vec<f32> = (0..model.config().vocab_size as u32)
.map(|i| {
if model.config().suppress_tokens.contains(&i)
|| timestamps && i == no_timestamps_token
{
f32::NEG_INFINITY
} else {
0f32
}
})
.collect();
let suppress_tokens = Tensor::new(suppress_tokens.as_slice(), device)?;
let sot_token = token_id(&tokenizer, m::SOT_TOKEN)?;
let transcribe_token = token_id(&tokenizer, m::TRANSCRIBE_TOKEN)?;
let translate_token = token_id(&tokenizer, m::TRANSLATE_TOKEN)?;
let eot_token = token_id(&tokenizer, m::EOT_TOKEN)?;
let no_speech_token = m::NO_SPEECH_TOKENS
.iter()
.find_map(|token| token_id(&tokenizer, token).ok());
let no_speech_token = match no_speech_token {
None => anyhow::bail!("unable to find any non-speech token"),
Some(n) => n,
};
Ok(Self {
model,
rng: rand::rngs::StdRng::seed_from_u64(seed),
tokenizer,
task,
timestamps,
verbose,
suppress_tokens,
sot_token,
transcribe_token,
translate_token,
eot_token,
no_speech_token,
language_token,
no_timestamps_token,
})
}
fn decode(&mut self, mel: &Tensor, t: f64) -> Result<DecodingResult> {
let model = &mut self.model;
let audio_features = model.encoder_forward(mel, true)?;
if self.verbose {
println!("audio features: {:?}", audio_features.dims());
}
let sample_len = model.config().max_target_positions / 2;
let mut sum_logprob = 0f64;
let mut no_speech_prob = f64::NAN;
let mut tokens = vec![self.sot_token];
if let Some(language_token) = self.language_token {
tokens.push(language_token);
}
match self.task {
None | Some(Task::Transcribe) => tokens.push(self.transcribe_token),
Some(Task::Translate) => tokens.push(self.translate_token),
}
if !self.timestamps {
tokens.push(self.no_timestamps_token);
}
for i in 0..sample_len {
let tokens_t = Tensor::new(tokens.as_slice(), mel.device())?;
// The model expects a batch dim but this inference loop does not handle
// it so we add it at this point.
let tokens_t = tokens_t.unsqueeze(0)?;
let ys = model.decoder_forward(&tokens_t, &audio_features, i == 0)?;
// Extract the no speech probability on the first iteration by looking at the first
// token logits and the probability for the according token.
if i == 0 {
let logits = model.decoder_final_linear(&ys.i(..1)?)?.i(0)?.i(0)?;
no_speech_prob = softmax(&logits, 0)?
.i(self.no_speech_token as usize)?
.to_scalar::<f32>()? as f64;
}
let (_, seq_len, _) = ys.dims3()?;
let logits = model
.decoder_final_linear(&ys.i((..1, seq_len - 1..))?)?
.i(0)?
.i(0)?;
// TODO: Besides suppress tokens, we should apply the heuristics from
// ApplyTimestampRules, i.e.:
// - Timestamps come in pairs, except before EOT.
// - Timestamps should be non-decreasing.
// - If the sum of the probabilities of timestamps is higher than any other tokens,
// only consider timestamps when sampling.
// https://github.com/openai/whisper/blob/e8622f9afc4eba139bf796c210f5c01081000472/whisper/decoding.py#L439
let logits = logits.broadcast_add(&self.suppress_tokens)?;
let next_token = if t > 0f64 {
let prs = softmax(&(&logits / t)?, 0)?;
let logits_v: Vec<f32> = prs.to_vec1()?;
let distr = rand::distributions::WeightedIndex::new(&logits_v)?;
distr.sample(&mut self.rng) as u32
} else {
let logits_v: Vec<f32> = logits.to_vec1()?;
logits_v
.iter()
.enumerate()
.max_by(|(_, u), (_, v)| u.total_cmp(v))
.map(|(i, _)| i as u32)
.unwrap()
};
tokens.push(next_token);
let prob = softmax(&logits, candle::D::Minus1)?
.i(next_token as usize)?
.to_scalar::<f32>()? as f64;
if next_token == self.eot_token || tokens.len() > model.config().max_target_positions {
break;
}
sum_logprob += prob.ln();
}
let text = self.tokenizer.decode(&tokens, true).map_err(E::msg)?;
let avg_logprob = sum_logprob / tokens.len() as f64;
Ok(DecodingResult {
tokens,
text,
avg_logprob,
no_speech_prob,
temperature: t,
compression_ratio: f64::NAN,
})
}
fn decode_with_fallback(&mut self, segment: &Tensor) -> Result<DecodingResult> {
for (i, &t) in m::TEMPERATURES.iter().enumerate() {
let dr: Result<DecodingResult> = self.decode(segment, t);
if i == m::TEMPERATURES.len() - 1 {
return dr;
}
// On errors, we try again with a different temperature.
match dr {
Ok(dr) => {
let needs_fallback = dr.compression_ratio > m::COMPRESSION_RATIO_THRESHOLD
|| dr.avg_logprob < m::LOGPROB_THRESHOLD;
if !needs_fallback || dr.no_speech_prob > m::NO_SPEECH_THRESHOLD {
return Ok(dr);
}
}
Err(err) => {
println!("Error running at {t}: {err}")
}
}
}
unreachable!()
}
fn run(&mut self, mel: &Tensor, times: Option<(f64, f64)>) -> Result<Vec<Segment>> {
let (_, _, content_frames) = mel.dims3()?;
let mut seek = 0;
let mut segments = vec![];
while seek < content_frames {
let start = std::time::Instant::now();
let time_offset = (seek * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
let segment_size = usize::min(content_frames - seek, m::N_FRAMES);
let mel_segment = mel.narrow(2, seek, segment_size)?;
let segment_duration = (segment_size * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
let dr = self.decode_with_fallback(&mel_segment)?;
seek += segment_size;
if dr.no_speech_prob > m::NO_SPEECH_THRESHOLD && dr.avg_logprob < m::LOGPROB_THRESHOLD {
println!("no speech detected, skipping {seek} {dr:?}");
continue;
}
let segment = Segment {
start: time_offset,
duration: segment_duration,
dr,
};
if self.timestamps {
println!(
"{:.1}s -- {:.1}s",
segment.start,
segment.start + segment.duration,
);
let mut tokens_to_decode = vec![];
let mut prev_timestamp_s = 0f32;
for &token in segment.dr.tokens.iter() {
if token == self.sot_token || token == self.eot_token {
continue;
}
// The no_timestamp_token is the last before the timestamp ones.
if token > self.no_timestamps_token {
let timestamp_s = (token - self.no_timestamps_token + 1) as f32 / 50.;
if !tokens_to_decode.is_empty() {
let text = self
.tokenizer
.decode(&tokens_to_decode, true)
.map_err(E::msg)?;
println!(" {:.1}s-{:.1}s: {}", prev_timestamp_s, timestamp_s, text);
tokens_to_decode.clear()
}
prev_timestamp_s = timestamp_s;
} else {
tokens_to_decode.push(token)
}
}
if !tokens_to_decode.is_empty() {
let text = self
.tokenizer
.decode(&tokens_to_decode, true)
.map_err(E::msg)?;
if !text.is_empty() {
println!(" {:.1}s-...: {}", prev_timestamp_s, text);
}
tokens_to_decode.clear()
}
} else {
match times {
Some((start, end)) => {
println!("{:.1}s -- {:.1}s: {}", start, end, segment.dr.text)
}
None => {
println!(
"{:.1}s -- {:.1}s: {}",
segment.start,
segment.start + segment.duration,
segment.dr.text,
)
}
}
}
if self.verbose {
println!("{seek}: {segment:?}, in {:?}", start.elapsed());
}
segments.push(segment)
}
Ok(segments)
}
fn set_language_token(&mut self, language_token: Option<u32>) {
self.language_token = language_token;
}
#[allow(dead_code)]
fn reset_kv_cache(&mut self) {
match &mut self.model {
Model::Normal(m) => m.reset_kv_cache(),
Model::Quantized(m) => m.reset_kv_cache(),
}
}
fn model(&mut self) -> &mut Model {
&mut self.model
}
}
pub fn token_id(tokenizer: &Tokenizer, token: &str) -> candle::Result<u32> {
match tokenizer.token_to_id(token) {
None => candle::bail!("no token-id for {token}"),
Some(id) => Ok(id),
}
}
#[derive(Clone, Copy, Debug, ValueEnum)]
enum Task {
Transcribe,
Translate,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq, ValueEnum)]
enum WhichModel {
Tiny,
#[value(name = "tiny.en")]
TinyEn,
Base,
#[value(name = "base.en")]
BaseEn,
Small,
#[value(name = "small.en")]
SmallEn,
Medium,
#[value(name = "medium.en")]
MediumEn,
Large,
LargeV2,
LargeV3,
LargeV3Turbo,
#[value(name = "distil-medium.en")]
DistilMediumEn,
#[value(name = "distil-large-v2")]
DistilLargeV2,
}
impl WhichModel {
fn is_multilingual(&self) -> bool {
match self {
Self::Tiny
| Self::Base
| Self::Small
| Self::Medium
| Self::Large
| Self::LargeV2
| Self::LargeV3
| Self::LargeV3Turbo
| Self::DistilLargeV2 => true,
Self::TinyEn | Self::BaseEn | Self::SmallEn | Self::MediumEn | Self::DistilMediumEn => {
false
}
}
}
fn model_and_revision(&self) -> (&'static str, &'static str) {
match self {
Self::Tiny => ("openai/whisper-tiny", "main"),
Self::TinyEn => ("openai/whisper-tiny.en", "refs/pr/15"),
Self::Base => ("openai/whisper-base", "refs/pr/22"),
Self::BaseEn => ("openai/whisper-base.en", "refs/pr/13"),
Self::Small => ("openai/whisper-small", "main"),
Self::SmallEn => ("openai/whisper-small.en", "refs/pr/10"),
Self::Medium => ("openai/whisper-medium", "main"),
Self::MediumEn => ("openai/whisper-medium.en", "main"),
Self::Large => ("openai/whisper-large", "refs/pr/36"),
Self::LargeV2 => ("openai/whisper-large-v2", "refs/pr/57"),
Self::LargeV3 => ("openai/whisper-large-v3", "main"),
Self::LargeV3Turbo => ("openai/whisper-large-v3-turbo", "main"),
Self::DistilMediumEn => ("distil-whisper/distil-medium.en", "main"),
Self::DistilLargeV2 => ("distil-whisper/distil-large-v2", "main"),
}
}
}
#[derive(Parser, Debug)]
#[command(author, version, about, long_about = None)]
struct Args {
/// Run on CPU rather than on GPU.
#[arg(long)]
cpu: bool,
#[arg(long)]
model_id: Option<String>,
/// The model to use, check out available models:
/// https://huggingface.co/models?search=whisper
#[arg(long)]
revision: Option<String>,
/// The model to be used, can be tiny, small, medium.
#[arg(long, default_value = "tiny.en")]
model: WhichModel,
/// The seed to use when generating random samples.
#[arg(long, default_value_t = 299792458)]
seed: u64,
/// Enable tracing (generates a trace-timestamp.json file).
#[arg(long)]
tracing: bool,
#[arg(long)]
quantized: bool,
/// Language.
#[arg(long)]
language: Option<String>,
/// Task, when no task is specified, the input tokens contain only the sot token which can
/// improve things when in no-timestamp mode.
#[arg(long)]
task: Option<Task>,
/// Timestamps mode, this is not fully implemented yet.
#[arg(long)]
timestamps: bool,
/// Print the full DecodingResult structure rather than just the text.
#[arg(long)]
verbose: bool,
/// The input device to use.
#[arg(long)]
device: Option<String>,
}
pub fn main() -> Result<()> {
use tracing_chrome::ChromeLayerBuilder;
use tracing_subscriber::prelude::*;
let args = Args::parse();
let _guard = if args.tracing {
let (chrome_layer, guard) = ChromeLayerBuilder::new().build();
tracing_subscriber::registry().with(chrome_layer).init();
Some(guard)
} else {
None
};
let device = candle_examples::device(args.cpu)?;
let (default_model, default_revision) = if args.quantized {
("lmz/candle-whisper", "main")
} else {
args.model.model_and_revision()
};
let default_model = default_model.to_string();
let default_revision = default_revision.to_string();
let (model_id, revision) = match (args.model_id, args.revision) {
(Some(model_id), Some(revision)) => (model_id, revision),
(Some(model_id), None) => (model_id, "main".to_string()),
(None, Some(revision)) => (default_model, revision),
(None, None) => (default_model, default_revision),
};
let (config_filename, tokenizer_filename, weights_filename) = {
let api = Api::new()?;
let repo = api.repo(Repo::with_revision(model_id, RepoType::Model, revision));
let (config, tokenizer, model) = if args.quantized {
let ext = match args.model {
WhichModel::TinyEn => "tiny-en",
WhichModel::Tiny => "tiny",
_ => unimplemented!("no quantized support for {:?}", args.model),
};
(
repo.get(&format!("config-{ext}.json"))?,
repo.get(&format!("tokenizer-{ext}.json"))?,
repo.get(&format!("model-{ext}-q80.gguf"))?,
)
} else {
let config = repo.get("config.json")?;
let tokenizer = repo.get("tokenizer.json")?;
let model = repo.get("model.safetensors")?;
(config, tokenizer, model)
};
(config, tokenizer, model)
};
let config: Config = serde_json::from_str(&std::fs::read_to_string(config_filename)?)?;
let tokenizer = Tokenizer::from_file(tokenizer_filename).map_err(E::msg)?;
let model = if args.quantized {
let vb = candle_transformers::quantized_var_builder::VarBuilder::from_gguf(
&weights_filename,
&device,
)?;
Model::Quantized(m::quantized_model::Whisper::load(&vb, config.clone())?)
} else {
let vb =
unsafe { VarBuilder::from_mmaped_safetensors(&[weights_filename], m::DTYPE, &device)? };
Model::Normal(m::model::Whisper::load(&vb, config.clone())?)
};
let mut decoder = Decoder::new(
model,
tokenizer.clone(),
args.seed,
&device,
/* language_token */ None,
args.task,
args.timestamps,
args.verbose,
)?;
let mel_bytes = match config.num_mel_bins {
80 => include_bytes!("../whisper/melfilters.bytes").as_slice(),
128 => include_bytes!("../whisper/melfilters128.bytes").as_slice(),
nmel => anyhow::bail!("unexpected num_mel_bins {nmel}"),
};
let mut mel_filters = vec![0f32; mel_bytes.len() / 4];
<byteorder::LittleEndian as byteorder::ByteOrder>::read_f32_into(mel_bytes, &mut mel_filters);
// Set up the input device and stream with the default input config.
let host = cpal::default_host();
let audio_device = match args.device.as_ref() {
None => host.default_input_device(),
Some(device) => host
.input_devices()?
.find(|x| x.name().map_or(false, |y| &y == device)),
}
.expect("failed to find the audio input device");
let audio_config = audio_device
.default_input_config()
.expect("Failed to get default input config");
println!("audio config {audio_config:?}");
let channel_count = audio_config.channels() as usize;
let in_sample_rate = audio_config.sample_rate().0 as usize;
let resample_ratio = 16000. / in_sample_rate as f64;
let mut resampler = rubato::FastFixedIn::new(
resample_ratio,
10.,
rubato::PolynomialDegree::Septic,
1024,
1,
)?;
let (tx, rx) = std::sync::mpsc::channel();
let stream = audio_device.build_input_stream(
&audio_config.config(),
move |pcm: &[f32], _: &cpal::InputCallbackInfo| {
let pcm = pcm
.iter()
.step_by(channel_count)
.copied()
.collect::<Vec<f32>>();
if !pcm.is_empty() {
tx.send(pcm).unwrap()
}
},
move |err| {
eprintln!("an error occurred on stream: {}", err);
},
None,
)?;
stream.play()?;
// loop to process the audio data forever (until the user stops the program)
println!("transcribing audio...");
let mut buffered_pcm = vec![];
let mut language_token_set = false;
while let Ok(pcm) = rx.recv() {
use rubato::Resampler;
buffered_pcm.extend_from_slice(&pcm);
if buffered_pcm.len() < 10 * in_sample_rate {
continue;
}
let mut resampled_pcm = vec![];
// resample the audio, one chunk of 1024 samples at a time.
// in case the audio input failed to produce an exact multiple of 1024 samples,
// process the remainder on the next iteration of the loop.
let full_chunks = buffered_pcm.len() / 1024;
let remainder = buffered_pcm.len() % 1024;
for chunk in 0..full_chunks {
let buffered_pcm = &buffered_pcm[chunk * 1024..(chunk + 1) * 1024];
let pcm = resampler.process(&[&buffered_pcm], None)?;
resampled_pcm.extend_from_slice(&pcm[0]);
}
let pcm = resampled_pcm;
println!("{} {}", buffered_pcm.len(), pcm.len());
if remainder == 0 {
buffered_pcm.clear();
} else {
// efficiently copy the remainder to the beginning of the `buffered_pcm` buffer and
// truncate it. That's more efficient then allocating a new vector and copying into it
println!("audio device produced partial chunk with {remainder} samples; processing the remainder on the next iteration of the loop");
buffered_pcm.copy_within(full_chunks * 1024.., 0);
buffered_pcm.truncate(remainder);
}
let mel = audio::pcm_to_mel(&config, &pcm, &mel_filters);
let mel_len = mel.len();
let mel = Tensor::from_vec(
mel,
(1, config.num_mel_bins, mel_len / config.num_mel_bins),
&device,
)?;
// on the first iteration, we detect the language and set the language token.
if !language_token_set {
let language_token = match (args.model.is_multilingual(), args.language.clone()) {
(true, None) => Some(multilingual::detect_language(
decoder.model(),
&tokenizer,
&mel,
)?),
(false, None) => None,
(true, Some(language)) => match token_id(&tokenizer, &format!("<|{language}|>")) {
Ok(token_id) => Some(token_id),
Err(_) => anyhow::bail!("language {language} is not supported"),
},
(false, Some(_)) => {
anyhow::bail!("a language cannot be set for non-multilingual models")
}
};
println!("language_token: {:?}", language_token);
decoder.set_language_token(language_token);
language_token_set = true;
}
decoder.run(&mel, None)?;
decoder.reset_kv_cache();
}
Ok(())
}
| candle/candle-examples/examples/whisper-microphone/main.rs/0 | {
"file_path": "candle/candle-examples/examples/whisper-microphone/main.rs",
"repo_id": "candle",
"token_count": 12459
} |
/******************************************************************************
* Copyright (c) 2023, Tri Dao.
******************************************************************************/
#pragma once
#include "cute/algorithm/copy.hpp"
#include "cutlass/cutlass.h"
#include "cutlass/layout/layout.h"
#include <cutlass/numeric_types.h>
using namespace cute;
template<int kHeadDim_, int kBlockM_, int kBlockN_, int kNWarps_, typename elem_type=cutlass::half_t>
struct Flash_kernel_traits_sm90 {
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 800
using Element = elem_type;
static constexpr bool Has_cp_async = true;
#else
using Element = cutlass::half_t;
static constexpr bool Has_cp_async = false;
#endif
using ElementAccum = float;
using index_t = uint32_t;
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 800
using MMA_Atom_Arch = std::conditional_t<
std::is_same_v<elem_type, cutlass::half_t>,
MMA_Atom<SM80_16x8x16_F32F16F16F32_TN>,
MMA_Atom<SM80_16x8x16_F32BF16BF16F32_TN>
>;
using ValLayoutMNK = Layout<Shape<_1, _2, _1>>;
#else
using MMA_Atom_Arch = MMA_Atom<SM75_16x8x8_F32F16F16F32_TN>;
using ValLayoutMNK = Layout<Shape<_1, _2, _2>>;
#endif
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 750
using SmemCopyAtom = Copy_Atom<SM75_U32x4_LDSM_N, elem_type>;
using SmemCopyAtomTransposed = Copy_Atom<SM75_U16x8_LDSM_T, elem_type>;
#else
using SmemCopyAtom = Copy_Atom<DefaultCopy, elem_type>;
using SmemCopyAtomTransposed = Copy_Atom<DefaultCopy, elem_type>;
#endif
};
template<int kHeadDim_, int kBlockM_, int kBlockN_, int kNWarps_, bool Is_Q_in_regs_=false, bool Share_Q_K_smem_=false, typename elem_type=cutlass::half_t,
typename Base=Flash_kernel_traits_sm90<kHeadDim_, kBlockM_, kBlockN_, kNWarps_, elem_type> >
struct Flash_fwd_kernel_traits : public Base {
using Element = typename Base::Element;
using ElementAccum = typename Base::ElementAccum;
using index_t = typename Base::index_t;
static constexpr bool Has_cp_async = Base::Has_cp_async;
using SmemCopyAtom = typename Base::SmemCopyAtom;
using SmemCopyAtomTransposed = typename Base::SmemCopyAtomTransposed;
static constexpr bool Share_Q_K_smem = Share_Q_K_smem_;
static constexpr bool Is_Q_in_regs = Is_Q_in_regs_ || Share_Q_K_smem;
// The number of threads.
static constexpr int kNWarps = kNWarps_;
static constexpr int kNThreads = kNWarps * 32;
static constexpr int kBlockM = kBlockM_;
static constexpr int kBlockN = kBlockN_;
static constexpr int kHeadDim = kHeadDim_;
static_assert(kHeadDim % 32 == 0);
static constexpr int kBlockKSmem = kHeadDim % 64 == 0 ? 64 : 32;
static constexpr int kBlockKGmem = kHeadDim % 128 == 0 ? 128 : (kHeadDim % 64 == 0 ? 64 : 32);
static constexpr int kSwizzle = kBlockKSmem == 32 ? 2 : 3;
using TiledMma = TiledMMA<
typename Base::MMA_Atom_Arch,
Layout<Shape<Int<kNWarps>,_1,_1>>, // 4x1x1 or 8x1x1 thread group
typename Base::ValLayoutMNK>; // 1x2x1 or 1x2x2 value group for 16x16x16 MMA and LDSM
using SmemLayoutAtomQ = decltype(
composition(Swizzle<kSwizzle, 3, 3>{},
// This has to be kBlockKSmem, using kHeadDim gives wrong results for d=128
Layout<Shape<_8, Int<kBlockKSmem>>,
Stride<Int<kBlockKSmem>, _1>>{}));
using SmemLayoutQ = decltype(tile_to_shape(
SmemLayoutAtomQ{},
Shape<Int<kBlockM>, Int<kHeadDim>>{}));
using SmemLayoutKV = decltype(tile_to_shape(
SmemLayoutAtomQ{},
Shape<Int<kBlockN>, Int<kHeadDim>>{}));
using SmemLayoutAtomVtransposed = decltype(
composition(Swizzle<kSwizzle, 3, 3>{},
// This has to be kBlockN and not 8, otherwise we get wrong results for d=128
Layout<Shape<Int<kBlockKSmem>, Int<kBlockN>>,
Stride<_1, Int<kBlockKSmem>>>{}));
using SmemLayoutVtransposed = decltype(tile_to_shape(
SmemLayoutAtomVtransposed{},
Shape<Int<kHeadDim>, Int<kBlockN>>{}));
// Maybe the VtransposeNoSwizzle just needs to have the right shape
// And the strides don't matter?
using SmemLayoutVtransposedNoSwizzle = decltype(SmemLayoutVtransposed{}.layout_fn());
using SmemLayoutAtomO = decltype(
composition(Swizzle<kSwizzle, 3, 3>{},
Layout<Shape<Int<8>, Int<kBlockKSmem>>,
Stride<Int<kBlockKSmem>, _1>>{}));
using SmemLayoutO = decltype(tile_to_shape(
SmemLayoutAtomO{},
Shape<Int<kBlockM>, Int<kHeadDim>>{}));
using SmemCopyAtomO = Copy_Atom<DefaultCopy, elem_type>;
static constexpr int kSmemQCount = size(SmemLayoutQ{});
static constexpr int kSmemKVCount = size(SmemLayoutKV{}) * 2;
static constexpr int kSmemQSize = kSmemQCount * sizeof(Element);
static constexpr int kSmemKVSize = kSmemKVCount * sizeof(Element);
static constexpr int kSmemSize = Share_Q_K_smem ? std::max(kSmemQSize, kSmemKVSize) : kSmemQSize + kSmemKVSize;
static constexpr int kGmemElemsPerLoad = sizeof(cute::uint128_t) / sizeof(Element);
static_assert(kHeadDim % kGmemElemsPerLoad == 0, "kHeadDim must be a multiple of kGmemElemsPerLoad");
// Using kBlockKSmem here is 6-10% faster than kBlockKGmem for d=128 because of bank conflicts.
// For example, for d=128, smem is split into 2 "pages", each page takes care of columns
// 0-63 and 64-127. If we have 16 threads per row for gmem read, when we write to smem,
// thread 0 - 7 will write to the first page and thread 8 - 15 will write to the second page,
// to the same banks.
static constexpr int kGmemThreadsPerRow = kBlockKSmem / kGmemElemsPerLoad;
static_assert(kNThreads % kGmemThreadsPerRow == 0, "kNThreads must be a multiple of kGmemThreadsPerRow");
using GmemLayoutAtom = Layout<Shape <Int<kNThreads / kGmemThreadsPerRow>, Int<kGmemThreadsPerRow>>,
Stride<Int<kGmemThreadsPerRow>, _1>>;
// We use CACHEGLOBAL instead of CACHEALWAYS for both Q and K/V, since we won't be reading
// from the same address by the same threadblock. This is slightly faster.
using Gmem_copy_struct = std::conditional_t<
Has_cp_async,
SM80_CP_ASYNC_CACHEGLOBAL<cute::uint128_t>,
DefaultCopy
>;
using GmemTiledCopyQKV = decltype(
make_tiled_copy(Copy_Atom<Gmem_copy_struct, elem_type>{},
GmemLayoutAtom{},
Layout<Shape<_1, _8>>{})); // Val layout, 8 vals per read
using GmemTiledCopyO = decltype(
make_tiled_copy(Copy_Atom<DefaultCopy, elem_type>{},
GmemLayoutAtom{},
Layout<Shape<_1, _8>>{})); // Val layout, 8 vals per store
static constexpr int kGmemThreadsPerRowP = kBlockN / kGmemElemsPerLoad;
static_assert(kNThreads % kGmemThreadsPerRowP == 0, "kNThreads must be a multiple of kGmemThreadsPerRowP");
using GmemLayoutAtomP = Layout<Shape <Int<kNThreads / kGmemThreadsPerRowP>, Int<kGmemThreadsPerRowP>>,
Stride<Int<kGmemThreadsPerRowP>, _1>>;
using GmemTiledCopyP = decltype(
make_tiled_copy(Copy_Atom<DefaultCopy, elem_type>{},
GmemLayoutAtomP{},
Layout<Shape<_1, _8>>{})); // Val layout, 8 vals per store
};
////////////////////////////////////////////////////////////////////////////////////////////////////
| candle/candle-flash-attn/kernels/kernel_traits_sm90.h/0 | {
"file_path": "candle/candle-flash-attn/kernels/kernel_traits_sm90.h",
"repo_id": "candle",
"token_count": 3269
} |
#include "cuda_utils.cuh"
#define BINARY_OP_OUT(TYPENAME, OUT_TYPENAME, FN_NAME, FUNC) \
extern "C" __global__ void FN_NAME( \
const size_t numel, \
const size_t num_dims, \
const size_t *dims_and_strides, \
const TYPENAME *lhs, \
const TYPENAME *rhs, \
OUT_TYPENAME *out \
) { \
const size_t *dims = dims_and_strides; \
const size_t *lhs_strides = dims_and_strides + 1 * num_dims; \
const size_t *rhs_strides = dims_and_strides + 2 * num_dims; \
bool lhs_cont = dims_and_strides == nullptr || is_contiguous(num_dims, dims, lhs_strides); \
bool rhs_cont = dims_and_strides == nullptr || is_contiguous(num_dims, dims, rhs_strides); \
if (lhs_cont && rhs_cont) { \
for (unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; i < numel; i += blockDim.x * gridDim.x) { \
TYPENAME x = lhs[i]; \
TYPENAME y = rhs[i]; \
out[i] = FUNC; \
} \
} else if (lhs_cont) { \
for (unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; i < numel; i += blockDim.x * gridDim.x) { \
unsigned int tmp_i = i; \
unsigned int rhs_i = 0; \
for (int d = num_dims - 1; d >= 0; d--) { \
unsigned int i_dim = tmp_i % dims[d]; \
rhs_i += i_dim * rhs_strides[d]; \
tmp_i /= dims[d]; \
} \
TYPENAME x = lhs[i]; \
TYPENAME y = rhs[rhs_i]; \
out[i] = FUNC; \
} \
} else if (rhs_cont) { \
for (unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; i < numel; i += blockDim.x * gridDim.x) { \
unsigned int tmp_i = i; \
unsigned int lhs_i = 0; \
for (int d = num_dims - 1; d >= 0; d--) { \
unsigned int i_dim = tmp_i % dims[d]; \
lhs_i += i_dim * lhs_strides[d]; \
tmp_i /= dims[d]; \
} \
TYPENAME x = lhs[lhs_i]; \
TYPENAME y = rhs[i]; \
out[i] = FUNC; \
} \
} else { \
for (unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; i < numel; i += blockDim.x * gridDim.x) { \
unsigned int tmp_i = i; \
unsigned int lhs_i = 0; \
unsigned int rhs_i = 0; \
for (int d = num_dims - 1; d >= 0; d--) { \
unsigned int i_dim = tmp_i % dims[d]; \
lhs_i += i_dim * lhs_strides[d]; \
rhs_i += i_dim * rhs_strides[d]; \
tmp_i /= dims[d]; \
} \
TYPENAME x = lhs[lhs_i]; \
TYPENAME y = rhs[rhs_i]; \
out[i] = FUNC; \
} \
} \
} \
#define BINARY_OP(TYPENAME, FN_NAME, FUNC) \
BINARY_OP_OUT(TYPENAME, TYPENAME, FN_NAME, FUNC)
| candle/candle-kernels/src/binary_op_macros.cuh/0 | {
"file_path": "candle/candle-kernels/src/binary_op_macros.cuh",
"repo_id": "candle",
"token_count": 1561
} |
#include <metal_stdlib>
METAL_FUNC uint get_strided_index(
uint idx,
constant size_t &num_dims,
constant size_t *dims,
constant size_t *strides
) {
uint strided_i = 0;
for (uint d = 0; d < num_dims; d++) {
uint dim_idx = num_dims - 1 - d;
strided_i += (idx % dims[dim_idx]) * strides[dim_idx];
idx /= dims[dim_idx];
}
return strided_i;
}
using namespace metal;
#define AFFINE(FN_NAME, T) \
kernel void FN_NAME( \
constant size_t &dim, \
constant float &mul, \
constant float &add, \
device const T *input, \
device T *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
output[id] = T(fma(float(input[id]), mul, add)); \
} \
kernel void FN_NAME##_strided( \
constant size_t &dim, \
constant size_t &num_dims, \
constant size_t *dims, \
constant size_t *strides, \
constant float &mul, \
constant float &add, \
device const T *input, \
device T *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
output[id] = T(fma(float(input[get_strided_index(id, num_dims, dims, strides)]), mul, add)); \
}
#define POWF(FN_NAME, TYPENAME) \
kernel void FN_NAME( \
constant size_t &dim, \
constant float &mul, \
device const TYPENAME *input, \
device TYPENAME *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
output[id] = TYPENAME(pow(input[id], TYPENAME(mul))); \
} \
kernel void FN_NAME##_strided( \
constant size_t &dim, \
constant size_t &num_dims, \
constant size_t *dims, \
constant size_t *strides, \
constant float &mul, \
device const TYPENAME *input, \
device TYPENAME *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
output[id] = TYPENAME(pow(input[get_strided_index(id, num_dims, dims, strides)], TYPENAME(mul))); \
}
#define ELU(FN_NAME, TYPENAME) \
kernel void FN_NAME( \
constant size_t &dim, \
constant float &mul, \
device const TYPENAME *input, \
device TYPENAME *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
const TYPENAME x = input[id]; \
output[id] = TYPENAME((x > 0)?x: mul * (exp(x) - 1)); \
} \
kernel void FN_NAME##_strided( \
constant size_t &dim, \
constant size_t &num_dims, \
constant size_t *dims, \
constant size_t *strides, \
constant float &mul, \
device const TYPENAME *input, \
device TYPENAME *output, \
uint id [[ thread_position_in_grid ]] \
) { \
if (id >= dim) { \
return; \
} \
const TYPENAME x = input[get_strided_index(id, num_dims, dims, strides)]; \
output[id] = TYPENAME((x > 0)?x: mul * (exp(x) - 1)); \
} \
AFFINE(affine_u8, uint8_t)
AFFINE(affine_u32, uint32_t)
AFFINE(affine_f32, float)
AFFINE(affine_f16, half)
POWF(powf_f32, float)
POWF(powf_f16, half)
ELU(elu_f32, float)
ELU(elu_f16, half)
#if defined(__HAVE_BFLOAT__)
AFFINE(affine_bf16, bfloat);
POWF(powf_bf16, bfloat);
ELU(elu_bf16, bfloat);
#endif
| candle/candle-metal-kernels/src/affine.metal/0 | {
"file_path": "candle/candle-metal-kernels/src/affine.metal",
"repo_id": "candle",
"token_count": 1498
} |
#include <metal_stdlib>
using namespace metal;
METAL_FUNC uint get_strided_index(
uint idx,
constant size_t &num_dims,
constant size_t *dims,
constant size_t *strides
) {
uint strided_i = 0;
for (uint d = 0; d < num_dims; d++) {
uint dim_idx = num_dims - 1 - d;
strided_i += (idx % dims[dim_idx]) * strides[dim_idx];
idx /= dims[dim_idx];
}
return strided_i;
}
template<typename T, typename ID>
METAL_FUNC void where_cond(
constant size_t &numel,
constant size_t &num_dims,
constant size_t *dims,
constant size_t *strides,
constant size_t *strides_t,
constant size_t *strides_f,
device const ID *ids,
device const T *t,
device const T *f,
device T *out,
uint i [[ thread_position_in_grid ]]
) {
if (i >= numel){
return;
}
uint strided_i = get_strided_index(i, num_dims, dims, strides);
uint strided_i_t = get_strided_index(i, num_dims, dims, strides_t);
uint strided_i_f = get_strided_index(i, num_dims, dims, strides_f);
out[i] = ids[strided_i] ? t[strided_i_t] : f[strided_i_f];
}
#define WHERE_OP(T, ID, FN_NAME) \
kernel void FN_NAME( \
constant size_t &numel, \
constant size_t &num_dims, \
constant size_t *dims, \
constant size_t *strides, \
constant size_t *strides_t, \
constant size_t *strides_f, \
device const ID *ids, \
device const T *t, \
device const T *f, \
device T *out, \
uint i [[ thread_position_in_grid ]] \
) { \
where_cond<T, ID>(numel, num_dims, dims, strides, strides_t, strides_f, ids, t, f, out, i); \
} \
WHERE_OP(half, uint32_t, where_u32_f16)
WHERE_OP(float, uint32_t, where_u32_f32)
WHERE_OP(uint8_t, uint32_t, where_u32_u8)
WHERE_OP(uint32_t, uint32_t, where_u32_u32)
WHERE_OP(half, uint8_t, where_u8_f16)
WHERE_OP(float, uint8_t, where_u8_f32)
WHERE_OP(uint8_t, uint8_t, where_u8_u8)
WHERE_OP(uint32_t, uint8_t, where_u8_u32)
#if __METAL_VERSION__ >= 220
WHERE_OP(int64_t, uint8_t, where_u8_i64)
WHERE_OP(int64_t, uint32_t, where_u32_i64)
WHERE_OP(half, int64_t, where_i64_f16)
WHERE_OP(float, int64_t, where_i64_f32)
WHERE_OP(uint8_t, int64_t, where_i64_u8)
WHERE_OP(uint32_t, int64_t, where_i64_u32)
WHERE_OP(int64_t, int64_t, where_i64_i64)
#if defined(__HAVE_BFLOAT__)
WHERE_OP(bfloat, int64_t, where_i64_bf16)
#endif
#endif
#if defined(__HAVE_BFLOAT__)
WHERE_OP(bfloat, uint8_t, where_u8_bf16)
WHERE_OP(bfloat, uint32_t, where_u32_bf16)
#endif
| candle/candle-metal-kernels/src/ternary.metal/0 | {
"file_path": "candle/candle-metal-kernels/src/ternary.metal",
"repo_id": "candle",
"token_count": 2256
} |
#[cfg(feature = "mkl")]
extern crate intel_mkl_src;
#[cfg(feature = "accelerate")]
extern crate accelerate_src;
use candle::{DType, Device, Result, Tensor};
use candle_nn::{linear, AdamW, Linear, Module, Optimizer, ParamsAdamW, VarBuilder, VarMap};
fn gen_data() -> Result<(Tensor, Tensor)> {
// Generate some sample linear data.
let w_gen = Tensor::new(&[[3f32, 1.]], &Device::Cpu)?;
let b_gen = Tensor::new(-2f32, &Device::Cpu)?;
let gen = Linear::new(w_gen, Some(b_gen));
let sample_xs = Tensor::new(&[[2f32, 1.], [7., 4.], [-4., 12.], [5., 8.]], &Device::Cpu)?;
let sample_ys = gen.forward(&sample_xs)?;
Ok((sample_xs, sample_ys))
}
fn main() -> Result<()> {
let (sample_xs, sample_ys) = gen_data()?;
// Use backprop to run a linear regression between samples and get the coefficients back.
let varmap = VarMap::new();
let vb = VarBuilder::from_varmap(&varmap, DType::F32, &Device::Cpu);
let model = linear(2, 1, vb.pp("linear"))?;
let params = ParamsAdamW {
lr: 0.1,
..Default::default()
};
let mut opt = AdamW::new(varmap.all_vars(), params)?;
for step in 0..10000 {
let ys = model.forward(&sample_xs)?;
let loss = ys.sub(&sample_ys)?.sqr()?.sum_all()?;
opt.backward_step(&loss)?;
println!("{step} {}", loss.to_vec0::<f32>()?);
}
Ok(())
}
| candle/candle-nn/examples/basic_optimizer.rs/0 | {
"file_path": "candle/candle-nn/examples/basic_optimizer.rs",
"repo_id": "candle",
"token_count": 595
} |
//! Various optimization algorithms.
use candle::{Result, Tensor, Var};
/// The interface optimizers should implement.
pub trait Optimizer: Sized {
type Config: Sized;
fn new(vars: Vec<Var>, config: Self::Config) -> Result<Self>;
fn step(&mut self, grads: &candle::backprop::GradStore) -> Result<()>;
fn learning_rate(&self) -> f64;
fn set_learning_rate(&mut self, lr: f64);
fn empty(config: Self::Config) -> Result<Self> {
Self::new(vec![], config)
}
fn backward_step(&mut self, loss: &Tensor) -> Result<()> {
let grads = loss.backward()?;
self.step(&grads)
}
fn from_slice(vars: &[&Var], config: Self::Config) -> Result<Self> {
let vars: Vec<_> = vars.iter().map(|&v| v.clone()).collect();
Self::new(vars, config)
}
}
/// Optimizer for Stochastic Gradient Descent.
///
/// Contrary to the PyTorch implementation of SGD, this version does not support momentum.
#[derive(Debug)]
pub struct SGD {
vars: Vec<Var>,
learning_rate: f64,
}
impl Optimizer for SGD {
type Config = f64;
fn new(vars: Vec<Var>, learning_rate: f64) -> Result<Self> {
let vars = vars
.into_iter()
.filter(|var| var.dtype().is_float())
.collect();
Ok(Self {
vars,
learning_rate,
})
}
fn learning_rate(&self) -> f64 {
self.learning_rate
}
fn step(&mut self, grads: &candle::backprop::GradStore) -> Result<()> {
for var in self.vars.iter() {
if let Some(grad) = grads.get(var) {
var.set(&var.sub(&(grad * self.learning_rate)?)?)?;
}
}
Ok(())
}
fn set_learning_rate(&mut self, lr: f64) {
self.learning_rate = lr
}
}
impl SGD {
pub fn into_inner(self) -> Vec<Var> {
self.vars
}
pub fn push(&mut self, var: &Var) {
self.vars.push(var.clone())
}
}
#[derive(Clone, Debug)]
pub struct ParamsAdamW {
pub lr: f64,
pub beta1: f64,
pub beta2: f64,
pub eps: f64,
pub weight_decay: f64,
}
impl Default for ParamsAdamW {
fn default() -> Self {
Self {
lr: 0.001,
beta1: 0.9,
beta2: 0.999,
eps: 1e-8,
weight_decay: 0.01,
}
}
}
#[derive(Debug)]
struct VarAdamW {
var: Var,
first_moment: Var,
second_moment: Var,
}
#[derive(Debug)]
pub struct AdamW {
vars: Vec<VarAdamW>,
step_t: usize,
params: ParamsAdamW,
}
impl Optimizer for AdamW {
type Config = ParamsAdamW;
fn new(vars: Vec<Var>, params: ParamsAdamW) -> Result<Self> {
let vars = vars
.into_iter()
.filter(|var| var.dtype().is_float())
.map(|var| {
let dtype = var.dtype();
let shape = var.shape();
let device = var.device();
let first_moment = Var::zeros(shape, dtype, device)?;
let second_moment = Var::zeros(shape, dtype, device)?;
Ok(VarAdamW {
var,
first_moment,
second_moment,
})
})
.collect::<Result<Vec<_>>>()?;
Ok(Self {
vars,
params,
step_t: 0,
})
}
fn learning_rate(&self) -> f64 {
self.params.lr
}
fn set_learning_rate(&mut self, lr: f64) {
self.params.lr = lr
}
fn step(&mut self, grads: &candle::backprop::GradStore) -> Result<()> {
self.step_t += 1;
let lr = self.params.lr;
let lambda = self.params.weight_decay;
let lr_lambda = lr * lambda;
let beta1 = self.params.beta1;
let beta2 = self.params.beta2;
let scale_m = 1f64 / (1f64 - beta1.powi(self.step_t as i32));
let scale_v = 1f64 / (1f64 - beta2.powi(self.step_t as i32));
for var in self.vars.iter() {
let theta = &var.var;
let m = &var.first_moment;
let v = &var.second_moment;
if let Some(g) = grads.get(theta) {
// This involves locking 3 RWLocks per params, if the parameters are large this
// should not be an issue but this may be problematic with models with lots of
// small parameters.
let next_m = ((m.as_tensor() * beta1)? + (g * (1.0 - beta1))?)?;
let next_v = ((v.as_tensor() * beta2)? + (g.sqr()? * (1.0 - beta2))?)?;
let m_hat = (&next_m * scale_m)?;
let v_hat = (&next_v * scale_v)?;
let next_theta = (theta.as_tensor() * (1f64 - lr_lambda))?;
let adjusted_grad = (m_hat / (v_hat.sqrt()? + self.params.eps)?)?;
let next_theta = (next_theta - (adjusted_grad * lr)?)?;
m.set(&next_m)?;
v.set(&next_v)?;
theta.set(&next_theta)?;
}
}
Ok(())
}
}
impl AdamW {
pub fn new_lr(vars: Vec<Var>, learning_rate: f64) -> Result<Self> {
let params = ParamsAdamW {
lr: learning_rate,
..ParamsAdamW::default()
};
Self::new(vars, params)
}
pub fn params(&self) -> &ParamsAdamW {
&self.params
}
pub fn set_params(&mut self, params: ParamsAdamW) {
self.params = params;
}
}
| candle/candle-nn/src/optim.rs/0 | {
"file_path": "candle/candle-nn/src/optim.rs",
"repo_id": "candle",
"token_count": 2798
} |
[package]
name = "candle-onnx"
version = "0.8.2"
edition = "2021"
description = "ONNX support for Candle"
repository = "https://github.com/huggingface/candle"
keywords = ["blas", "tensor", "machine-learning"]
categories = ["science"]
license = "MIT OR Apache-2.0"
[dependencies]
candle = { path = "../candle-core", package = "candle-core", version = "0.8.2" }
candle-nn = { path = "../candle-nn", version = "0.8.2" }
prost = "0.12.1"
[build-dependencies]
prost-build = "0.12.1"
[dev-dependencies]
anyhow = { version = "1", features = ["backtrace"] }
clap = { version = "4.2.4", features = ["derive"] }
| candle/candle-onnx/Cargo.toml/0 | {
"file_path": "candle/candle-onnx/Cargo.toml",
"repo_id": "candle",
"token_count": 242
} |
# Generated content DO NOT EDIT
from .. import functional
avg_pool2d = functional.avg_pool2d
gelu = functional.gelu
max_pool2d = functional.max_pool2d
relu = functional.relu
silu = functional.silu
softmax = functional.softmax
tanh = functional.tanh
| candle/candle-pyo3/py_src/candle/functional/__init__.py/0 | {
"file_path": "candle/candle-pyo3/py_src/candle/functional/__init__.py",
"repo_id": "candle",
"token_count": 84
} |
# Generated content DO NOT EDIT
from typing import Any, Callable, Dict, List, Optional, Tuple, Union, Sequence
from os import PathLike
from candle.typing import _ArrayLike, Device, Scalar, Index, Shape
from candle import Tensor, DType, QTensor
@staticmethod
def cuda_is_available() -> bool:
"""
Returns true if the 'cuda' backend is available.
"""
pass
@staticmethod
def get_num_threads() -> int:
"""
Returns the number of threads used by the candle.
"""
pass
@staticmethod
def has_accelerate() -> bool:
"""
Returns true if candle was compiled with 'accelerate' support.
"""
pass
@staticmethod
def has_mkl() -> bool:
"""
Returns true if candle was compiled with MKL support.
"""
pass
@staticmethod
def load_ggml(path, device=None) -> Tuple[Dict[str, QTensor], Dict[str, Any], List[str]]:
"""
Load a GGML file. Returns a tuple of three objects: a dictionary mapping tensor names to tensors,
a dictionary mapping hyperparameter names to hyperparameter values, and a vocabulary.
"""
pass
@staticmethod
def load_gguf(path, device=None) -> Tuple[Dict[str, QTensor], Dict[str, Any]]:
"""
Loads a GGUF file. Returns a tuple of two dictionaries: the first maps tensor names to tensors,
and the second maps metadata keys to metadata values.
"""
pass
@staticmethod
def load_safetensors(path: Union[str, PathLike]) -> Dict[str, Tensor]:
"""
Loads a safetensors file. Returns a dictionary mapping tensor names to tensors.
"""
pass
@staticmethod
def save_gguf(path, tensors, metadata):
"""
Save quanitzed tensors and metadata to a GGUF file.
"""
pass
@staticmethod
def save_safetensors(path: Union[str, PathLike], tensors: Dict[str, Tensor]) -> None:
"""
Saves a dictionary of tensors to a safetensors file.
"""
pass
| candle/candle-pyo3/py_src/candle/utils/__init__.pyi/0 | {
"file_path": "candle/candle-pyo3/py_src/candle/utils/__init__.pyi",
"repo_id": "candle",
"token_count": 654
} |
import candle
from candle import Tensor, QTensor
from candle.utils import load_safetensors, save_gguf, load_gguf, save_safetensors
from pathlib import Path
TEST_DIR = Path(__file__).parent.parent / "_workdir"
TEST_DIR.mkdir(exist_ok=True)
def test_can_roundtrip_safetensors():
tensors = {
"a": candle.randn((16, 256)),
"b": candle.randn((16, 16)),
}
file = str(TEST_DIR / "test.safetensors")
save_safetensors(file, tensors)
loaded_tensors = load_safetensors(file)
assert set(tensors.keys()) == set(loaded_tensors.keys())
for key in tensors.keys():
assert tensors[key].values() == loaded_tensors[key].values(), "Values are not equal"
assert tensors[key].shape == loaded_tensors[key].shape, "Shapes are not equal"
assert str(tensors[key].dtype) == str(loaded_tensors[key].dtype), "Dtypes are not equal"
def test_can_roundtrip_gguf():
metadata = {
"a": 1,
"b": "foo",
"c": [1, 2, 3],
"d": [[1, 2], [3, 4]],
}
tensors = {
"a": candle.randn((16, 256)).quantize("q4_0"),
"b": candle.randn((16, 16)).quantize("f32"),
}
file = str(TEST_DIR / "test.gguf")
save_gguf(file, tensors, metadata)
loaded_tensors, loaded_metadata = load_gguf(file)
assert set(metadata.keys()) == set(loaded_metadata.keys())
for key in metadata.keys():
assert metadata[key] == loaded_metadata[key]
assert set(tensors.keys()) == set(loaded_tensors.keys())
for key in tensors.keys():
assert tensors[key].dequantize().values() == loaded_tensors[key].dequantize().values(), "Values are not equal"
assert tensors[key].shape == loaded_tensors[key].shape, "Shapes are not equal"
assert str(tensors[key].ggml_dtype) == str(loaded_tensors[key].ggml_dtype), "Dtypes are not equal"
| candle/candle-pyo3/tests/native/test_utils.py/0 | {
"file_path": "candle/candle-pyo3/tests/native/test_utils.py",
"repo_id": "candle",
"token_count": 774
} |
//! Contrastive Language-Image Pre-Training
//!
//! Contrastive Language-Image Pre-Training (CLIP) is an architecture trained on
//! pairs of images with related texts.
//!
//! - [GH](https://github.com/openai/CLIP)
//! - [Code](https://github.com/huggingface/transformers/tree/f6fa0f0bf0796ac66f201f23bdb8585de1609add/src/transformers/models/clip)
use candle::{DType, Device, IndexOp, Result, Tensor, D};
use candle_nn as nn;
use candle_nn::Module;
use super::EncoderConfig;
#[derive(Debug, Clone, Copy)]
pub enum Activation {
QuickGelu,
}
impl Module for Activation {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
match self {
Activation::QuickGelu => xs * nn::ops::sigmoid(&(xs * 1.702f64)?)?,
}
}
}
#[derive(Debug, Clone)]
pub struct ClipTextConfig {
pub vocab_size: usize,
pub embed_dim: usize,
pub activation: Activation,
pub intermediate_size: usize,
pub max_position_embeddings: usize,
pub pad_with: Option<String>,
pub num_hidden_layers: usize,
pub num_attention_heads: usize,
#[allow(dead_code)]
pub projection_dim: usize,
}
impl ClipTextConfig {
// The config details can be found in the "text_config" section of this json file:
// https://huggingface.co/openai/clip-vit-large-patch14/blob/main/config.json
pub fn vit_base_patch32() -> Self {
Self {
vocab_size: 49408,
embed_dim: 512,
intermediate_size: 2048,
max_position_embeddings: 77,
pad_with: None,
num_hidden_layers: 12,
num_attention_heads: 8,
projection_dim: 512,
activation: Activation::QuickGelu,
}
}
}
// ClipTextEmbeddings mostly based on the existing implementation in the stable diffision model.
// TODO rewrite to be more similar to https://github.com/huggingface/transformers/blob/f6fa0f0bf0796ac66f201f23bdb8585de1609add/src/transformers/models/clip/modeling_clip.py#L142
#[derive(Clone, Debug)]
struct ClipTextEmbeddings {
token_embedding: candle_nn::Embedding,
position_embedding: candle_nn::Embedding,
position_ids: Tensor,
}
impl ClipTextEmbeddings {
fn new(vs: candle_nn::VarBuilder, c: &ClipTextConfig) -> Result<Self> {
let token_embedding =
candle_nn::embedding(c.vocab_size, c.embed_dim, vs.pp("token_embedding"))?;
let position_embedding: nn::Embedding = candle_nn::embedding(
c.max_position_embeddings,
c.embed_dim,
vs.pp("position_embedding"),
)?;
let position_ids =
Tensor::arange(0u32, c.max_position_embeddings as u32, vs.device())?.unsqueeze(0)?;
Ok(Self {
token_embedding,
position_embedding,
position_ids,
})
}
}
impl Module for ClipTextEmbeddings {
fn forward(&self, input_ids: &Tensor) -> Result<Tensor> {
let seq_length = input_ids.dim(D::Minus1)?;
let inputs_embeds = self.token_embedding.forward(input_ids)?;
let position_ids = self.position_ids.narrow(1, 0, seq_length)?;
let position_embedding = self.position_embedding.forward(&position_ids)?;
inputs_embeds.broadcast_add(&position_embedding)
}
}
#[derive(Clone, Debug)]
struct ClipAttention {
k_proj: candle_nn::Linear,
v_proj: candle_nn::Linear,
q_proj: candle_nn::Linear,
out_proj: candle_nn::Linear,
head_dim: usize,
scale: f64,
num_attention_heads: usize,
}
impl ClipAttention {
fn new(vs: candle_nn::VarBuilder, c: &EncoderConfig) -> Result<Self> {
let embed_dim = c.embed_dim();
let num_attention_heads = c.num_attention_heads();
let k_proj = candle_nn::linear(embed_dim, embed_dim, vs.pp("k_proj"))?;
let v_proj = candle_nn::linear(embed_dim, embed_dim, vs.pp("v_proj"))?;
let q_proj = candle_nn::linear(embed_dim, embed_dim, vs.pp("q_proj"))?;
let out_proj = candle_nn::linear(embed_dim, embed_dim, vs.pp("out_proj"))?;
let head_dim = embed_dim / num_attention_heads;
let scale = (head_dim as f64).powf(-0.5);
Ok(ClipAttention {
k_proj,
v_proj,
q_proj,
out_proj,
head_dim,
scale,
num_attention_heads,
})
}
fn shape(&self, xs: &Tensor, seq_len: usize, bsz: usize) -> Result<Tensor> {
xs.reshape((bsz, seq_len, self.num_attention_heads, self.head_dim))?
.transpose(1, 2)?
.contiguous()
}
fn forward(&self, xs: &Tensor, causal_attention_mask: Option<&Tensor>) -> Result<Tensor> {
let in_dtype = xs.dtype();
let (bsz, seq_len, embed_dim) = xs.dims3()?;
let query_states = (self.q_proj.forward(xs)? * self.scale)?;
let proj_shape = (bsz * self.num_attention_heads, seq_len, self.head_dim);
let query_states = self
.shape(&query_states, seq_len, bsz)?
.reshape(proj_shape)?
.to_dtype(DType::F32)?;
let key_states = self
.shape(&self.k_proj.forward(xs)?, seq_len, bsz)?
.reshape(proj_shape)?
.to_dtype(DType::F32)?;
let value_states = self
.shape(&self.v_proj.forward(xs)?, seq_len, bsz)?
.reshape(proj_shape)?
.to_dtype(DType::F32)?;
let attn_weights = query_states.matmul(&key_states.transpose(1, 2)?)?;
let src_len = key_states.dim(1)?;
let attn_weights = if let Some(causal_attention_mask) = causal_attention_mask {
attn_weights
.reshape((bsz, self.num_attention_heads, seq_len, src_len))?
.broadcast_add(causal_attention_mask)?
.reshape((bsz * self.num_attention_heads, seq_len, src_len))?
} else {
attn_weights
};
let attn_weights = candle_nn::ops::softmax(&attn_weights, D::Minus1)?;
let attn_output = attn_weights.matmul(&value_states)?.to_dtype(in_dtype)?;
let attn_output = attn_output
.reshape((bsz, self.num_attention_heads, seq_len, self.head_dim))?
.transpose(1, 2)?
.reshape((bsz, seq_len, embed_dim))?;
self.out_proj.forward(&attn_output)
}
}
#[derive(Clone, Debug)]
struct ClipMlp {
fc1: candle_nn::Linear,
fc2: candle_nn::Linear,
activation: Activation,
}
impl ClipMlp {
fn new(vs: candle_nn::VarBuilder, c: &EncoderConfig) -> Result<Self> {
let fc1 = candle_nn::linear(c.embed_dim(), c.intermediate_size(), vs.pp("fc1"))?;
let fc2 = candle_nn::linear(c.intermediate_size(), c.embed_dim(), vs.pp("fc2"))?;
Ok(ClipMlp {
fc1,
fc2,
activation: c.activation(),
})
}
}
impl ClipMlp {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
let xs = self.fc1.forward(xs)?;
self.fc2.forward(&self.activation.forward(&xs)?)
}
}
#[derive(Clone, Debug)]
struct ClipEncoderLayer {
self_attn: ClipAttention,
layer_norm1: candle_nn::LayerNorm,
mlp: ClipMlp,
layer_norm2: candle_nn::LayerNorm,
}
impl ClipEncoderLayer {
fn new(vs: candle_nn::VarBuilder, c: &EncoderConfig) -> Result<Self> {
let self_attn = ClipAttention::new(vs.pp("self_attn"), c)?;
let layer_norm1 = candle_nn::layer_norm(c.embed_dim(), 1e-5, vs.pp("layer_norm1"))?;
let mlp = ClipMlp::new(vs.pp("mlp"), c)?;
let layer_norm2 = candle_nn::layer_norm(c.embed_dim(), 1e-5, vs.pp("layer_norm2"))?;
Ok(ClipEncoderLayer {
self_attn,
layer_norm1,
mlp,
layer_norm2,
})
}
fn forward(&self, xs: &Tensor, causal_attention_mask: Option<&Tensor>) -> Result<Tensor> {
let residual = xs;
let xs = self.layer_norm1.forward(xs)?;
let xs = self.self_attn.forward(&xs, causal_attention_mask)?;
let xs = (xs + residual)?;
let residual = &xs;
let xs = self.layer_norm2.forward(&xs)?;
let xs = self.mlp.forward(&xs)?;
xs + residual
}
}
#[derive(Clone, Debug)]
pub struct ClipEncoder {
layers: Vec<ClipEncoderLayer>,
}
impl ClipEncoder {
pub fn new(vs: candle_nn::VarBuilder, c: &EncoderConfig) -> Result<Self> {
let vs = vs.pp("layers");
let mut layers: Vec<ClipEncoderLayer> = Vec::new();
for index in 0..c.num_hidden_layers() {
let layer = ClipEncoderLayer::new(vs.pp(index.to_string()), c)?;
layers.push(layer)
}
Ok(ClipEncoder { layers })
}
pub fn forward(&self, xs: &Tensor, causal_attention_mask: Option<&Tensor>) -> Result<Tensor> {
let mut xs = xs.clone();
for layer in self.layers.iter() {
xs = layer.forward(&xs, causal_attention_mask)?;
}
Ok(xs)
}
// required by LLaVA
pub fn output_hidden_states(
&self,
xs: &Tensor,
causal_attention_mask: Option<&Tensor>,
) -> Result<Vec<Tensor>> {
let mut xs = xs.clone();
let mut hidden_states = Vec::new();
for layer in self.layers.iter() {
xs = layer.forward(&xs, causal_attention_mask)?;
hidden_states.push(xs.clone());
}
Ok(hidden_states)
}
}
/// A CLIP transformer based model.
#[derive(Clone, Debug)]
pub struct ClipTextTransformer {
embeddings: ClipTextEmbeddings,
encoder: ClipEncoder,
final_layer_norm: candle_nn::LayerNorm,
}
impl ClipTextTransformer {
pub fn new(vs: candle_nn::VarBuilder, c: &ClipTextConfig) -> Result<Self> {
let embeddings = ClipTextEmbeddings::new(vs.pp("embeddings"), c)?;
let encoder = ClipEncoder::new(vs.pp("encoder"), &EncoderConfig::Text(c.clone()))?;
let final_layer_norm = candle_nn::layer_norm(c.embed_dim, 1e-5, vs.pp("final_layer_norm"))?;
Ok(ClipTextTransformer {
embeddings,
encoder,
final_layer_norm,
})
}
// TODO: rewrite to newer version
fn build_causal_attention_mask(
bsz: usize,
seq_len: usize,
mask_after: usize,
device: &Device,
) -> Result<Tensor> {
let mask: Vec<_> = (0..seq_len)
.flat_map(|i| {
(0..seq_len).map(move |j| {
if j > i || j > mask_after {
f32::MIN
} else {
0.
}
})
})
.collect();
let mask = Tensor::from_slice(&mask, (seq_len, seq_len), device)?;
mask.broadcast_as((bsz, 1, seq_len, seq_len))
}
pub fn forward_with_mask(&self, input_ids: &Tensor, mask_after: usize) -> Result<Tensor> {
let (bsz, seq_len) = input_ids.dims2()?;
let input_ids = self.embeddings.forward(input_ids)?;
let causal_attention_mask =
Self::build_causal_attention_mask(bsz, seq_len, mask_after, input_ids.device())?;
let input_ids = self
.encoder
.forward(&input_ids, Some(&causal_attention_mask))?;
self.final_layer_norm.forward(&input_ids)
}
}
impl Module for ClipTextTransformer {
fn forward(&self, input_ids: &Tensor) -> Result<Tensor> {
let output = self.forward_with_mask(input_ids, usize::MAX)?;
let sequence_max_indices = input_ids.argmax(D::Minus1)?.to_dtype(DType::I64)?;
let mut indices = Vec::new();
for (batch_idx, &seq_idx) in sequence_max_indices.to_vec1::<i64>()?.iter().enumerate() {
let index = output.i((batch_idx, seq_idx as usize))?.unsqueeze(0)?;
indices.push(index);
}
Tensor::cat(&indices, 0)
}
}
| candle/candle-transformers/src/models/clip/text_model.rs/0 | {
"file_path": "candle/candle-transformers/src/models/clip/text_model.rs",
"repo_id": "candle",
"token_count": 5660
} |
//! Falcon language model inference implementation
//!
//! See ["Falcon: a new approach to large language models"](https://huggingface.co/blog/falcon)
//!
//! Based on implementation from [Huggingface Transformers](https://github.com/huggingface/transformers/blob/main/src/transformers/models/falcon)
use candle::{DType, Device, Result, Tensor, D};
use candle_nn::{embedding, linear_b as linear, Embedding, LayerNorm, Linear, Module, VarBuilder};
use serde::Deserialize;
const MAX_SEQ_LEN: usize = 5000;
fn layer_norm(size: usize, eps: f64, vb: VarBuilder) -> Result<LayerNorm> {
let (weight, bias) = match (vb.get(size, "weight"), vb.get(size, "bias")) {
(Ok(weight), Ok(bias)) => (weight, bias),
(Err(err), _) | (_, Err(err)) => {
if let (Ok(weight), Ok(bias)) = (vb.get(size, "gamma"), vb.get(size, "beta")) {
(weight, bias)
} else {
return Err(err);
}
}
};
Ok(LayerNorm::new(weight, bias, eps))
}
// https://raw.githubusercontent.com/huggingface/transformers/030c863aaa0165e98352b61697430bf69bf33755/src/transformers/models/falcon/configuration_falcon.py
#[derive(Clone, Debug, Deserialize)]
pub struct Config {
pub vocab_size: usize,
pub hidden_size: usize,
pub num_hidden_layers: usize,
pub num_attention_heads: usize,
pub layer_norm_epsilon: f64,
pub initializer_range: f64,
pub use_cache: bool,
pub bos_token_id: u32,
pub eos_token_id: u32,
pub hidden_dropout: f64,
pub attention_dropout: f64,
pub n_head_kv: Option<usize>,
pub alibi: bool,
pub new_decoder_architecture: bool,
pub multi_query: bool,
pub parallel_attn: bool,
pub bias: bool,
}
impl Default for Config {
fn default() -> Self {
Self {
vocab_size: 65024,
hidden_size: 4544,
num_hidden_layers: 32,
num_attention_heads: 71,
layer_norm_epsilon: 1e-5,
initializer_range: 0.02,
use_cache: true,
bos_token_id: 11,
eos_token_id: 11,
hidden_dropout: 0.0,
attention_dropout: 0.0,
n_head_kv: None,
alibi: false,
new_decoder_architecture: false,
multi_query: true,
parallel_attn: true,
bias: false,
}
}
}
impl Config {
pub fn validate(&self) -> Result<()> {
if self.alibi {
candle::bail!("alibi is not supported");
}
if self.new_decoder_architecture {
candle::bail!("new_decoder_architecture is not supported");
}
if self.n_head_kv.is_some() {
candle::bail!("n_head_kv is not supported");
}
Ok(())
}
// https://huggingface.co/tiiuae/falcon-7b/blob/main/config.json
pub fn falcon7b() -> Self {
// This is currently on par with the defaults, the defaults come from the Python default
// arguments for the config initialization whereas the following come from the json config.
Self {
vocab_size: 65024,
hidden_size: 4544,
num_hidden_layers: 32,
num_attention_heads: 71,
layer_norm_epsilon: 1e-5,
initializer_range: 0.02,
use_cache: true,
bos_token_id: 11,
eos_token_id: 11,
hidden_dropout: 0.,
attention_dropout: 0.,
n_head_kv: None,
alibi: false,
new_decoder_architecture: false,
multi_query: true,
parallel_attn: true,
bias: false,
}
}
fn head_dim(&self) -> usize {
self.hidden_size / self.num_attention_heads
}
fn rotary(&self) -> bool {
!self.alibi
}
}
fn rotate_half(x: &Tensor) -> Result<Tensor> {
let l = x.dim(D::Minus1)?;
let x1 = x.narrow(D::Minus1, 0, l / 2)?;
let x2 = x.narrow(D::Minus1, l / 2, l - l / 2)?;
let x21 = Tensor::cat(&[&x2.neg()?, &x1], D::Minus1)?;
Ok(x21)
}
#[derive(Debug, Clone)]
struct FalconRotaryEmbedding {
inv_freq: Tensor,
cache: Option<(usize, Tensor, Tensor)>,
}
impl FalconRotaryEmbedding {
fn load(device: &Device, cfg: &Config) -> Result<Self> {
let head_dim = cfg.head_dim();
let inv_freq: Vec<_> = (0..head_dim)
.step_by(2)
.map(|i| 1f32 / 10000f32.powf(i as f32 / head_dim as f32))
.collect();
Ok(Self {
inv_freq: Tensor::new(inv_freq.as_slice(), device)?,
cache: None,
})
}
fn cos_sin(
&mut self,
seq_len: usize,
device: &Device,
dtype: DType,
) -> Result<(Tensor, Tensor)> {
match &self.cache {
Some((s, cos, sin)) if *s == seq_len => {
return Ok((cos.clone(), sin.clone()));
}
_ => {}
}
let t = Tensor::arange(0, seq_len as u32, device)?.to_dtype(dtype)?;
let inv_freq = self.inv_freq.to_dtype(dtype)?;
let freqs = t.unsqueeze(1)?.matmul(&inv_freq.unsqueeze(0)?)?;
let emb = Tensor::cat(&[&freqs, &freqs], D::Minus1)?;
let cos = emb.cos()?;
let sin = emb.sin()?;
self.cache = Some((seq_len, cos.clone(), sin.clone()));
Ok((cos, sin))
}
fn forward(
&mut self,
query: &Tensor,
key: &Tensor,
past_kv_len: usize,
) -> Result<(Tensor, Tensor)> {
let (_batch, seq_len, _head_dim) = query.dims3()?;
let (cos, sin) = self.cos_sin(MAX_SEQ_LEN, query.device(), query.dtype())?;
let cos = cos.narrow(0, past_kv_len, seq_len)?;
let sin = sin.narrow(0, past_kv_len, seq_len)?;
let qs = (query.broadcast_mul(&cos)? + &rotate_half(query)?.broadcast_mul(&sin)?)?;
let ks = (key.broadcast_mul(&cos)? + &rotate_half(key)?.broadcast_mul(&sin)?)?;
Ok((qs, ks))
}
}
fn masked_fill(on_false: &Tensor, mask: &Tensor, on_true: f32) -> Result<Tensor> {
let shape = mask.shape();
let on_true = Tensor::new(on_true, on_false.device())?
.to_dtype(on_false.dtype())?
.broadcast_as(shape.dims())?;
let m = mask.where_cond(&on_true, on_false)?;
Ok(m)
}
#[derive(Debug, Clone)]
struct FalconAttention {
query_key_value: Linear,
dense: Linear,
maybe_rotary: Option<FalconRotaryEmbedding>,
kv_cache: Option<(Tensor, Tensor)>,
inv_norm_factor: f64,
multi_query: bool,
use_cache: bool,
num_heads: usize,
head_dim: usize,
n_head_kv: usize,
}
impl FalconAttention {
fn load(vb: VarBuilder, cfg: &Config) -> Result<Self> {
let maybe_rotary = if cfg.rotary() {
let rotary = FalconRotaryEmbedding::load(vb.device(), cfg)?;
Some(rotary)
} else {
None
};
let head_dim = cfg.head_dim();
let hidden_size = cfg.hidden_size;
let qkv_out_dim = if cfg.multi_query {
hidden_size + 2 * head_dim
} else {
3 * hidden_size
};
let query_key_value = linear(hidden_size, qkv_out_dim, cfg.bias, vb.pp("query_key_value"))?;
let dense = linear(hidden_size, hidden_size, cfg.bias, vb.pp("dense"))?;
Ok(Self {
query_key_value,
dense,
maybe_rotary,
kv_cache: None,
inv_norm_factor: 1. / (head_dim as f64).sqrt(),
multi_query: cfg.multi_query,
use_cache: cfg.use_cache,
num_heads: cfg.num_attention_heads,
n_head_kv: cfg.n_head_kv.unwrap_or(1),
head_dim,
})
}
fn split_heads(&self, fused_qkv: &Tensor) -> Result<(Tensor, Tensor, Tensor)> {
let (b_sz, seq_len, _) = fused_qkv.dims3()?;
if !self.multi_query {
let fused_qkv = fused_qkv.reshape((b_sz, seq_len, self.num_heads, 3, self.head_dim))?;
let q = fused_qkv.narrow(D::Minus2, 0, 1)?.squeeze(D::Minus2)?;
let k = fused_qkv.narrow(D::Minus2, 1, 1)?.squeeze(D::Minus2)?;
let v = fused_qkv.narrow(D::Minus2, 2, 1)?.squeeze(D::Minus2)?;
Ok((q, k, v))
} else {
let fused_qkv =
fused_qkv.reshape((b_sz, seq_len, self.num_heads + 2, self.head_dim))?;
let d = fused_qkv.dim(D::Minus2)?;
let q = fused_qkv.narrow(D::Minus2, 0, d - 2)?;
let k = fused_qkv.narrow(D::Minus2, d - 2, 1)?;
let v = fused_qkv.narrow(D::Minus2, d - 1, 1)?;
Ok((q, k, v))
}
}
fn forward(&mut self, x: &Tensor, mask: Option<&Tensor>, past_kv_len: usize) -> Result<Tensor> {
let fused_qkv = self.query_key_value.forward(x)?;
let head_dim = self.head_dim;
let (query, key, value) = self.split_heads(&fused_qkv)?;
let (b_sz, seq_len, _, _) = query.dims4()?;
let query = query
.transpose(1, 2)?
.reshape((b_sz * self.num_heads, seq_len, head_dim))?;
let key = key
.transpose(1, 2)?
.reshape((b_sz * self.n_head_kv, seq_len, head_dim))?;
let value = value
.transpose(1, 2)?
.reshape((b_sz * self.n_head_kv, seq_len, head_dim))?;
let (query, key) = if let Some(r) = &mut self.maybe_rotary {
r.forward(&query, &key, past_kv_len)?
} else {
(query, key)
};
let (mut key, mut value) = (key, value);
if self.use_cache {
if let Some((cache_k, cache_v)) = &self.kv_cache {
// TODO: we could trim the tensors to MAX_SEQ_LEN so that this would work for
// arbitrarily large sizes.
key = Tensor::cat(&[cache_k, &key], 1)?.contiguous()?;
value = Tensor::cat(&[cache_v, &value], 1)?.contiguous()?;
}
self.kv_cache = Some((key.clone(), value.clone()))
}
let query = query.reshape((b_sz * self.num_heads, seq_len, head_dim))?;
let all_len = past_kv_len + seq_len;
let key = key.reshape((b_sz * self.n_head_kv, all_len, head_dim))?;
let value = value.reshape((b_sz * self.n_head_kv, all_len, head_dim))?;
let (key, value) = if self.n_head_kv == 1 {
(
key.broadcast_as((b_sz * self.num_heads, all_len, head_dim))?,
value.broadcast_as((b_sz * self.num_heads, all_len, head_dim))?,
)
} else {
(key, value)
};
// Only handle the case where alibi is None here, and non-flash attention.
let attention_scores = (query.matmul(&key.t()?)? * self.inv_norm_factor)?;
let attention_scores = match mask {
None => attention_scores,
Some(mask) => {
let mask = masked_fill(&mask.to_dtype(DType::F32)?, mask, -1e9)?
.to_dtype(query.dtype())?;
attention_scores.broadcast_add(&mask.squeeze(1)?)?
}
};
let attention_scores =
candle_nn::ops::softmax(&attention_scores.to_dtype(DType::F32)?, D::Minus1)?
.to_dtype(x.dtype())?;
let attn_output = attention_scores
.matmul(&value)?
.reshape((b_sz, self.num_heads, seq_len, head_dim))?
.transpose(1, 2)?
.reshape((b_sz, seq_len, self.num_heads * head_dim))?;
let attn_output = self.dense.forward(&attn_output)?;
Ok(attn_output)
}
fn clear_kv_cache(&mut self) {
self.kv_cache = None
}
}
#[derive(Debug, Clone)]
struct FalconMlp {
dense_h_to_4h: Linear,
dense_4h_to_h: Linear,
}
impl FalconMlp {
fn load(vb: VarBuilder, cfg: &Config) -> Result<Self> {
let h = cfg.hidden_size;
let b = cfg.bias;
let dense_h_to_4h = linear(h, 4 * h, b, vb.pp("dense_h_to_4h"))?;
let dense_4h_to_h = linear(4 * h, h, b, vb.pp("dense_4h_to_h"))?;
Ok(Self {
dense_h_to_4h,
dense_4h_to_h,
})
}
fn forward(&self, x: &Tensor) -> Result<Tensor> {
let x = self.dense_h_to_4h.forward(x)?.gelu()?;
let x = self.dense_4h_to_h.forward(&x)?;
Ok(x)
}
}
#[derive(Debug, Clone)]
struct FalconDecoderLayer {
inp_layernorm: LayerNorm,
self_attention: FalconAttention,
post_attention_layernorm: Option<LayerNorm>,
mlp: FalconMlp,
parallel_attn: bool,
}
impl FalconDecoderLayer {
fn load(vb: VarBuilder, cfg: &Config) -> Result<Self> {
let mlp = FalconMlp::load(vb.pp("mlp"), cfg)?;
let inp_layernorm = layer_norm(
cfg.hidden_size,
cfg.layer_norm_epsilon,
vb.pp("input_layernorm"),
)?;
let self_attention = FalconAttention::load(vb.pp("self_attention"), cfg)?;
let post_attention_layernorm = if cfg.parallel_attn {
None
} else {
let ln = layer_norm(
cfg.hidden_size,
cfg.layer_norm_epsilon,
vb.pp("post_attention_layernorm"),
)?;
Some(ln)
};
Ok(Self {
inp_layernorm,
self_attention,
post_attention_layernorm,
mlp,
parallel_attn: cfg.parallel_attn,
})
}
fn forward(&mut self, x: &Tensor, mask: Option<&Tensor>, past_kv_len: usize) -> Result<Tensor> {
let residual = x.clone();
let ln_attn = self.inp_layernorm.forward(x)?;
let attn_output = self.self_attention.forward(&ln_attn, mask, past_kv_len)?;
let (residual, ln_mlp) = match &self.post_attention_layernorm {
None => (residual, ln_attn),
Some(pal) => {
// This should include some dropout.
let residual = (&attn_output + &residual)?;
let ln_mlp = pal.forward(&residual)?;
(residual, ln_mlp)
}
};
let mlp_output = self.mlp.forward(&ln_mlp)?;
let mlp_output = if self.parallel_attn {
(mlp_output + attn_output)?
} else {
mlp_output
};
let output = (mlp_output + residual)?;
Ok(output)
}
pub fn clear_kv_cache(&mut self) {
self.self_attention.clear_kv_cache()
}
}
#[derive(Debug, Clone)]
pub struct Falcon {
word_embeddings: Embedding,
blocks: Vec<FalconDecoderLayer>,
ln_f: LayerNorm,
lm_head: Linear,
config: Config,
}
fn make_causal_mask(t: usize) -> Result<Tensor> {
let mask: Vec<_> = (0..t)
.flat_map(|i| (0..t).map(move |j| u8::from(j > i)))
.collect();
let mask = Tensor::from_slice(&mask, (t, t), &Device::Cpu)?;
Ok(mask)
}
fn prepare_attn_mask(b_sz: usize, seq_len: usize) -> Result<Tensor> {
// let mask = Tensor::ones((b_sz, seq_len), DType::U32, &Device::Cpu)?;
let mask = make_causal_mask(seq_len)?;
let mask = mask.broadcast_as((b_sz, 1, seq_len, seq_len))?;
Ok(mask)
}
impl Falcon {
pub fn config(&self) -> &Config {
&self.config
}
pub fn load(vb: VarBuilder, cfg: Config) -> Result<Self> {
let word_embeddings = embedding(
cfg.vocab_size,
cfg.hidden_size,
vb.pp("transformer.word_embeddings"),
)?;
let blocks = (0..cfg.num_hidden_layers)
.map(|i| FalconDecoderLayer::load(vb.pp(format!("transformer.h.{i}")), &cfg))
.collect::<Result<Vec<_>>>()?;
let ln_f = layer_norm(
cfg.hidden_size,
cfg.layer_norm_epsilon,
vb.pp("transformer.ln_f"),
)?;
let lm_head = linear(cfg.hidden_size, cfg.vocab_size, false, vb.pp("lm_head"))?;
Ok(Self {
word_embeddings,
blocks,
ln_f,
lm_head,
config: cfg,
})
}
pub fn forward(&mut self, input_ids: &Tensor) -> Result<Tensor> {
let (b_sz, seq_len) = input_ids.dims2()?;
let mut hidden_state = self.word_embeddings.forward(input_ids)?;
let past_kv_len = match &self.blocks[0].self_attention.kv_cache {
Some((k, _)) => k.dim(1)?,
None => 0,
};
let causal_mask = if seq_len <= 1 {
None
} else {
Some(prepare_attn_mask(b_sz, seq_len)?.to_device(input_ids.device())?)
};
for block in self.blocks.iter_mut() {
hidden_state = block.forward(&hidden_state, causal_mask.as_ref(), past_kv_len)?;
}
let hidden_state = self.ln_f.forward(&hidden_state)?;
let hidden_state = hidden_state.narrow(1, seq_len - 1, 1)?;
let logits = self.lm_head.forward(&hidden_state)?.squeeze(1)?;
Ok(logits)
}
pub fn clear_kv_cache(&mut self) {
for block in self.blocks.iter_mut() {
block.clear_kv_cache()
}
}
}
| candle/candle-transformers/src/models/falcon.rs/0 | {
"file_path": "candle/candle-transformers/src/models/falcon.rs",
"repo_id": "candle",
"token_count": 8880
} |
//! Llama2 inference implementation.
//!
//! See ["LLaMA 2: Open Foundation and Fine-Tuned Chat Models"](https://arxiv.org/abs/2307.09288)
//!
//! Based on the [llama2.c](https://github.com/karpathy/llama2.c) implementation
use byteorder::{LittleEndian, ReadBytesExt};
use candle::{DType, Device, IndexOp, Result, Shape, Tensor};
use candle_nn::VarBuilder;
use super::llama2_c::Config;
pub struct TransformerWeights {
// token embedding table
token_embedding_table: Tensor, // (vocab_size, dim)
// weights for rmsnorms
rms_att_weight: Tensor, // (layer, dim) rmsnorm weights
rms_ffn_weight: Tensor, // (layer, dim)
// weights for matmuls
wq: Tensor, // (layer, dim, dim)
wk: Tensor, // (layer, dim, dim)
wv: Tensor, // (layer, dim, dim)
wo: Tensor, // (layer, dim, dim)
// weights for ffn
w1: Tensor, // (layer, hidden_dim, dim)
w2: Tensor, // (layer, dim, hidden_dim)
w3: Tensor, // (layer, hidden_dim, dim)
// final rmsnorm
rms_final_weight: Tensor, // (dim,)
// freq_cis for RoPE relatively positional embeddings
freq_cis_real: Tensor, // (seq_len, head_size/2)
freq_cis_imag: Tensor, // (seq_len, head_size/2)
}
fn read_i32<R: std::io::Read>(r: &mut R) -> Result<i32> {
let mut buf = [0u8; 4];
r.read_exact(&mut buf)?;
Ok(i32::from_le_bytes(buf))
}
fn read_tensor<R: std::io::Read, S: Into<Shape>>(
r: &mut R,
shape: S,
dev: &Device,
) -> Result<Tensor> {
let shape = shape.into();
let mut data_t = vec![0f32; shape.elem_count()];
r.read_f32_into::<LittleEndian>(&mut data_t)?;
let tensor = Tensor::from_vec(data_t, shape, dev)?;
Ok(tensor)
}
impl Config {
pub fn from_reader<R: std::io::Read>(r: &mut R) -> Result<Self> {
let dim = read_i32(r)? as usize;
let hidden_dim = read_i32(r)? as usize;
let n_layers = read_i32(r)? as usize;
let n_heads = read_i32(r)? as usize;
let n_kv_heads = read_i32(r)? as usize;
let vocab_size = read_i32(r)? as usize;
let seq_len = read_i32(r)? as usize;
Ok(Self {
dim,
hidden_dim,
n_layers,
n_heads,
n_kv_heads,
vocab_size,
seq_len,
norm_eps: 1e-5,
})
}
pub fn head_size(&self) -> usize {
self.dim / self.n_heads
}
}
impl TransformerWeights {
pub fn from_reader<R: std::io::Read>(r: &mut R, c: &Config, dev: &Device) -> Result<Self> {
let token_embedding_table = read_tensor(r, (c.vocab_size, c.dim), dev)?;
let rms_att_weight = read_tensor(r, (c.n_layers, c.dim), dev)?;
let wq = read_tensor(r, (c.n_layers, c.dim, c.dim), dev)?;
let wk = read_tensor(r, (c.n_layers, c.dim, c.dim), dev)?;
let wv = read_tensor(r, (c.n_layers, c.dim, c.dim), dev)?;
let wo = read_tensor(r, (c.n_layers, c.dim, c.dim), dev)?;
let rms_ffn_weight = read_tensor(r, (c.n_layers, c.dim), dev)?;
let w1 = read_tensor(r, (c.n_layers, c.hidden_dim, c.dim), dev)?;
let w2 = read_tensor(r, (c.n_layers, c.dim, c.hidden_dim), dev)?;
let w3 = read_tensor(r, (c.n_layers, c.hidden_dim, c.dim), dev)?;
let rms_final_weight = read_tensor(r, c.dim, dev)?;
let head_size = c.head_size();
let freq_cis_real = read_tensor(r, (c.seq_len, head_size / 2), dev)?;
let freq_cis_imag = read_tensor(r, (c.seq_len, head_size / 2), dev)?;
Ok(Self {
token_embedding_table,
rms_att_weight,
wq,
wk,
wv,
wo,
rms_ffn_weight,
w1,
w2,
w3,
rms_final_weight,
freq_cis_real,
freq_cis_imag,
})
}
pub fn var_builder(&self, cfg: &Config, device: &Device) -> Result<VarBuilder<'static>> {
// TODO: As of 2023-08-04, gemm is slower than expected when multiplying a matrix of
// size (1, k) with the transpose of a matrix of size (k, n) as it ends up transposing the
// second matrix back. We detect this case here and as a temporary hack make the weight
// matrix column major rather than row major. This ends up speeding up text generation from
// 120 token/s to 220 token/s on a Ryzen 2600X.
let tr = device.is_cpu() && !candle::utils::has_mkl();
let tr = |x: Tensor| if tr { x.t()?.contiguous()?.t() } else { Ok(x) };
let mut ws = std::collections::HashMap::new();
let mut insert = |name: &str, t: Tensor| {
ws.insert(name.to_string(), t);
};
insert("rot.freq_cis_real", self.freq_cis_real.clone());
insert("rot.freq_cis_imag", self.freq_cis_imag.clone());
insert(
"model.embed_tokens.weight",
self.token_embedding_table.clone(),
);
insert("lm_head.weight", tr(self.token_embedding_table.clone())?);
insert("model.norm.weight", self.rms_final_weight.clone());
for layer in 0..cfg.n_layers {
ws.insert(
format!("model.layers.{layer}.self_attn.q_proj.weight"),
tr(self.wq.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.self_attn.k_proj.weight"),
tr(self.wk.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.self_attn.v_proj.weight"),
tr(self.wv.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.self_attn.o_proj.weight"),
tr(self.wo.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.mlp.gate_proj.weight"),
tr(self.w1.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.mlp.down_proj.weight"),
tr(self.w2.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.mlp.up_proj.weight"),
tr(self.w3.i(layer)?)?,
);
ws.insert(
format!("model.layers.{layer}.input_layernorm.weight"),
self.rms_att_weight.i(layer)?,
);
ws.insert(
format!("model.layers.{layer}.post_attention_layernorm.weight"),
self.rms_ffn_weight.i(layer)?,
);
}
let vb = VarBuilder::from_tensors(ws, DType::F32, device);
Ok(vb)
}
}
| candle/candle-transformers/src/models/llama2_c_weights.rs/0 | {
"file_path": "candle/candle-transformers/src/models/llama2_c_weights.rs",
"repo_id": "candle",
"token_count": 3405
} |
use candle::{Module, Result, Tensor, D};
use candle_nn as nn;
use super::projections::{AttnProjections, Mlp, Qkv, QkvOnlyAttnProjections};
pub struct ModulateIntermediates {
gate_msa: Tensor,
shift_mlp: Tensor,
scale_mlp: Tensor,
gate_mlp: Tensor,
}
pub struct DiTBlock {
norm1: LayerNormNoAffine,
attn: AttnProjections,
norm2: LayerNormNoAffine,
mlp: Mlp,
ada_ln_modulation: nn::Sequential,
}
pub struct LayerNormNoAffine {
eps: f64,
}
impl LayerNormNoAffine {
pub fn new(eps: f64) -> Self {
Self { eps }
}
}
impl Module for LayerNormNoAffine {
fn forward(&self, x: &Tensor) -> Result<Tensor> {
nn::LayerNorm::new_no_bias(Tensor::ones_like(x)?, self.eps).forward(x)
}
}
impl DiTBlock {
pub fn new(hidden_size: usize, num_heads: usize, vb: nn::VarBuilder) -> Result<Self> {
let norm1 = LayerNormNoAffine::new(1e-6);
let attn = AttnProjections::new(hidden_size, num_heads, vb.pp("attn"))?;
let norm2 = LayerNormNoAffine::new(1e-6);
let mlp_ratio = 4;
let mlp = Mlp::new(hidden_size, hidden_size * mlp_ratio, vb.pp("mlp"))?;
let n_mods = 6;
let ada_ln_modulation = nn::seq().add(nn::Activation::Silu).add(nn::linear(
hidden_size,
n_mods * hidden_size,
vb.pp("adaLN_modulation.1"),
)?);
Ok(Self {
norm1,
attn,
norm2,
mlp,
ada_ln_modulation,
})
}
pub fn pre_attention(&self, x: &Tensor, c: &Tensor) -> Result<(Qkv, ModulateIntermediates)> {
let modulation = self.ada_ln_modulation.forward(c)?;
let chunks = modulation.chunk(6, D::Minus1)?;
let (shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp) = (
chunks[0].clone(),
chunks[1].clone(),
chunks[2].clone(),
chunks[3].clone(),
chunks[4].clone(),
chunks[5].clone(),
);
let norm_x = self.norm1.forward(x)?;
let modulated_x = modulate(&norm_x, &shift_msa, &scale_msa)?;
let qkv = self.attn.pre_attention(&modulated_x)?;
Ok((
qkv,
ModulateIntermediates {
gate_msa,
shift_mlp,
scale_mlp,
gate_mlp,
},
))
}
pub fn post_attention(
&self,
attn: &Tensor,
x: &Tensor,
mod_interm: &ModulateIntermediates,
) -> Result<Tensor> {
let attn_out = self.attn.post_attention(attn)?;
let x = x.add(&attn_out.broadcast_mul(&mod_interm.gate_msa.unsqueeze(1)?)?)?;
let norm_x = self.norm2.forward(&x)?;
let modulated_x = modulate(&norm_x, &mod_interm.shift_mlp, &mod_interm.scale_mlp)?;
let mlp_out = self.mlp.forward(&modulated_x)?;
let x = x.add(&mlp_out.broadcast_mul(&mod_interm.gate_mlp.unsqueeze(1)?)?)?;
Ok(x)
}
}
pub struct SelfAttnModulateIntermediates {
gate_msa: Tensor,
shift_mlp: Tensor,
scale_mlp: Tensor,
gate_mlp: Tensor,
gate_msa2: Tensor,
}
pub struct SelfAttnDiTBlock {
norm1: LayerNormNoAffine,
attn: AttnProjections,
attn2: AttnProjections,
norm2: LayerNormNoAffine,
mlp: Mlp,
ada_ln_modulation: nn::Sequential,
}
impl SelfAttnDiTBlock {
pub fn new(hidden_size: usize, num_heads: usize, vb: nn::VarBuilder) -> Result<Self> {
let norm1 = LayerNormNoAffine::new(1e-6);
let attn = AttnProjections::new(hidden_size, num_heads, vb.pp("attn"))?;
let attn2 = AttnProjections::new(hidden_size, num_heads, vb.pp("attn2"))?;
let norm2 = LayerNormNoAffine::new(1e-6);
let mlp_ratio = 4;
let mlp = Mlp::new(hidden_size, hidden_size * mlp_ratio, vb.pp("mlp"))?;
let n_mods = 9;
let ada_ln_modulation = nn::seq().add(nn::Activation::Silu).add(nn::linear(
hidden_size,
n_mods * hidden_size,
vb.pp("adaLN_modulation.1"),
)?);
Ok(Self {
norm1,
attn,
attn2,
norm2,
mlp,
ada_ln_modulation,
})
}
pub fn pre_attention(
&self,
x: &Tensor,
c: &Tensor,
) -> Result<(Qkv, Qkv, SelfAttnModulateIntermediates)> {
let modulation = self.ada_ln_modulation.forward(c)?;
let chunks = modulation.chunk(9, D::Minus1)?;
let (
shift_msa,
scale_msa,
gate_msa,
shift_mlp,
scale_mlp,
gate_mlp,
shift_msa2,
scale_msa2,
gate_msa2,
) = (
chunks[0].clone(),
chunks[1].clone(),
chunks[2].clone(),
chunks[3].clone(),
chunks[4].clone(),
chunks[5].clone(),
chunks[6].clone(),
chunks[7].clone(),
chunks[8].clone(),
);
let norm_x = self.norm1.forward(x)?;
let modulated_x = modulate(&norm_x, &shift_msa, &scale_msa)?;
let qkv = self.attn.pre_attention(&modulated_x)?;
let modulated_x2 = modulate(&norm_x, &shift_msa2, &scale_msa2)?;
let qkv2 = self.attn2.pre_attention(&modulated_x2)?;
Ok((
qkv,
qkv2,
SelfAttnModulateIntermediates {
gate_msa,
shift_mlp,
scale_mlp,
gate_mlp,
gate_msa2,
},
))
}
pub fn post_attention(
&self,
attn: &Tensor,
attn2: &Tensor,
x: &Tensor,
mod_interm: &SelfAttnModulateIntermediates,
) -> Result<Tensor> {
let attn_out = self.attn.post_attention(attn)?;
let x = x.add(&attn_out.broadcast_mul(&mod_interm.gate_msa.unsqueeze(1)?)?)?;
let attn_out2 = self.attn2.post_attention(attn2)?;
let x = x.add(&attn_out2.broadcast_mul(&mod_interm.gate_msa2.unsqueeze(1)?)?)?;
let norm_x = self.norm2.forward(&x)?;
let modulated_x = modulate(&norm_x, &mod_interm.shift_mlp, &mod_interm.scale_mlp)?;
let mlp_out = self.mlp.forward(&modulated_x)?;
let x = x.add(&mlp_out.broadcast_mul(&mod_interm.gate_mlp.unsqueeze(1)?)?)?;
Ok(x)
}
}
pub struct QkvOnlyDiTBlock {
norm1: LayerNormNoAffine,
attn: QkvOnlyAttnProjections,
ada_ln_modulation: nn::Sequential,
}
impl QkvOnlyDiTBlock {
pub fn new(hidden_size: usize, num_heads: usize, vb: nn::VarBuilder) -> Result<Self> {
let norm1 = LayerNormNoAffine::new(1e-6);
let attn = QkvOnlyAttnProjections::new(hidden_size, num_heads, vb.pp("attn"))?;
let n_mods = 2;
let ada_ln_modulation = nn::seq().add(nn::Activation::Silu).add(nn::linear(
hidden_size,
n_mods * hidden_size,
vb.pp("adaLN_modulation.1"),
)?);
Ok(Self {
norm1,
attn,
ada_ln_modulation,
})
}
pub fn pre_attention(&self, x: &Tensor, c: &Tensor) -> Result<Qkv> {
let modulation = self.ada_ln_modulation.forward(c)?;
let chunks = modulation.chunk(2, D::Minus1)?;
let (shift_msa, scale_msa) = (chunks[0].clone(), chunks[1].clone());
let norm_x = self.norm1.forward(x)?;
let modulated_x = modulate(&norm_x, &shift_msa, &scale_msa)?;
self.attn.pre_attention(&modulated_x)
}
}
pub struct FinalLayer {
norm_final: LayerNormNoAffine,
linear: nn::Linear,
ada_ln_modulation: nn::Sequential,
}
impl FinalLayer {
pub fn new(
hidden_size: usize,
patch_size: usize,
out_channels: usize,
vb: nn::VarBuilder,
) -> Result<Self> {
let norm_final = LayerNormNoAffine::new(1e-6);
let linear = nn::linear(
hidden_size,
patch_size * patch_size * out_channels,
vb.pp("linear"),
)?;
let ada_ln_modulation = nn::seq().add(nn::Activation::Silu).add(nn::linear(
hidden_size,
2 * hidden_size,
vb.pp("adaLN_modulation.1"),
)?);
Ok(Self {
norm_final,
linear,
ada_ln_modulation,
})
}
pub fn forward(&self, x: &Tensor, c: &Tensor) -> Result<Tensor> {
let modulation = self.ada_ln_modulation.forward(c)?;
let chunks = modulation.chunk(2, D::Minus1)?;
let (shift, scale) = (chunks[0].clone(), chunks[1].clone());
let norm_x = self.norm_final.forward(x)?;
let modulated_x = modulate(&norm_x, &shift, &scale)?;
let output = self.linear.forward(&modulated_x)?;
Ok(output)
}
}
fn modulate(x: &Tensor, shift: &Tensor, scale: &Tensor) -> Result<Tensor> {
let shift = shift.unsqueeze(1)?;
let scale = scale.unsqueeze(1)?;
let scale_plus_one = scale.add(&Tensor::ones_like(&scale)?)?;
shift.broadcast_add(&x.broadcast_mul(&scale_plus_one)?)
}
pub trait JointBlock {
fn forward(&self, context: &Tensor, x: &Tensor, c: &Tensor) -> Result<(Tensor, Tensor)>;
}
pub struct MMDiTJointBlock {
x_block: DiTBlock,
context_block: DiTBlock,
num_heads: usize,
use_flash_attn: bool,
}
impl MMDiTJointBlock {
pub fn new(
hidden_size: usize,
num_heads: usize,
use_flash_attn: bool,
vb: nn::VarBuilder,
) -> Result<Self> {
let x_block = DiTBlock::new(hidden_size, num_heads, vb.pp("x_block"))?;
let context_block = DiTBlock::new(hidden_size, num_heads, vb.pp("context_block"))?;
Ok(Self {
x_block,
context_block,
num_heads,
use_flash_attn,
})
}
}
impl JointBlock for MMDiTJointBlock {
fn forward(&self, context: &Tensor, x: &Tensor, c: &Tensor) -> Result<(Tensor, Tensor)> {
let (context_qkv, context_interm) = self.context_block.pre_attention(context, c)?;
let (x_qkv, x_interm) = self.x_block.pre_attention(x, c)?;
let (context_attn, x_attn) =
joint_attn(&context_qkv, &x_qkv, self.num_heads, self.use_flash_attn)?;
let context_out =
self.context_block
.post_attention(&context_attn, context, &context_interm)?;
let x_out = self.x_block.post_attention(&x_attn, x, &x_interm)?;
Ok((context_out, x_out))
}
}
pub struct MMDiTXJointBlock {
x_block: SelfAttnDiTBlock,
context_block: DiTBlock,
num_heads: usize,
use_flash_attn: bool,
}
impl MMDiTXJointBlock {
pub fn new(
hidden_size: usize,
num_heads: usize,
use_flash_attn: bool,
vb: nn::VarBuilder,
) -> Result<Self> {
let x_block = SelfAttnDiTBlock::new(hidden_size, num_heads, vb.pp("x_block"))?;
let context_block = DiTBlock::new(hidden_size, num_heads, vb.pp("context_block"))?;
Ok(Self {
x_block,
context_block,
num_heads,
use_flash_attn,
})
}
}
impl JointBlock for MMDiTXJointBlock {
fn forward(&self, context: &Tensor, x: &Tensor, c: &Tensor) -> Result<(Tensor, Tensor)> {
let (context_qkv, context_interm) = self.context_block.pre_attention(context, c)?;
let (x_qkv, x_qkv2, x_interm) = self.x_block.pre_attention(x, c)?;
let (context_attn, x_attn) =
joint_attn(&context_qkv, &x_qkv, self.num_heads, self.use_flash_attn)?;
let x_attn2 = attn(&x_qkv2, self.num_heads, self.use_flash_attn)?;
let context_out =
self.context_block
.post_attention(&context_attn, context, &context_interm)?;
let x_out = self
.x_block
.post_attention(&x_attn, &x_attn2, x, &x_interm)?;
Ok((context_out, x_out))
}
}
pub struct ContextQkvOnlyJointBlock {
x_block: DiTBlock,
context_block: QkvOnlyDiTBlock,
num_heads: usize,
use_flash_attn: bool,
}
impl ContextQkvOnlyJointBlock {
pub fn new(
hidden_size: usize,
num_heads: usize,
use_flash_attn: bool,
vb: nn::VarBuilder,
) -> Result<Self> {
let x_block = DiTBlock::new(hidden_size, num_heads, vb.pp("x_block"))?;
let context_block = QkvOnlyDiTBlock::new(hidden_size, num_heads, vb.pp("context_block"))?;
Ok(Self {
x_block,
context_block,
num_heads,
use_flash_attn,
})
}
pub fn forward(&self, context: &Tensor, x: &Tensor, c: &Tensor) -> Result<Tensor> {
let context_qkv = self.context_block.pre_attention(context, c)?;
let (x_qkv, x_interm) = self.x_block.pre_attention(x, c)?;
let (_, x_attn) = joint_attn(&context_qkv, &x_qkv, self.num_heads, self.use_flash_attn)?;
let x_out = self.x_block.post_attention(&x_attn, x, &x_interm)?;
Ok(x_out)
}
}
// A QKV-attention that is compatible with the interface of candle_flash_attn::flash_attn
// Flash attention regards q, k, v dimensions as (batch_size, seqlen, nheads, headdim)
fn flash_compatible_attention(
q: &Tensor,
k: &Tensor,
v: &Tensor,
softmax_scale: f32,
) -> Result<Tensor> {
let q_dims_for_matmul = q.transpose(1, 2)?.dims().to_vec();
let rank = q_dims_for_matmul.len();
let q = q.transpose(1, 2)?.flatten_to(rank - 3)?;
let k = k.transpose(1, 2)?.flatten_to(rank - 3)?;
let v = v.transpose(1, 2)?.flatten_to(rank - 3)?;
let attn_weights = (q.matmul(&k.t()?)? * softmax_scale as f64)?;
let attn_scores = candle_nn::ops::softmax_last_dim(&attn_weights)?.matmul(&v)?;
attn_scores.reshape(q_dims_for_matmul)?.transpose(1, 2)
}
#[cfg(feature = "flash-attn")]
fn flash_attn(
q: &Tensor,
k: &Tensor,
v: &Tensor,
softmax_scale: f32,
causal: bool,
) -> Result<Tensor> {
candle_flash_attn::flash_attn(q, k, v, softmax_scale, causal)
}
#[cfg(not(feature = "flash-attn"))]
fn flash_attn(_: &Tensor, _: &Tensor, _: &Tensor, _: f32, _: bool) -> Result<Tensor> {
unimplemented!("compile with '--features flash-attn'")
}
fn joint_attn(
context_qkv: &Qkv,
x_qkv: &Qkv,
num_heads: usize,
use_flash_attn: bool,
) -> Result<(Tensor, Tensor)> {
let qkv = Qkv {
q: Tensor::cat(&[&context_qkv.q, &x_qkv.q], 1)?,
k: Tensor::cat(&[&context_qkv.k, &x_qkv.k], 1)?,
v: Tensor::cat(&[&context_qkv.v, &x_qkv.v], 1)?,
};
let seqlen = qkv.q.dim(1)?;
let attn = attn(&qkv, num_heads, use_flash_attn)?;
let context_qkv_seqlen = context_qkv.q.dim(1)?;
let context_attn = attn.narrow(1, 0, context_qkv_seqlen)?;
let x_attn = attn.narrow(1, context_qkv_seqlen, seqlen - context_qkv_seqlen)?;
Ok((context_attn, x_attn))
}
fn attn(qkv: &Qkv, num_heads: usize, use_flash_attn: bool) -> Result<Tensor> {
let batch_size = qkv.q.dim(0)?;
let seqlen = qkv.q.dim(1)?;
let qkv = Qkv {
q: qkv.q.reshape((batch_size, seqlen, num_heads, ()))?,
k: qkv.k.reshape((batch_size, seqlen, num_heads, ()))?,
v: qkv.v.clone(),
};
let headdim = qkv.q.dim(D::Minus1)?;
let softmax_scale = 1.0 / (headdim as f64).sqrt();
let attn = if use_flash_attn {
flash_attn(&qkv.q, &qkv.k, &qkv.v, softmax_scale as f32, false)?
} else {
flash_compatible_attention(&qkv.q, &qkv.k, &qkv.v, softmax_scale as f32)?
};
attn.reshape((batch_size, seqlen, ()))
}
| candle/candle-transformers/src/models/mmdit/blocks.rs/0 | {
"file_path": "candle/candle-transformers/src/models/mmdit/blocks.rs",
"repo_id": "candle",
"token_count": 8057
} |
//! Open Contrastive Language-Image Pre-Training
//!
//! Open Contrastive Language-Image Pre-Training (OpenCLIP) is an architecture trained on
//! pairs of images with related texts.
//!
//! - 💻 [GH Link](https://github.com/mlfoundations/open_clip)
//! - 📝 [Paper](https://arxiv.org/abs/2212.07143)
//!
//! ## Overview
//!
//! 
pub mod text_model;
| candle/candle-transformers/src/models/openclip/mod.rs/0 | {
"file_path": "candle/candle-transformers/src/models/openclip/mod.rs",
"repo_id": "candle",
"token_count": 154
} |
//! Module containing quantized MixFormer model implementation.
//!
//! MixFormer is an efficient transformer variant for text generation that uses
//! mixture-of-experts and parallel attention/feed-forward blocks.
//! This implementation provides quantization for reduced memory usage.
//!
//! Key features:
//! - Parallel attention and feed-forward computation
//! - Rotary positional embeddings
//! - Optional key-value caching
//! - Support for 8-bit quantization
//!
use crate::quantized_nn::{layer_norm, linear, Linear};
pub use crate::quantized_var_builder::VarBuilder;
use candle::{DType, Device, IndexOp, Module, Result, Tensor, D};
use candle_nn::Activation;
pub use crate::models::mixformer::Config;
const MAX_SEQ_LEN: usize = 4096;
#[derive(Debug, Clone)]
struct Embedding {
wte: crate::quantized_nn::Embedding,
}
impl Embedding {
fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let wte = crate::quantized_nn::Embedding::new(cfg.vocab_size, cfg.n_embd, vb.pp("wte"))?;
Ok(Self { wte })
}
}
impl Module for Embedding {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
self.wte.forward(xs)
}
}
fn get_mask(size: usize, device: &Device) -> Result<Tensor> {
let mask: Vec<_> = (0..size)
.flat_map(|i| (0..size).map(move |j| u8::from(j > i)))
.collect();
Tensor::from_slice(&mask, (size, size), device)
}
fn masked_fill(on_false: &Tensor, mask: &Tensor, on_true: f32) -> Result<Tensor> {
let shape = mask.shape();
let on_true = Tensor::new(on_true, on_false.device())?.broadcast_as(shape.dims())?;
let m = mask.where_cond(&on_true, on_false)?;
Ok(m)
}
#[derive(Debug, Clone)]
struct RotaryEmbedding {
sin: Tensor,
cos: Tensor,
}
impl RotaryEmbedding {
fn new(dim: usize, max_seq_len: usize, dev: &Device) -> Result<Self> {
let inv_freq: Vec<_> = (0..dim)
.step_by(2)
.map(|i| 1f32 / 10000f32.powf(i as f32 / dim as f32))
.collect();
let inv_freq_len = inv_freq.len();
let inv_freq = Tensor::from_vec(inv_freq, (1, inv_freq_len), dev)?;
let t = Tensor::arange(0u32, max_seq_len as u32, dev)?
.to_dtype(DType::F32)?
.reshape((max_seq_len, 1))?;
let freqs = t.matmul(&inv_freq)?;
Ok(Self {
sin: freqs.sin()?,
cos: freqs.cos()?,
})
}
fn apply_rotary_emb_qkv(
&self,
qkv: &Tensor,
seqlen_offset: usize,
) -> Result<(Tensor, Tensor, Tensor)> {
let (_b_size, seqlen, three, _, _headdim) = qkv.dims5()?;
if three != 3 {
candle::bail!("unexpected shape for qkv {:?}", qkv.shape())
}
let (_rotary_seqlen, rotary_dim) = self.cos.dims2()?;
let rotary_dim = rotary_dim * 2;
let q_rot = qkv.i((.., .., 0, .., ..rotary_dim))?;
let q_pass = qkv.i((.., .., 0, .., rotary_dim..))?;
let k_rot = qkv.i((.., .., 1, .., ..rotary_dim))?;
let k_pass = qkv.i((.., .., 1, .., rotary_dim..))?;
let q12 = q_rot.chunk(2, D::Minus1)?;
let k12 = k_rot.chunk(2, D::Minus1)?;
let (q1, q2) = (&q12[0], &q12[1]);
let (k1, k2) = (&k12[0], &k12[1]);
let c = self.cos.narrow(0, seqlen_offset, seqlen)?.unsqueeze(1)?;
let s = self.sin.narrow(0, seqlen_offset, seqlen)?.unsqueeze(1)?;
let q_rot = Tensor::cat(
&[
(q1.broadcast_mul(&c)? - q2.broadcast_mul(&s)?)?,
(q1.broadcast_mul(&s)? + q2.broadcast_mul(&c)?)?,
],
D::Minus1,
)?;
let k_rot = Tensor::cat(
&[
(k1.broadcast_mul(&c)? - k2.broadcast_mul(&s)?)?,
(k1.broadcast_mul(&s)? + k2.broadcast_mul(&c)?)?,
],
D::Minus1,
)?;
let q = Tensor::cat(&[&q_rot, &q_pass], D::Minus1)?;
let k = Tensor::cat(&[&k_rot, &k_pass], D::Minus1)?;
let v = qkv.i((.., .., 2))?;
Ok((q, k, v))
}
}
#[derive(Debug, Clone)]
#[allow(clippy::upper_case_acronyms)]
struct MLP {
fc1: Linear,
fc2: Linear,
act: Activation,
}
impl MLP {
fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let n_inner = cfg.n_inner.unwrap_or(4 * cfg.n_embd);
let fc1 = linear(cfg.n_embd, n_inner, vb.pp("fc1"))?;
let fc2 = linear(n_inner, cfg.n_embd, vb.pp("fc2"))?;
Ok(Self {
fc1,
fc2,
act: cfg.activation_function,
})
}
}
impl Module for MLP {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
xs.apply(&self.fc1)?.apply(&self.act)?.apply(&self.fc2)
}
}
#[derive(Debug, Clone)]
struct CausalLMHead {
ln: candle_nn::LayerNorm,
linear: Linear,
}
impl CausalLMHead {
fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let ln = layer_norm(cfg.n_embd, cfg.layer_norm_epsilon, vb.pp("ln"))?;
let linear = linear(cfg.n_embd, cfg.vocab_size, vb.pp("linear"))?;
Ok(Self { ln, linear })
}
}
impl Module for CausalLMHead {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
xs.apply(&self.ln)?
.apply(&self.linear)?
.to_dtype(DType::F32)
}
}
#[derive(Debug, Clone)]
#[allow(clippy::upper_case_acronyms)]
struct MHA {
wqkv: Linear,
out_proj: Linear,
rotary_emb: RotaryEmbedding,
kv_cache: Option<(Tensor, Tensor)>,
head_dim: usize,
n_head: usize,
softmax_scale: f64,
span: tracing::Span,
}
impl MHA {
fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let head_dim = cfg.n_embd / cfg.n_head;
let op_size = cfg.n_embd;
let wqkv = linear(cfg.n_embd, 3 * op_size, vb.pp("Wqkv"))?;
let out_proj = linear(op_size, cfg.n_embd, vb.pp("out_proj"))?;
let rotary_emb = RotaryEmbedding::new(cfg.rotary_dim, MAX_SEQ_LEN, vb.device())?;
let softmax_scale = 1f64 / (head_dim as f64).sqrt();
Ok(Self {
wqkv,
out_proj,
head_dim,
n_head: cfg.n_head,
kv_cache: None,
rotary_emb,
softmax_scale,
span: tracing::span!(tracing::Level::TRACE, "mha"),
})
}
fn forward(&mut self, xs: &Tensor, mask: Option<&Tensor>) -> Result<Tensor> {
let _enter = self.span.enter();
let (b_size, seq_len, _n_embd) = xs.dims3()?;
let qkv = self
.wqkv
.forward(xs)?
.reshape((b_size, seq_len, 3, (), self.head_dim))?;
let seqlen_offset = match &self.kv_cache {
None => 0,
Some((prev_k, _)) => prev_k.dim(1)?,
};
// In the python implementation, a single tensor is returned with the third axis of size 3.
let (q, k, v) = self.rotary_emb.apply_rotary_emb_qkv(&qkv, seqlen_offset)?;
let (k, v) = match &self.kv_cache {
None => (k, v),
Some((prev_k, prev_v)) => {
let k = Tensor::cat(&[prev_k, &k], 1)?;
let v = Tensor::cat(&[prev_v, &v], 1)?;
(k, v)
}
};
self.kv_cache = Some((k.clone(), v.clone()));
// scores = torch.einsum('bthd,bshd->bhts', q, k * softmax_scale)
let q = q.transpose(1, 2)?.flatten_to(1)?; // b*h, t, d
let k = k.transpose(1, 2)?.flatten_to(1)?; // b*h, s, d
let v = v.transpose(1, 2)?.flatten_to(1)?; // b*h, s, d
let attn_weights = (q.matmul(&k.t()?)? * self.softmax_scale)?; // b*h, t, s
// causal_mask = torch.triu(torch.full((seqlen_q, seqlen_k), -10000.0, device=scores.device), 1)
// scores = scores + causal_mask.to(dtype=scores.dtype)
let attn_weights = match mask {
None => attn_weights,
Some(mask) => masked_fill(
&attn_weights,
&mask.broadcast_left(b_size * self.n_head)?,
f32::NEG_INFINITY,
)?,
};
let attn_weights = candle_nn::ops::softmax_last_dim(&attn_weights)?;
// output = torch.einsum('bhts,bshd->bthd', attention_drop, v)
// attn_weights: b*h,t,s, v: b*h,s,d
let attn_output = attn_weights.matmul(&v)?;
// b*h,t,d
let attn_output = attn_output
.reshape((b_size, (), seq_len, self.head_dim))?
.transpose(1, 2)?
.flatten_from(D::Minus2)?;
attn_output.apply(&self.out_proj)
}
fn clear_kv_cache(&mut self) {
self.kv_cache = None
}
}
#[derive(Debug, Clone)]
struct ParallelBlock {
ln: candle_nn::LayerNorm,
mixer: MHA,
mlp: MLP,
span: tracing::Span,
}
impl ParallelBlock {
fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let ln = layer_norm(cfg.n_embd, cfg.layer_norm_epsilon, vb.pp("ln"))?;
let mixer = MHA::new(cfg, vb.pp("mixer"))?;
let mlp = MLP::new(cfg, vb.pp("mlp"))?;
Ok(Self {
ln,
mixer,
mlp,
span: tracing::span!(tracing::Level::TRACE, "block"),
})
}
fn forward(&mut self, xs: &Tensor, mask: Option<&Tensor>) -> Result<Tensor> {
let _enter = self.span.enter();
let residual = xs;
let xs = xs.apply(&self.ln)?;
let attn_outputs = self.mixer.forward(&xs, mask)?;
let feed_forward_hidden_states = self.mlp.forward(&xs)?;
attn_outputs + feed_forward_hidden_states + residual
}
fn clear_kv_cache(&mut self) {
self.mixer.clear_kv_cache()
}
}
#[derive(Debug, Clone)]
pub struct MixFormerSequentialForCausalLM {
embedding: Embedding,
blocks: Vec<ParallelBlock>,
head: CausalLMHead,
span: tracing::Span,
}
impl MixFormerSequentialForCausalLM {
pub fn new_v2(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let vb_head = vb.pp("lm_head");
let vb = vb.pp("transformer");
let embedding = Embedding::new(cfg, vb.pp("embd"))?;
let mut blocks = Vec::new();
for i in 0..cfg.n_layer {
let block = ParallelBlock::new(cfg, vb.pp("h").pp(i))?;
blocks.push(block)
}
let head = CausalLMHead::new(cfg, vb_head)?;
Ok(Self {
embedding,
blocks,
head,
span: tracing::span!(tracing::Level::TRACE, "mixformer"),
})
}
pub fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let vb = vb.pp("layers");
let embedding = Embedding::new(cfg, vb.pp(0))?;
let mut blocks = Vec::new();
for i in 0..cfg.n_layer {
let block = ParallelBlock::new(cfg, vb.pp(i + 1))?;
blocks.push(block);
}
let head = CausalLMHead::new(cfg, vb.pp(cfg.n_layer + 1))?;
Ok(Self {
embedding,
blocks,
head,
span: tracing::span!(tracing::Level::TRACE, "mixformer"),
})
}
pub fn forward(&mut self, xs: &Tensor) -> Result<Tensor> {
let _enter = self.span.enter();
let (_b_size, seq_len) = xs.dims2()?;
let mut xs = xs.apply(&self.embedding)?;
let mask = if seq_len <= 1 {
None
} else {
Some(get_mask(seq_len, xs.device())?)
};
for block in self.blocks.iter_mut() {
xs = block.forward(&xs, mask.as_ref())?;
}
xs.narrow(1, seq_len - 1, 1)?.apply(&self.head)?.squeeze(1)
}
pub fn forward_with_img(
&mut self,
bos_token: &Tensor,
xs: &Tensor,
img_embeds: &Tensor,
) -> Result<Tensor> {
let _enter = self.span.enter();
let xs = xs.apply(&self.embedding)?;
let bos_token = bos_token.apply(&self.embedding)?;
// Python implementation sequence order is <bos token embedding><img embedding><rest of text embedding>
// https://github.com/vikhyat/moondream/blob/a9d788a20d1543fb1479edc54106e88cff7759d3/moondream/moondream.py#L43-L56
let mut xs = Tensor::cat(&[bos_token, img_embeds.clone(), xs], 1)?;
let (_b_size, seq_len, _embds) = xs.dims3()?;
let mask = Some(get_mask(seq_len, xs.device())?);
for block in self.blocks.iter_mut() {
xs = block.forward(&xs, mask.as_ref())?
}
let xs = xs
.narrow(1, seq_len - 1, 1)?
.apply(&self.head)?
.squeeze(1)?;
Ok(xs)
}
pub fn clear_kv_cache(&mut self) {
self.blocks.iter_mut().for_each(|b| b.clear_kv_cache())
}
}
| candle/candle-transformers/src/models/quantized_mixformer.rs/0 | {
"file_path": "candle/candle-transformers/src/models/quantized_mixformer.rs",
"repo_id": "candle",
"token_count": 6503
} |
//! RWKV v5 model implementation.
//!
//! The [RWKV model](https://wiki.rwkv.com/) is a recurrent neural network model
//! with performance on par with transformer architectures. Several variants are
//! available, candle implements the v5 and v6 versions and can be used with
//! Eagle 7B([blog post](https://blog.rwkv.com/p/eagle-7b-soaring-past-transformers)).
//!
//! Key characteristics:
//! - Time-mix attention mechanism
//! - Channel-mix feed-forward network
//! - Linear attention
//! - Group normalization
//! - Token shift mechanism
//!
//! References:
//! - [RWKV Language Model](https://github.com/BlinkDL/RWKV-LM)
//! - [RWKV v5 Release](https://github.com/BlinkDL/ChatRWKV/tree/main)
//!
//! # Example
//!
//! ```bash
//! cargo run --example rwkv --release -- \
//! --prompt "The smallest prime is "
//!
//! > avx: true, neon: false, simd128: false, f16c: true
//! > temp: 0.00 repeat-penalty: 1.10 repeat-last-n: 64
//! > The smallest prime is ϕ(2) = 2.
//! > The smallest composite is ϕ(3) = 3.
//! > The smallest perfect number is ϕ(5) = 5.
//! > The smallest perfect square is ϕ(4) = 4.
//! > The smallest perfect cube is ϕ(6) = 6.
//! ```
use super::with_tracing::{layer_norm, linear_no_bias as linear, LayerNorm, Linear};
use candle::{DType, Device, IndexOp, Result, Tensor};
use candle_nn::{embedding, Embedding, Module, VarBuilder};
use std::collections::{HashMap, HashSet};
fn default_num_attention_heads() -> usize {
64
}
// https://huggingface.co/RWKV/HF_v5-Eagle-7B/blob/main/configuration_rwkv5.py
#[derive(Debug, Clone, serde::Deserialize)]
pub struct Config {
pub vocab_size: usize,
pub hidden_size: usize,
pub num_hidden_layers: usize,
pub attention_hidden_size: usize,
#[serde(default = "default_num_attention_heads")]
pub num_attention_heads: usize,
pub head_size: usize,
pub intermediate_size: Option<usize>,
pub layer_norm_epsilon: f64,
pub rescale_every: usize,
}
pub struct StatePerLayer {
pub extract_key_value: Tensor,
pub linear_attention: Tensor,
pub feed_forward: Tensor,
}
pub struct State {
pub per_layer: Vec<StatePerLayer>,
pub pos: usize,
}
impl State {
pub fn new(batch_size: usize, cfg: &Config, dev: &Device) -> Result<Self> {
let mut per_layer = Vec::with_capacity(cfg.num_hidden_layers);
// Certainly a weird convention but taken from modeling_rwkv5.py
let num_attention_heads = cfg.hidden_size / cfg.num_attention_heads;
for _layer_idx in 0..cfg.num_hidden_layers {
let extract_key_value = Tensor::zeros((batch_size, cfg.hidden_size), DType::F32, dev)?;
let linear_attention = Tensor::zeros(
(
batch_size,
num_attention_heads,
cfg.hidden_size / num_attention_heads,
cfg.hidden_size / num_attention_heads,
),
DType::F32,
dev,
)?;
let feed_forward = Tensor::zeros((batch_size, cfg.hidden_size), DType::F32, dev)?;
per_layer.push(StatePerLayer {
extract_key_value,
linear_attention,
feed_forward,
});
}
Ok(Self { per_layer, pos: 0 })
}
}
#[derive(Debug, Clone)]
struct SelfAttention {
key: Linear,
receptance: Linear,
value: Linear,
gate: Linear,
output: Linear,
ln_x: candle_nn::GroupNorm,
time_mix_key: Tensor,
time_mix_value: Tensor,
time_mix_receptance: Tensor,
time_decay: Tensor,
time_faaaa: Tensor,
time_mix_gate: Tensor,
layer_id: usize,
n_attn_heads: usize,
}
impl SelfAttention {
pub fn new(layer_id: usize, cfg: &Config, vb: VarBuilder) -> Result<Self> {
let hidden_size = cfg.hidden_size;
let attn_hidden_size = cfg.attention_hidden_size;
let key = linear(hidden_size, attn_hidden_size, vb.pp("key"))?;
let receptance = linear(hidden_size, attn_hidden_size, vb.pp("receptance"))?;
let value = linear(hidden_size, attn_hidden_size, vb.pp("value"))?;
let gate = linear(hidden_size, attn_hidden_size, vb.pp("gate"))?;
let output = linear(attn_hidden_size, hidden_size, vb.pp("output"))?;
let ln_x = candle_nn::group_norm(
hidden_size / cfg.head_size,
hidden_size,
1e-5,
vb.pp("ln_x"),
)?;
let time_mix_key = vb.get((1, 1, cfg.hidden_size), "time_mix_key")?;
let time_mix_value = vb.get((1, 1, cfg.hidden_size), "time_mix_value")?;
let time_mix_receptance = vb.get((1, 1, cfg.hidden_size), "time_mix_receptance")?;
let n_attn_heads = cfg.hidden_size / cfg.head_size;
let time_decay = vb.get((n_attn_heads, cfg.head_size), "time_decay")?;
let time_faaaa = vb.get((n_attn_heads, cfg.head_size), "time_faaaa")?;
let time_mix_gate = vb.get((1, 1, cfg.hidden_size), "time_mix_gate")?;
Ok(Self {
key,
value,
receptance,
gate,
output,
ln_x,
time_mix_key,
time_mix_value,
time_mix_receptance,
time_decay,
time_faaaa,
time_mix_gate,
layer_id,
n_attn_heads,
})
}
pub fn forward(&self, xs: &Tensor, state: &mut State) -> Result<Tensor> {
let h = self.time_decay.dim(0)?;
let (b, t, s) = xs.dims3()?;
let s = s / h;
let (receptance, key, value, gate) = {
// extract key-value
let shifted = state.per_layer[self.layer_id].extract_key_value.clone();
let shifted = if shifted.rank() == 2 {
shifted.unsqueeze(1)?
} else {
shifted
};
let key = ((xs * &self.time_mix_key)? + &shifted * (1.0 - &self.time_mix_key)?)?;
let value = ((xs * &self.time_mix_value)? + &shifted * (1.0 - &self.time_mix_value)?)?;
let receptance = ((xs * &self.time_mix_receptance)?
+ &shifted * (1.0 - &self.time_mix_receptance)?)?;
let gate = ((xs * &self.time_mix_gate)? + &shifted * (1.0 - &self.time_mix_gate)?)?;
let key = self.key.forward(&key)?;
let value = self.value.forward(&value)?;
let receptance = self.receptance.forward(&receptance)?;
let gate = candle_nn::ops::silu(&self.gate.forward(&gate)?)?;
state.per_layer[self.layer_id].extract_key_value = xs.i((.., t - 1))?;
(receptance, key, value, gate)
};
// linear attention
let mut state_ = state.per_layer[self.layer_id].linear_attention.clone();
let key = key.reshape((b, t, h, s))?.permute((0, 2, 3, 1))?;
let value = value.reshape((b, t, h, s))?.transpose(1, 2)?;
let receptance = receptance.reshape((b, t, h, s))?.transpose(1, 2)?;
let time_decay = self
.time_decay
.exp()?
.neg()?
.exp()?
.reshape(((), 1, 1))?
.reshape((self.n_attn_heads, (), 1))?;
let time_faaaa =
self.time_faaaa
.reshape(((), 1, 1))?
.reshape((self.n_attn_heads, (), 1))?;
let mut out: Vec<Tensor> = Vec::with_capacity(t);
for t_ in 0..t {
let rt = receptance.i((.., .., t_..t_ + 1))?.contiguous()?;
let kt = key.i((.., .., .., t_..t_ + 1))?.contiguous()?;
let vt = value.i((.., .., t_..t_ + 1))?.contiguous()?;
let at = kt.matmul(&vt)?;
let rhs = (time_faaaa.broadcast_mul(&at)? + &state_)?;
let out_ = rt.matmul(&rhs)?.squeeze(2)?;
state_ = (&at + time_decay.broadcast_mul(&state_))?;
out.push(out_)
}
let out = Tensor::cat(&out, 1)?.reshape((b * t, h * s, 1))?;
let out = out.apply(&self.ln_x)?.reshape((b, t, h * s))?;
let out = (out * gate)?.apply(&self.output)?;
state.per_layer[self.layer_id].linear_attention = state_;
Ok(out)
}
}
#[derive(Debug, Clone)]
struct FeedForward {
time_mix_key: Tensor,
time_mix_receptance: Tensor,
key: Linear,
receptance: Linear,
value: Linear,
layer_id: usize,
}
impl FeedForward {
pub fn new(layer_id: usize, cfg: &Config, vb: VarBuilder) -> Result<Self> {
let int_size = cfg
.intermediate_size
.unwrap_or(((cfg.hidden_size as f64 * 3.5) as usize) / 32 * 32);
let key = linear(cfg.hidden_size, int_size, vb.pp("key"))?;
let receptance = linear(cfg.hidden_size, cfg.hidden_size, vb.pp("receptance"))?;
let value = linear(int_size, cfg.hidden_size, vb.pp("value"))?;
let time_mix_key = vb.get((1, 1, cfg.hidden_size), "time_mix_key")?;
let time_mix_receptance = vb.get((1, 1, cfg.hidden_size), "time_mix_receptance")?;
Ok(Self {
key,
receptance,
value,
time_mix_key,
time_mix_receptance,
layer_id,
})
}
pub fn forward(&self, xs: &Tensor, state: &mut State) -> Result<Tensor> {
let shifted = &state.per_layer[self.layer_id].feed_forward;
let key = (xs.broadcast_mul(&self.time_mix_key)?
+ shifted.broadcast_mul(&(1.0 - &self.time_mix_key)?)?)?;
let receptance = (xs.broadcast_mul(&self.time_mix_receptance)?
+ shifted.broadcast_mul(&(1.0 - &self.time_mix_receptance)?)?)?;
let key = key.apply(&self.key)?.relu()?.sqr()?;
let value = key.apply(&self.value)?;
let receptance = candle_nn::ops::sigmoid(&receptance.apply(&self.receptance)?)?;
state.per_layer[self.layer_id].feed_forward = xs.i((.., xs.dim(1)? - 1))?;
let xs = (receptance * value)?;
Ok(xs)
}
}
#[derive(Debug, Clone)]
struct Block {
pre_ln: Option<LayerNorm>,
ln1: LayerNorm,
ln2: LayerNorm,
attention: SelfAttention,
feed_forward: FeedForward,
}
impl Block {
pub fn new(layer_id: usize, cfg: &Config, vb: VarBuilder) -> Result<Self> {
let ln1 = layer_norm(cfg.hidden_size, cfg.layer_norm_epsilon, vb.pp("ln1"))?;
let ln2 = layer_norm(cfg.hidden_size, cfg.layer_norm_epsilon, vb.pp("ln2"))?;
let pre_ln = if layer_id == 0 {
let ln = layer_norm(cfg.hidden_size, cfg.layer_norm_epsilon, vb.pp("pre_ln"))?;
Some(ln)
} else {
None
};
let attention = SelfAttention::new(layer_id, cfg, vb.pp("attention"))?;
let feed_forward = FeedForward::new(layer_id, cfg, vb.pp("feed_forward"))?;
Ok(Self {
pre_ln,
ln1,
ln2,
attention,
feed_forward,
})
}
pub fn forward(&self, xs: &Tensor, state: &mut State) -> Result<Tensor> {
let xs = match self.pre_ln.as_ref() {
None => xs.clone(),
Some(pre_ln) => xs.apply(pre_ln)?,
};
let attention = self.attention.forward(&xs.apply(&self.ln1)?, state)?;
let xs = (xs + attention)?;
let feed_forward = self.feed_forward.forward(&xs.apply(&self.ln2)?, state)?;
let xs = (xs + feed_forward)?;
Ok(xs)
}
}
#[derive(Debug, Clone)]
pub struct Model {
embeddings: Embedding,
blocks: Vec<Block>,
ln_out: LayerNorm,
head: Linear,
rescale_every: usize,
layers_are_rescaled: bool,
}
impl Model {
pub fn new(cfg: &Config, vb: VarBuilder) -> Result<Self> {
let vb_m = vb.pp("rwkv");
let embeddings = embedding(cfg.vocab_size, cfg.hidden_size, vb_m.pp("embeddings"))?;
let mut blocks = Vec::with_capacity(cfg.num_hidden_layers);
let vb_b = vb_m.pp("blocks");
for block_index in 0..cfg.num_hidden_layers {
let block = Block::new(block_index, cfg, vb_b.pp(block_index))?;
blocks.push(block)
}
let ln_out = layer_norm(cfg.hidden_size, 1e-5, vb_m.pp("ln_out"))?;
let head = linear(cfg.hidden_size, cfg.vocab_size, vb.pp("head"))?;
Ok(Self {
embeddings,
blocks,
ln_out,
head,
rescale_every: cfg.rescale_every,
layers_are_rescaled: false, // This seem to only happen for the f16/bf16 dtypes.
})
}
pub fn forward(&self, xs: &Tensor, state: &mut State) -> Result<Tensor> {
let (_b_size, _seq_len) = xs.dims2()?;
let mut xs = xs.apply(&self.embeddings)?;
for (block_idx, block) in self.blocks.iter().enumerate() {
xs = block.forward(&xs, state)?;
if self.layers_are_rescaled && (block_idx + 1) % self.rescale_every == 0 {
xs = (xs / 2.)?
}
}
let xs = xs.apply(&self.ln_out)?.apply(&self.head)?;
state.pos += 1;
Ok(xs)
}
}
type Bytes = Vec<u8>;
// https://github.com/BlinkDL/ChatRWKV/blob/095e812aef15a1f74107f6c39d13578a2412dc46/RWKV_v5_demo.py#L14
pub struct Tokenizer {
table: Vec<Vec<Vec<Bytes>>>,
good: Vec<HashSet<u8>>,
idx2token: HashMap<u32, Vec<u8>>,
token2idx: HashMap<Vec<u8>, u32>,
}
impl Tokenizer {
pub fn new<P: AsRef<std::path::Path>>(p: P) -> Result<Self> {
let file = std::fs::File::open(p)?;
let token2idx: HashMap<String, u32> =
serde_json::from_reader(file).map_err(candle::Error::wrap)?;
let token2idx = token2idx
.into_iter()
.map(|(key, value)| (key.into_bytes(), value))
.collect::<HashMap<_, _>>();
let idx2token = token2idx
.iter()
.map(|(key, value)| (*value, key.to_vec()))
.collect::<HashMap<_, _>>();
let max_idx = token2idx.values().copied().max().unwrap_or(0);
let mut table = vec![vec![vec![]; 256]; 256];
let mut good = vec![HashSet::new(); 256];
for idx in (0..(1 + max_idx)).rev() {
let s = match idx2token.get(&idx) {
None => continue,
Some(s) => s,
};
if s.len() >= 2 {
let (s0, s1) = (s[0], s[1]);
table[s0 as usize][s1 as usize].push(s.to_vec());
good[s0 as usize].insert(s1);
}
}
Ok(Self {
table,
good,
idx2token,
token2idx,
})
}
pub fn decode_bytes(&self, tokens: &[u32]) -> Vec<u8> {
let mut v = Vec::new();
for token_id in tokens.iter() {
if let Some(token) = self.idx2token.get(token_id) {
v.extend_from_slice(token.as_slice())
}
}
v
}
pub fn decode(&self, tokens: &[u32]) -> Result<String> {
let bytes = self.decode_bytes(tokens);
String::from_utf8(bytes).map_err(candle::Error::wrap)
}
pub fn encode_bytes(&self, bytes: &[u8]) -> Result<Vec<u32>> {
let mut tokens = Vec::new();
let mut i = 0;
while i < bytes.len() {
let mut s = vec![bytes[i]];
if i + 1 < bytes.len() && self.good[bytes[i] as usize].contains(&bytes[i + 1]) {
let table = &self.table[bytes[i] as usize][bytes[i + 1] as usize];
for table_elem in table.iter() {
if bytes[i..].starts_with(table_elem) {
s = table_elem.to_vec();
break;
}
}
}
i += s.len();
let token = match self.token2idx.get(&s) {
None => candle::bail!("unexpected token '{}' {s:?}", String::from_utf8_lossy(&s)),
Some(token) => *token,
};
tokens.push(token)
}
Ok(tokens)
}
pub fn encode(&self, str: &str) -> Result<Vec<u32>> {
self.encode_bytes(str.as_bytes())
}
}
| candle/candle-transformers/src/models/rwkv_v5.rs/0 | {
"file_path": "candle/candle-transformers/src/models/rwkv_v5.rs",
"repo_id": "candle",
"token_count": 8119
} |
//! Ancestral sampling with Euler method steps.
//!
//! Based on the original [`k-diffusion` implementation by Katherine Crowson]( https://github.com/crowsonkb/k-diffusion/blob/481677d114f6ea445aa009cf5bd7a9cdee909e47/k_diffusion/sampling.py#L72).
//!
use super::{
schedulers::{
betas_for_alpha_bar, BetaSchedule, PredictionType, Scheduler, SchedulerConfig,
TimestepSpacing,
},
utils::interp,
};
use candle::{bail, Error, Result, Tensor};
/// The configuration for the EulerAncestral Discrete scheduler.
#[derive(Debug, Clone, Copy)]
pub struct EulerAncestralDiscreteSchedulerConfig {
/// The value of beta at the beginning of training.n
pub beta_start: f64,
/// The value of beta at the end of training.
pub beta_end: f64,
/// How beta evolved during training.
pub beta_schedule: BetaSchedule,
/// Adjust the indexes of the inference schedule by this value.
pub steps_offset: usize,
/// prediction type of the scheduler function, one of `epsilon` (predicting
/// the noise of the diffusion process), `sample` (directly predicting the noisy sample`)
/// or `v_prediction` (see [section 2.4](https://imagen.research.google/video/paper.pdf))
pub prediction_type: PredictionType,
/// number of diffusion steps used to train the model
pub train_timesteps: usize,
/// time step spacing for the diffusion process
pub timestep_spacing: TimestepSpacing,
}
impl Default for EulerAncestralDiscreteSchedulerConfig {
fn default() -> Self {
Self {
beta_start: 0.00085f64,
beta_end: 0.012f64,
beta_schedule: BetaSchedule::ScaledLinear,
steps_offset: 1,
prediction_type: PredictionType::Epsilon,
train_timesteps: 1000,
timestep_spacing: TimestepSpacing::Leading,
}
}
}
impl SchedulerConfig for EulerAncestralDiscreteSchedulerConfig {
fn build(&self, inference_steps: usize) -> Result<Box<dyn Scheduler>> {
Ok(Box::new(EulerAncestralDiscreteScheduler::new(
inference_steps,
*self,
)?))
}
}
/// The EulerAncestral Discrete scheduler.
#[derive(Debug, Clone)]
pub struct EulerAncestralDiscreteScheduler {
timesteps: Vec<usize>,
sigmas: Vec<f64>,
init_noise_sigma: f64,
pub config: EulerAncestralDiscreteSchedulerConfig,
}
// clip_sample: False, set_alpha_to_one: False
impl EulerAncestralDiscreteScheduler {
/// Creates a new EulerAncestral Discrete scheduler given the number of steps to be
/// used for inference as well as the number of steps that was used
/// during training.
pub fn new(
inference_steps: usize,
config: EulerAncestralDiscreteSchedulerConfig,
) -> Result<Self> {
let step_ratio = config.train_timesteps / inference_steps;
let timesteps: Vec<usize> = match config.timestep_spacing {
TimestepSpacing::Leading => (0..(inference_steps))
.map(|s| s * step_ratio + config.steps_offset)
.rev()
.collect(),
TimestepSpacing::Trailing => std::iter::successors(Some(config.train_timesteps), |n| {
if *n > step_ratio {
Some(n - step_ratio)
} else {
None
}
})
.map(|n| n - 1)
.collect(),
TimestepSpacing::Linspace => {
super::utils::linspace(0.0, (config.train_timesteps - 1) as f64, inference_steps)?
.to_vec1::<f64>()?
.iter()
.map(|&f| f as usize)
.rev()
.collect()
}
};
let betas = match config.beta_schedule {
BetaSchedule::ScaledLinear => super::utils::linspace(
config.beta_start.sqrt(),
config.beta_end.sqrt(),
config.train_timesteps,
)?
.sqr()?,
BetaSchedule::Linear => {
super::utils::linspace(config.beta_start, config.beta_end, config.train_timesteps)?
}
BetaSchedule::SquaredcosCapV2 => betas_for_alpha_bar(config.train_timesteps, 0.999)?,
};
let betas = betas.to_vec1::<f64>()?;
let mut alphas_cumprod = Vec::with_capacity(betas.len());
for &beta in betas.iter() {
let alpha = 1.0 - beta;
alphas_cumprod.push(alpha * *alphas_cumprod.last().unwrap_or(&1f64))
}
let sigmas: Vec<f64> = alphas_cumprod
.iter()
.map(|&f| ((1. - f) / f).sqrt())
.collect();
let sigmas_xa: Vec<_> = (0..sigmas.len()).map(|i| i as f64).collect();
let mut sigmas_int = interp(
×teps.iter().map(|&t| t as f64).collect::<Vec<_>>(),
&sigmas_xa,
&sigmas,
);
sigmas_int.push(0.0);
// standard deviation of the initial noise distribution
// f64 does not implement Ord such that there is no `max`, so we need to use this workaround
let init_noise_sigma = *sigmas_int
.iter()
.chain(std::iter::once(&0.0))
.reduce(|a, b| if a > b { a } else { b })
.expect("init_noise_sigma could not be reduced from sigmas - this should never happen");
Ok(Self {
sigmas: sigmas_int,
timesteps,
init_noise_sigma,
config,
})
}
}
impl Scheduler for EulerAncestralDiscreteScheduler {
fn timesteps(&self) -> &[usize] {
self.timesteps.as_slice()
}
/// Ensures interchangeability with schedulers that need to scale the denoising model input
/// depending on the current timestep.
///
/// Scales the denoising model input by `(sigma**2 + 1) ** 0.5` to match the K-LMS algorithm
fn scale_model_input(&self, sample: Tensor, timestep: usize) -> Result<Tensor> {
let step_index = match self.timesteps.iter().position(|&t| t == timestep) {
Some(i) => i,
None => bail!("timestep out of this schedulers bounds: {timestep}"),
};
let sigma = self
.sigmas
.get(step_index)
.expect("step_index out of sigma bounds - this shouldn't happen");
sample / ((sigma.powi(2) + 1.).sqrt())
}
/// Performs a backward step during inference.
fn step(&mut self, model_output: &Tensor, timestep: usize, sample: &Tensor) -> Result<Tensor> {
let step_index = self
.timesteps
.iter()
.position(|&p| p == timestep)
.ok_or_else(|| Error::Msg("timestep out of this schedulers bounds".to_string()))?;
let sigma_from = &self.sigmas[step_index];
let sigma_to = &self.sigmas[step_index + 1];
// 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
let pred_original_sample = match self.config.prediction_type {
PredictionType::Epsilon => (sample - (model_output * *sigma_from))?,
PredictionType::VPrediction => {
((model_output * (-sigma_from / (sigma_from.powi(2) + 1.0).sqrt()))?
+ (sample / (sigma_from.powi(2) + 1.0))?)?
}
PredictionType::Sample => bail!("prediction_type not implemented yet: sample"),
};
let sigma_up = (sigma_to.powi(2) * (sigma_from.powi(2) - sigma_to.powi(2))
/ sigma_from.powi(2))
.sqrt();
let sigma_down = (sigma_to.powi(2) - sigma_up.powi(2)).sqrt();
// 2. convert to a ODE derivative
let derivative = ((sample - pred_original_sample)? / *sigma_from)?;
let dt = sigma_down - *sigma_from;
let prev_sample = (sample + derivative * dt)?;
let noise = prev_sample.randn_like(0.0, 1.0)?;
prev_sample + noise * sigma_up
}
fn add_noise(&self, original: &Tensor, noise: Tensor, timestep: usize) -> Result<Tensor> {
let step_index = self
.timesteps
.iter()
.position(|&p| p == timestep)
.ok_or_else(|| Error::Msg("timestep out of this schedulers bounds".to_string()))?;
let sigma = self
.sigmas
.get(step_index)
.expect("step_index out of sigma bounds - this shouldn't happen");
original + (noise * *sigma)?
}
fn init_noise_sigma(&self) -> f64 {
match self.config.timestep_spacing {
TimestepSpacing::Trailing | TimestepSpacing::Linspace => self.init_noise_sigma,
TimestepSpacing::Leading => (self.init_noise_sigma.powi(2) + 1.0).sqrt(),
}
}
}
| candle/candle-transformers/src/models/stable_diffusion/euler_ancestral_discrete.rs/0 | {
"file_path": "candle/candle-transformers/src/models/stable_diffusion/euler_ancestral_discrete.rs",
"repo_id": "candle",
"token_count": 4097
} |
import init, { Model } from "./build/m.js";
async function fetchArrayBuffer(url, cacheFile = true) {
if (!cacheFile) return new Uint8Array(await (await fetch(url)).arrayBuffer());
const cacheName = "blip-candle-cache";
const cache = await caches.open(cacheName);
const cachedResponse = await cache.match(url);
if (cachedResponse) {
const data = await cachedResponse.arrayBuffer();
return new Uint8Array(data);
}
const res = await fetch(url, { cache: "force-cache" });
cache.put(url, res.clone());
return new Uint8Array(await res.arrayBuffer());
}
class Blip {
static instance = {};
static async getInstance(
weightsURL,
tokenizerURL,
configURL,
modelID,
quantized
) {
if (!this.instance[modelID]) {
await init();
self.postMessage({ status: "loading", message: "Loading Model" });
const [weightsArrayU8, tokenizerArrayU8, configArrayU8] =
await Promise.all([
fetchArrayBuffer(weightsURL),
fetchArrayBuffer(tokenizerURL),
fetchArrayBuffer(configURL),
]);
this.instance[modelID] = new Model(
weightsArrayU8,
tokenizerArrayU8,
configArrayU8,
quantized
);
} else {
self.postMessage({ status: "ready", message: "Model Already Loaded" });
}
return this.instance[modelID];
}
}
self.addEventListener("message", async (event) => {
const { weightsURL, tokenizerURL, configURL, modelID, imageURL, quantized } =
event.data;
try {
self.postMessage({ status: "status", message: "Loading Blip Model..." });
const model = await Blip.getInstance(
weightsURL,
tokenizerURL,
configURL,
modelID,
quantized
);
self.postMessage({
status: "status",
message: "Running Blip Inference...",
});
const imageArrayU8 = await fetchArrayBuffer(imageURL, false);
const output = model.generate_caption_from_image(imageArrayU8);
self.postMessage({
status: "complete",
message: "complete",
output: output,
});
} catch (e) {
self.postMessage({ error: e });
}
});
| candle/candle-wasm-examples/blip/blipWorker.js/0 | {
"file_path": "candle/candle-wasm-examples/blip/blipWorker.js",
"repo_id": "candle",
"token_count": 815
} |
use candle::{DType, Device, Tensor};
use candle_nn::VarBuilder;
use candle_transformers::generation::LogitsProcessor;
use candle_transformers::models::mixformer::{Config, MixFormerSequentialForCausalLM as MixFormer};
use candle_transformers::models::quantized_mixformer::MixFormerSequentialForCausalLM as QMixFormer;
use candle_wasm_example_phi::console_log;
use js_sys::Date;
use serde::Deserialize;
use tokenizers::Tokenizer;
use wasm_bindgen::prelude::*;
enum SelectedModel {
MixFormer(MixFormer),
Quantized(QMixFormer),
}
#[wasm_bindgen]
pub struct Model {
model: SelectedModel,
tokenizer: Tokenizer,
logits_processor: LogitsProcessor,
tokens: Vec<u32>,
repeat_penalty: f32,
repeat_last_n: usize,
}
#[derive(Debug, Clone, PartialEq, Deserialize)]
pub struct ModelName {
pub _name_or_path: String,
}
#[wasm_bindgen]
impl Model {
#[wasm_bindgen(constructor)]
pub fn load(
weights: Vec<u8>,
tokenizer: Vec<u8>,
config: Vec<u8>,
quantized: bool,
) -> Result<Model, JsError> {
console_error_panic_hook::set_once();
console_log!("loading model");
let device = Device::Cpu;
let name: ModelName = serde_json::from_slice(&config)?;
let config: Config = serde_json::from_slice(&config)?;
console_log!("config loaded {:?}", name);
let tokenizer =
Tokenizer::from_bytes(&tokenizer).map_err(|m| JsError::new(&m.to_string()))?;
let start = Date::now();
console_log!("weights len: {:?}", weights.len());
let model = if quantized {
let vb = candle_transformers::quantized_var_builder::VarBuilder::from_gguf_buffer(
&weights, &device,
)?;
console_log!("weights loaded");
if name._name_or_path == "microsoft/phi-2" {
let model = QMixFormer::new_v2(&config, vb)?;
SelectedModel::Quantized(model)
} else {
let model = QMixFormer::new(&config, vb)?;
SelectedModel::Quantized(model)
}
} else {
let device = &Device::Cpu;
let vb = VarBuilder::from_buffered_safetensors(weights, DType::F32, device)?;
let model = MixFormer::new(&config, vb)?;
SelectedModel::MixFormer(model)
};
console_log!("model loaded in {:?}s", (Date::now() - start) / 1000.);
let logits_processor = LogitsProcessor::new(299792458, None, None);
Ok(Self {
model,
tokenizer,
tokens: vec![],
logits_processor,
repeat_penalty: 1.,
repeat_last_n: 64,
})
}
#[wasm_bindgen]
pub fn init_with_prompt(
&mut self,
prompt: String,
temp: f64,
top_p: f64,
repeat_penalty: f32,
repeat_last_n: usize,
seed: u64,
) -> Result<String, JsError> {
match &mut self.model {
SelectedModel::MixFormer(m) => m.clear_kv_cache(),
SelectedModel::Quantized(m) => m.clear_kv_cache(),
};
let temp = if temp <= 0. { None } else { Some(temp) };
let top_p = if top_p <= 0. || top_p >= 1. {
None
} else {
Some(top_p)
};
self.logits_processor = LogitsProcessor::new(seed, temp, top_p);
self.repeat_penalty = repeat_penalty;
self.repeat_last_n = repeat_last_n;
self.tokens.clear();
let tokens = self
.tokenizer
.encode(prompt, true)
.map_err(|m| JsError::new(&m.to_string()))?
.get_ids()
.to_vec();
let text = self
.process(&tokens)
.map_err(|m| JsError::new(&m.to_string()))?;
Ok(text)
}
#[wasm_bindgen]
pub fn next_token(&mut self) -> Result<String, JsError> {
let last_token = *self.tokens.last().unwrap();
let text = self
.process(&[last_token])
.map_err(|m| JsError::new(&m.to_string()))?;
Ok(text)
}
}
impl Model {
fn process(&mut self, tokens: &[u32]) -> candle::Result<String> {
let dev = Device::Cpu;
let input = Tensor::new(tokens, &dev)?.unsqueeze(0)?;
let logits = match &mut self.model {
SelectedModel::MixFormer(m) => m.forward(&input)?,
SelectedModel::Quantized(m) => m.forward(&input)?,
};
let logits = logits.squeeze(0)?.to_dtype(DType::F32)?;
let logits = if self.repeat_penalty == 1. {
logits
} else {
let start_at = tokens.len().saturating_sub(self.repeat_last_n);
candle_transformers::utils::apply_repeat_penalty(
&logits,
self.repeat_penalty,
&tokens[start_at..],
)?
};
let next_token = self.logits_processor.sample(&logits)?;
self.tokens.push(next_token);
let token = match self.tokenizer.decode(&[next_token], false) {
Ok(token) => token,
Err(e) => {
console_log!("error decoding token: {:?}", e);
"".to_string()
}
};
// console_log!("token: {:?}: {:?}", token, next_token);
Ok(token)
}
}
fn main() {
console_error_panic_hook::set_once();
}
| candle/candle-wasm-examples/phi/src/bin/m.rs/0 | {
"file_path": "candle/candle-wasm-examples/phi/src/bin/m.rs",
"repo_id": "candle",
"token_count": 2646
} |
use candle::{DType, Device, Tensor};
use candle_nn::VarBuilder;
use candle_transformers::generation::LogitsProcessor;
pub use candle_transformers::models::t5::{Config, T5EncoderModel, T5ForConditionalGeneration};
use candle_wasm_example_t5::console_log;
use tokenizers::Tokenizer;
use wasm_bindgen::prelude::*;
#[wasm_bindgen]
pub struct ModelEncoder {
model: T5EncoderModel,
tokenizer: Tokenizer,
}
#[wasm_bindgen]
pub struct ModelConditionalGeneration {
model: T5ForConditionalGeneration,
tokenizer: Tokenizer,
config: Config,
}
#[wasm_bindgen]
impl ModelConditionalGeneration {
#[wasm_bindgen(constructor)]
pub fn load(
weights: Vec<u8>,
tokenizer: Vec<u8>,
config: Vec<u8>,
) -> Result<ModelConditionalGeneration, JsError> {
console_error_panic_hook::set_once();
console_log!("loading model");
let device = &Device::Cpu;
let vb = VarBuilder::from_buffered_safetensors(weights, DType::F32, device)?;
let mut config: Config = serde_json::from_slice(&config)?;
let tokenizer =
Tokenizer::from_bytes(&tokenizer).map_err(|m| JsError::new(&m.to_string()))?;
let model = T5ForConditionalGeneration::load(vb, &config)?;
config.use_cache = false;
Ok(Self {
model,
tokenizer,
config,
})
}
pub fn decode(&mut self, input: JsValue) -> Result<JsValue, JsError> {
let input: ConditionalGenerationParams =
serde_wasm_bindgen::from_value(input).map_err(|m| JsError::new(&m.to_string()))?;
let device = &Device::Cpu;
self.model.clear_kv_cache();
let mut output_token_ids = [self.config.pad_token_id as u32].to_vec();
let prompt = input.prompt;
let repeat_penalty = input.repeat_penalty;
let repeat_last_n = input.repeat_last_n;
let seed = input.seed;
let max_length = usize::clamp(input.max_length.unwrap_or(512), 0, 512);
let temperature = if input.temperature <= 0. {
None
} else {
Some(input.temperature)
};
let top_p = if input.top_p <= 0. || input.top_p >= 1. {
None
} else {
Some(input.top_p)
};
let mut logits_processor = LogitsProcessor::new(seed, temperature, top_p);
let tokens = self
.tokenizer
.encode(prompt, true)
.map_err(|m| JsError::new(&m.to_string()))?
.get_ids()
.to_vec();
let input_token_ids = Tensor::new(&tokens[..], device)?.unsqueeze(0)?;
let encoder_output = self.model.encode(&input_token_ids)?;
let mut decoded = String::new();
for index in 0.. {
if output_token_ids.len() > max_length {
break;
}
let decoder_token_ids = if index == 0 {
Tensor::new(output_token_ids.as_slice(), device)?.unsqueeze(0)?
} else {
let last_token = *output_token_ids.last().unwrap();
Tensor::new(&[last_token], device)?.unsqueeze(0)?
};
let logits = self
.model
.decode(&decoder_token_ids, &encoder_output)?
.squeeze(0)?;
let logits = if repeat_penalty == 1. {
logits
} else {
let start_at = output_token_ids.len().saturating_sub(repeat_last_n);
candle_transformers::utils::apply_repeat_penalty(
&logits,
repeat_penalty,
&output_token_ids[start_at..],
)?
};
let next_token_id = logits_processor.sample(&logits)?;
if next_token_id as usize == self.config.eos_token_id {
break;
}
output_token_ids.push(next_token_id);
if let Some(text) = self.tokenizer.id_to_token(next_token_id) {
let text = text.replace('▁', " ").replace("<0x0A>", "\n");
decoded += &text;
}
}
Ok(serde_wasm_bindgen::to_value(
&ConditionalGenerationOutput {
generation: decoded,
},
)?)
}
}
#[wasm_bindgen]
impl ModelEncoder {
#[wasm_bindgen(constructor)]
pub fn load(
weights: Vec<u8>,
tokenizer: Vec<u8>,
config: Vec<u8>,
) -> Result<ModelEncoder, JsError> {
console_error_panic_hook::set_once();
console_log!("loading model");
let device = &Device::Cpu;
let vb = VarBuilder::from_buffered_safetensors(weights, DType::F32, device)?;
let mut config: Config = serde_json::from_slice(&config)?;
config.use_cache = false;
let tokenizer =
Tokenizer::from_bytes(&tokenizer).map_err(|m| JsError::new(&m.to_string()))?;
let model = T5EncoderModel::load(vb, &config)?;
Ok(Self { model, tokenizer })
}
pub fn decode(&mut self, input: JsValue) -> Result<JsValue, JsError> {
let device = &Device::Cpu;
let input: DecoderParams =
serde_wasm_bindgen::from_value(input).map_err(|m| JsError::new(&m.to_string()))?;
self.model.clear_kv_cache();
let sentences = input.sentences;
let normalize_embeddings = input.normalize_embeddings;
let n_sentences = sentences.len();
let mut all_embeddings = Vec::with_capacity(n_sentences);
for sentence in sentences {
let tokens = self
.tokenizer
.encode(sentence, true)
.map_err(|m| JsError::new(&m.to_string()))?
.get_ids()
.to_vec();
let token_ids = Tensor::new(&tokens[..], device)?.unsqueeze(0)?;
let embeddings = self.model.forward(&token_ids)?;
console_log!("generated embeddings {:?}", embeddings.shape());
// Apply some avg-pooling by taking the mean embedding value for all tokens (including padding)
let (_n_sentence, n_tokens, _hidden_size) = embeddings.dims3()?;
let embeddings = (embeddings.sum(1)? / (n_tokens as f64))?;
let embeddings = if normalize_embeddings {
embeddings.broadcast_div(&embeddings.sqr()?.sum_keepdim(1)?.sqrt()?)?
} else {
embeddings
};
console_log!("{:?}", embeddings.shape());
all_embeddings.push(embeddings.squeeze(0)?.to_vec1::<f32>()?);
}
Ok(serde_wasm_bindgen::to_value(&DecoderOutput {
embeddings: all_embeddings,
})?)
}
}
#[derive(serde::Serialize, serde::Deserialize)]
struct ConditionalGenerationOutput {
generation: String,
}
#[derive(serde::Serialize, serde::Deserialize)]
struct DecoderOutput {
embeddings: Vec<Vec<f32>>,
}
#[derive(serde::Serialize, serde::Deserialize)]
pub struct DecoderParams {
sentences: Vec<String>,
normalize_embeddings: bool,
}
#[derive(serde::Serialize, serde::Deserialize)]
pub struct ConditionalGenerationParams {
prompt: String,
temperature: f64,
seed: u64,
top_p: f64,
repeat_penalty: f32,
repeat_last_n: usize,
max_length: Option<usize>,
}
fn main() {
console_error_panic_hook::set_once();
}
| candle/candle-wasm-examples/t5/src/bin/m.rs/0 | {
"file_path": "candle/candle-wasm-examples/t5/src/bin/m.rs",
"repo_id": "candle",
"token_count": 3593
} |
use crate::languages::LANGUAGES;
use anyhow::Error as E;
use candle::{safetensors::Load, DType, Device, IndexOp, Tensor, D};
use candle_nn::{ops::softmax, VarBuilder};
pub use candle_transformers::models::whisper::{self as m, Config};
use rand::{distributions::Distribution, rngs::StdRng, SeedableRng};
use serde::{Deserialize, Serialize};
use tokenizers::Tokenizer;
use wasm_bindgen::prelude::*;
use yew_agent::{HandlerId, Public, WorkerLink};
#[wasm_bindgen]
extern "C" {
// Use `js_namespace` here to bind `console.log(..)` instead of just
// `log(..)`
#[wasm_bindgen(js_namespace = console)]
pub fn log(s: &str);
}
#[macro_export]
macro_rules! console_log {
// Note that this is using the `log` function imported above during
// `bare_bones`
($($t:tt)*) => ($crate::worker::log(&format_args!($($t)*).to_string()))
}
pub const DTYPE: DType = DType::F32;
pub enum Model {
Normal(m::model::Whisper),
Quantized(m::quantized_model::Whisper),
}
// Maybe we should use some traits rather than doing the dispatch for all these.
impl Model {
pub fn config(&self) -> &Config {
match self {
Self::Normal(m) => &m.config,
Self::Quantized(m) => &m.config,
}
}
pub fn encoder_forward(&mut self, x: &Tensor, flush: bool) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.encoder.forward(x, flush),
Self::Quantized(m) => m.encoder.forward(x, flush),
}
}
pub fn decoder_forward(
&mut self,
x: &Tensor,
xa: &Tensor,
flush: bool,
) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.decoder.forward(x, xa, flush),
Self::Quantized(m) => m.decoder.forward(x, xa, flush),
}
}
pub fn decoder_final_linear(&self, x: &Tensor) -> candle::Result<Tensor> {
match self {
Self::Normal(m) => m.decoder.final_linear(x),
Self::Quantized(m) => m.decoder.final_linear(x),
}
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DecodingResult {
pub tokens: Vec<u32>,
pub text: String,
pub avg_logprob: f64,
pub no_speech_prob: f64,
temperature: f64,
compression_ratio: f64,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Segment {
pub start: f64,
pub duration: f64,
pub dr: DecodingResult,
}
pub struct Decoder {
model: Model,
rng: rand::rngs::StdRng,
task: Option<Task>,
language: Option<String>,
is_multilingual: bool,
mel_filters: Vec<f32>,
timestamps: bool,
tokenizer: Tokenizer,
suppress_tokens: Tensor,
sot_token: u32,
transcribe_token: u32,
translate_token: u32,
eot_token: u32,
no_speech_token: u32,
no_timestamps_token: u32,
}
impl Decoder {
#[allow(clippy::too_many_arguments)]
fn new(
model: Model,
tokenizer: Tokenizer,
mel_filters: Vec<f32>,
device: &Device,
task: Option<Task>,
language: Option<String>,
is_multilingual: bool,
timestamps: bool,
) -> anyhow::Result<Self> {
let suppress_tokens: Vec<f32> = (0..model.config().vocab_size as u32)
.map(|i| {
if model.config().suppress_tokens.contains(&i) {
f32::NEG_INFINITY
} else {
0f32
}
})
.collect();
let no_timestamps_token = token_id(&tokenizer, m::NO_TIMESTAMPS_TOKEN)?;
let suppress_tokens = Tensor::new(suppress_tokens.as_slice(), device)?;
let sot_token = token_id(&tokenizer, m::SOT_TOKEN)?;
let transcribe_token = token_id(&tokenizer, m::TRANSCRIBE_TOKEN)?;
let translate_token = token_id(&tokenizer, m::TRANSLATE_TOKEN)?;
let eot_token = token_id(&tokenizer, m::EOT_TOKEN)?;
let no_speech_token = m::NO_SPEECH_TOKENS
.iter()
.find_map(|token| token_id(&tokenizer, token).ok());
let no_speech_token = match no_speech_token {
None => anyhow::bail!("unable to find any non-speech token"),
Some(n) => n,
};
let seed = 299792458;
Ok(Self {
model,
rng: StdRng::seed_from_u64(seed),
tokenizer,
mel_filters,
task,
timestamps,
language,
is_multilingual,
suppress_tokens,
sot_token,
transcribe_token,
translate_token,
eot_token,
no_speech_token,
no_timestamps_token,
})
}
fn decode(&mut self, mel: &Tensor, t: f64) -> anyhow::Result<DecodingResult> {
let model = &mut self.model;
let language_token = match (self.is_multilingual, &self.language) {
(true, None) => Some(detect_language(model, &self.tokenizer, mel)?),
(false, None) => None,
(true, Some(language)) => {
match token_id(&self.tokenizer, &format!("<|{:?}|>", self.language)) {
Ok(token_id) => Some(token_id),
Err(_) => anyhow::bail!("language {language} is not supported"),
}
}
(false, Some(_)) => {
anyhow::bail!("a language cannot be set for non-multilingual models")
}
};
let audio_features = model.encoder_forward(mel, true)?;
println!("audio features: {:?}", audio_features.dims());
let sample_len = model.config().max_target_positions / 2;
let mut sum_logprob = 0f64;
let mut no_speech_prob = f64::NAN;
let mut tokens = vec![self.sot_token];
if let Some(language_token) = language_token {
tokens.push(language_token);
}
match self.task {
None | Some(Task::Transcribe) => tokens.push(self.transcribe_token),
Some(Task::Translate) => tokens.push(self.translate_token),
}
if !self.timestamps {
tokens.push(self.no_timestamps_token);
}
for i in 0..sample_len {
let tokens_t = Tensor::new(tokens.as_slice(), mel.device())?;
// The model expects a batch dim but this inference loop does not handle
// it so we add it at this point.
let tokens_t = tokens_t.unsqueeze(0)?;
let ys = model.decoder_forward(&tokens_t, &audio_features, i == 0)?;
// Extract the no speech probability on the first iteration by looking at the first
// token logits and the probability for the according token.
if i == 0 {
let logits = model.decoder_final_linear(&ys.i(..1)?)?.i(0)?.i(0)?;
no_speech_prob = softmax(&logits, 0)?
.i(self.no_speech_token as usize)?
.to_scalar::<f32>()? as f64;
}
let (_, seq_len, _) = ys.dims3()?;
let logits = model
.decoder_final_linear(&ys.i((..1, seq_len - 1..))?)?
.i(0)?
.i(0)?;
// TODO: Besides suppress tokens, we should apply the heuristics from
// ApplyTimestampRules, i.e.:
// - Timestamps come in pairs, except before EOT.
// - Timestamps should be non-decreasing.
// - If the sum of the probabilities of timestamps is higher than any other tokens,
// only consider timestamps when sampling.
// https://github.com/openai/whisper/blob/e8622f9afc4eba139bf796c210f5c01081000472/whisper/decoding.py#L439
let logits = logits.broadcast_add(&self.suppress_tokens)?;
let next_token = if t > 0f64 {
let prs = softmax(&(&logits / t)?, 0)?;
let logits_v: Vec<f32> = prs.to_vec1()?;
let distr = rand::distributions::WeightedIndex::new(&logits_v)?;
distr.sample(&mut self.rng) as u32
} else {
let logits_v: Vec<f32> = logits.to_vec1()?;
logits_v
.iter()
.enumerate()
.max_by(|(_, u), (_, v)| u.total_cmp(v))
.map(|(i, _)| i as u32)
.unwrap()
};
tokens.push(next_token);
let prob = softmax(&logits, candle::D::Minus1)?
.i(next_token as usize)?
.to_scalar::<f32>()? as f64;
if next_token == self.eot_token || tokens.len() > model.config().max_target_positions {
break;
}
sum_logprob += prob.ln();
}
let text = self.tokenizer.decode(&tokens, true).map_err(E::msg)?;
let avg_logprob = sum_logprob / tokens.len() as f64;
Ok(DecodingResult {
tokens,
text,
avg_logprob,
no_speech_prob,
temperature: t,
compression_ratio: f64::NAN,
})
}
fn decode_with_fallback(&mut self, segment: &Tensor) -> anyhow::Result<DecodingResult> {
for (i, &t) in m::TEMPERATURES.iter().enumerate() {
let dr: Result<DecodingResult, _> = self.decode(segment, t);
if i == m::TEMPERATURES.len() - 1 {
return dr;
}
// On errors, we try again with a different temperature.
match dr {
Ok(dr) => {
let needs_fallback = dr.compression_ratio > m::COMPRESSION_RATIO_THRESHOLD
|| dr.avg_logprob < m::LOGPROB_THRESHOLD;
if !needs_fallback || dr.no_speech_prob > m::NO_SPEECH_THRESHOLD {
return Ok(dr);
}
}
Err(err) => {
console_log!("Error running at {t}: {err}")
}
}
}
unreachable!()
}
fn run(&mut self, mel: &Tensor) -> anyhow::Result<Vec<Segment>> {
let (_, _, content_frames) = mel.dims3()?;
let mut seek = 0;
let mut segments = vec![];
while seek < content_frames {
let time_offset = (seek * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
let segment_size = usize::min(content_frames - seek, m::N_FRAMES);
let mel_segment = mel.narrow(2, seek, segment_size)?;
let segment_duration = (segment_size * m::HOP_LENGTH) as f64 / m::SAMPLE_RATE as f64;
let dr = self.decode_with_fallback(&mel_segment)?;
seek += segment_size;
if dr.no_speech_prob > m::NO_SPEECH_THRESHOLD && dr.avg_logprob < m::LOGPROB_THRESHOLD {
console_log!("no speech detected, skipping {seek} {dr:?}");
continue;
}
let segment = Segment {
start: time_offset,
duration: segment_duration,
dr,
};
console_log!("{seek}: {segment:?}");
segments.push(segment)
}
Ok(segments)
}
pub fn load(md: ModelData) -> anyhow::Result<Self> {
let device = Device::Cpu;
let tokenizer = Tokenizer::from_bytes(&md.tokenizer).map_err(E::msg)?;
let mel_filters = safetensors::tensor::SafeTensors::deserialize(&md.mel_filters)?;
let mel_filters = mel_filters.tensor("mel_80")?.load(&device)?;
console_log!("loaded mel filters {:?}", mel_filters.shape());
let mel_filters = mel_filters.flatten_all()?.to_vec1::<f32>()?;
let config: Config = serde_json::from_slice(&md.config)?;
let model = if md.quantized {
let vb = candle_transformers::quantized_var_builder::VarBuilder::from_gguf_buffer(
&md.weights,
&device,
)?;
Model::Quantized(m::quantized_model::Whisper::load(&vb, config)?)
} else {
let vb = VarBuilder::from_buffered_safetensors(md.weights, m::DTYPE, &device)?;
Model::Normal(m::model::Whisper::load(&vb, config)?)
};
console_log!("done loading model");
let task = match md.task.as_deref() {
Some("translate") => Some(Task::Translate),
_ => Some(Task::Transcribe),
};
let decoder = Self::new(
model,
tokenizer,
mel_filters,
&device,
task,
md.language,
md.is_multilingual,
md.timestamps,
)?;
Ok(decoder)
}
pub fn convert_and_run(&mut self, wav_input: &[u8]) -> anyhow::Result<Vec<Segment>> {
let device = Device::Cpu;
let mut wav_input = std::io::Cursor::new(wav_input);
let wav_reader = hound::WavReader::new(&mut wav_input)?;
let spec = wav_reader.spec();
console_log!("loaded wav data: {spec:?}");
if spec.sample_rate != m::SAMPLE_RATE as u32 {
anyhow::bail!("wav file must have a {} sampling rate", m::SAMPLE_RATE);
}
let mut data = wav_reader.into_samples::<i16>().collect::<Vec<_>>();
data.truncate(data.len() / spec.channels as usize);
let mut pcm_data = Vec::with_capacity(data.len());
for d in data.into_iter() {
let d = d?;
pcm_data.push(d as f32 / 32768.)
}
console_log!("pcm data loaded {}", pcm_data.len());
let mel = crate::audio::pcm_to_mel(self.model.config(), &pcm_data, &self.mel_filters)?;
let mel_len = mel.len();
let n_mels = self.model.config().num_mel_bins;
let mel = Tensor::from_vec(mel, (1, n_mels, mel_len / n_mels), &device)?;
console_log!("loaded mel: {:?}", mel.dims());
let segments = self.run(&mel)?;
Ok(segments)
}
}
/// Returns the token id for the selected language.
pub fn detect_language(model: &mut Model, tokenizer: &Tokenizer, mel: &Tensor) -> Result<u32, E> {
console_log!("detecting language");
let (_bsize, _, seq_len) = mel.dims3()?;
let mel = mel.narrow(
2,
0,
usize::min(seq_len, model.config().max_source_positions),
)?;
let device = mel.device();
let language_token_ids = LANGUAGES
.iter()
.map(|(t, _)| token_id(tokenizer, &format!("<|{t}|>")))
.map(|e| e.map_err(E::msg))
.collect::<Result<Vec<_>, E>>()?;
let sot_token = token_id(tokenizer, m::SOT_TOKEN)?;
let audio_features = model.encoder_forward(&mel, true)?;
let tokens = Tensor::new(&[[sot_token]], device)?;
let language_token_ids = Tensor::new(language_token_ids.as_slice(), device)?;
let ys = model.decoder_forward(&tokens, &audio_features, true)?;
let logits = model.decoder_final_linear(&ys.i(..1)?)?.i(0)?.i(0)?;
let logits = logits.index_select(&language_token_ids, 0)?;
let probs = candle_nn::ops::softmax(&logits, D::Minus1)?;
let probs = probs.to_vec1::<f32>()?;
let mut probs = LANGUAGES.iter().zip(probs.iter()).collect::<Vec<_>>();
probs.sort_by(|(_, p1), (_, p2)| p2.total_cmp(p1));
for ((_, language), p) in probs.iter().take(5) {
println!("{language}: {p}")
}
let token = &format!("<|{}|>", probs[0].0 .0);
let language = token_id(tokenizer, token)?;
console_log!("detected language: {language} {token}");
Ok(language)
}
pub fn token_id(tokenizer: &Tokenizer, token: &str) -> candle::Result<u32> {
match tokenizer.token_to_id(token) {
None => candle::bail!("no token-id for {token}"),
Some(id) => Ok(id),
}
}
#[derive(Serialize, Deserialize, Clone, Copy, Debug)]
pub enum Task {
Transcribe,
Translate,
}
// Communication to the worker happens through bincode, the model weights and configs are fetched
// on the main thread and transferred via the following structure.
#[derive(Serialize, Deserialize)]
pub struct ModelData {
pub weights: Vec<u8>,
pub tokenizer: Vec<u8>,
pub mel_filters: Vec<u8>,
pub config: Vec<u8>,
pub quantized: bool,
pub timestamps: bool,
pub is_multilingual: bool,
pub language: Option<String>,
pub task: Option<String>,
}
pub struct Worker {
link: WorkerLink<Self>,
decoder: Option<Decoder>,
}
#[derive(Serialize, Deserialize)]
pub enum WorkerInput {
ModelData(ModelData),
DecodeTask { wav_bytes: Vec<u8> },
}
#[derive(Serialize, Deserialize)]
pub enum WorkerOutput {
Decoded(Vec<Segment>),
WeightsLoaded,
}
impl yew_agent::Worker for Worker {
type Input = WorkerInput;
type Message = ();
type Output = Result<WorkerOutput, String>;
type Reach = Public<Self>;
fn create(link: WorkerLink<Self>) -> Self {
Self {
link,
decoder: None,
}
}
fn update(&mut self, _msg: Self::Message) {
// no messaging
}
fn handle_input(&mut self, msg: Self::Input, id: HandlerId) {
let output = match msg {
WorkerInput::ModelData(md) => match Decoder::load(md) {
Ok(decoder) => {
self.decoder = Some(decoder);
Ok(WorkerOutput::WeightsLoaded)
}
Err(err) => Err(format!("model creation error {err:?}")),
},
WorkerInput::DecodeTask { wav_bytes } => match &mut self.decoder {
None => Err("model has not been set".to_string()),
Some(decoder) => decoder
.convert_and_run(&wav_bytes)
.map(WorkerOutput::Decoded)
.map_err(|e| e.to_string()),
},
};
self.link.respond(id, output);
}
fn name_of_resource() -> &'static str {
"worker.js"
}
fn resource_path_is_relative() -> bool {
true
}
}
| candle/candle-wasm-examples/whisper/src/worker.rs/0 | {
"file_path": "candle/candle-wasm-examples/whisper/src/worker.rs",
"repo_id": "candle",
"token_count": 8825
} |
[package]
name = "candle-wasm-tests"
version.workspace = true
edition.workspace = true
description = "WASM tests for candle"
keywords.workspace = true
categories.workspace = true
[dependencies]
candle = { workspace = true }
rand = { workspace = true }
getrandom = { version = "0.2", features = ["js"] }
[dev-dependencies]
wasm-bindgen-test = "0.3.0"
| candle/candle-wasm-tests/Cargo.toml/0 | {
"file_path": "candle/candle-wasm-tests/Cargo.toml",
"repo_id": "candle",
"token_count": 122
} |
# Chat UI
**Find the docs at [hf.co/docs/chat-ui](https://huggingface.co/docs/chat-ui/index).**
A chat interface using open source models, eg OpenAssistant or Llama. It is a SvelteKit app and it powers the [HuggingChat app on hf.co/chat](https://huggingface.co/chat).
0. [Quickstart](#quickstart)
1. [No Setup Deploy](#no-setup-deploy)
2. [Setup](#setup)
3. [Launch](#launch)
4. [Web Search](#web-search)
5. [Text Embedding Models](#text-embedding-models)
6. [Extra parameters](#extra-parameters)
7. [Common issues](#common-issues)
8. [Deploying to a HF Space](#deploying-to-a-hf-space)
9. [Building](#building)
## Quickstart
### Docker image
You can deploy a chat-ui instance in a single command using the docker image. Get your huggingface token from [here](https://huggingface.co/settings/tokens).
```bash
docker run -p 3000 -e HF_TOKEN=hf_*** -v db:/data ghcr.io/huggingface/chat-ui-db:latest
```
Take a look at the [`.env` file](https://github.com/huggingface/chat-ui/blob/main/.env) and the readme to see all the environment variables that you can set. We have endpoint support for all OpenAI API compatible local services as well as many other providers like Anthropic, Cloudflare, Google Vertex AI, etc.
### Local setup
You can quickly start a locally running chat-ui & LLM text-generation server thanks to chat-ui's [llama.cpp server support](https://huggingface.co/docs/chat-ui/configuration/models/providers/llamacpp).
**Step 1 (Start llama.cpp server):**
Install llama.cpp w/ brew (for Mac):
```bash
# install llama.cpp
brew install llama.cpp
```
or [build directly from the source](https://github.com/ggerganov/llama.cpp/blob/master/docs/build.md) for your target device:
```
git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp && make
```
Next, start the server with the [LLM of your choice](https://huggingface.co/models?library=gguf):
```bash
# start llama.cpp server (using hf.co/microsoft/Phi-3-mini-4k-instruct-gguf as an example)
llama-server --hf-repo microsoft/Phi-3-mini-4k-instruct-gguf --hf-file Phi-3-mini-4k-instruct-q4.gguf -c 4096
```
A local LLaMA.cpp HTTP Server will start on `http://localhost:8080`. Read more [here](https://huggingface.co/docs/chat-ui/configuration/models/providers/llamacpp).
**Step 3 (make sure you have MongoDb running locally):**
```bash
docker run -d -p 27017:27017 --name mongo-chatui mongo:latest
```
Read more [here](#database).
**Step 4 (clone chat-ui):**
```bash
git clone https://github.com/huggingface/chat-ui
cd chat-ui
```
**Step 5 (tell chat-ui to use local llama.cpp server):**
Add the following to your `.env.local`:
```ini
MODELS=`[
{
"name": "microsoft/Phi-3-mini-4k-instruct",
"endpoints": [{
"type" : "llamacpp",
"baseURL": "http://localhost:8080"
}],
},
]`
```
Read more [here](https://huggingface.co/docs/chat-ui/configuration/models/providers/llamacpp).
**Step 6 (start chat-ui):**
```bash
npm install
npm run dev -- --open
```
Read more [here](#launch).
<img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/chat-ui/llamacpp-dark.png" height="auto"/>
## No Setup Deploy
If you don't want to configure, setup, and launch your own Chat UI yourself, you can use this option as a fast deploy alternative.
You can deploy your own customized Chat UI instance with any supported [LLM](https://huggingface.co/models?pipeline_tag=text-generation&sort=trending) of your choice on [Hugging Face Spaces](https://huggingface.co/spaces). To do so, use the chat-ui template [available here](https://huggingface.co/new-space?template=huggingchat/chat-ui-template).
Set `HF_TOKEN` in [Space secrets](https://huggingface.co/docs/hub/spaces-overview#managing-secrets) to deploy a model with gated access or a model in a private repository. It's also compatible with [Inference for PROs](https://huggingface.co/blog/inference-pro) curated list of powerful models with higher rate limits. Make sure to create your personal token first in your [User Access Tokens settings](https://huggingface.co/settings/tokens).
Read the full tutorial [here](https://huggingface.co/docs/hub/spaces-sdks-docker-chatui#chatui-on-spaces).
## Setup
The default config for Chat UI is stored in the `.env` file. You will need to override some values to get Chat UI to run locally. This is done in `.env.local`.
Start by creating a `.env.local` file in the root of the repository. The bare minimum config you need to get Chat UI to run locally is the following:
```env
MONGODB_URL=<the URL to your MongoDB instance>
HF_TOKEN=<your access token>
```
### Database
The chat history is stored in a MongoDB instance, and having a DB instance available is needed for Chat UI to work.
You can use a local MongoDB instance. The easiest way is to spin one up using docker:
```bash
docker run -d -p 27017:27017 --name mongo-chatui mongo:latest
```
In which case the url of your DB will be `MONGODB_URL=mongodb://localhost:27017`.
Alternatively, you can use a [free MongoDB Atlas](https://www.mongodb.com/pricing) instance for this, Chat UI should fit comfortably within their free tier. After which you can set the `MONGODB_URL` variable in `.env.local` to match your instance.
### Hugging Face Access Token
If you use a remote inference endpoint, you will need a Hugging Face access token to run Chat UI locally. You can get one from [your Hugging Face profile](https://huggingface.co/settings/tokens).
## Launch
After you're done with the `.env.local` file you can run Chat UI locally with:
```bash
npm install
npm run dev
```
## Web Search
Chat UI features a powerful Web Search feature. It works by:
1. Generating an appropriate search query from the user prompt.
2. Performing web search and extracting content from webpages.
3. Creating embeddings from texts using a text embedding model.
4. From these embeddings, find the ones that are closest to the user query using a vector similarity search. Specifically, we use `inner product` distance.
5. Get the corresponding texts to those closest embeddings and perform [Retrieval-Augmented Generation](https://huggingface.co/papers/2005.11401) (i.e. expand user prompt by adding those texts so that an LLM can use this information).
## Text Embedding Models
By default (for backward compatibility), when `TEXT_EMBEDDING_MODELS` environment variable is not defined, [transformers.js](https://huggingface.co/docs/transformers.js) embedding models will be used for embedding tasks, specifically, [Xenova/gte-small](https://huggingface.co/Xenova/gte-small) model.
You can customize the embedding model by setting `TEXT_EMBEDDING_MODELS` in your `.env.local` file. For example:
```env
TEXT_EMBEDDING_MODELS = `[
{
"name": "Xenova/gte-small",
"displayName": "Xenova/gte-small",
"description": "locally running embedding",
"chunkCharLength": 512,
"endpoints": [
{"type": "transformersjs"}
]
},
{
"name": "intfloat/e5-base-v2",
"displayName": "intfloat/e5-base-v2",
"description": "hosted embedding model",
"chunkCharLength": 768,
"preQuery": "query: ", # See https://huggingface.co/intfloat/e5-base-v2#faq
"prePassage": "passage: ", # See https://huggingface.co/intfloat/e5-base-v2#faq
"endpoints": [
{
"type": "tei",
"url": "http://127.0.0.1:8080/",
"authorization": "TOKEN_TYPE TOKEN" // optional authorization field. Example: "Basic VVNFUjpQQVNT"
}
]
}
]`
```
The required fields are `name`, `chunkCharLength` and `endpoints`.
Supported text embedding backends are: [`transformers.js`](https://huggingface.co/docs/transformers.js), [`TEI`](https://github.com/huggingface/text-embeddings-inference) and [`OpenAI`](https://platform.openai.com/docs/guides/embeddings). `transformers.js` models run locally as part of `chat-ui`, whereas `TEI` models run in a different environment & accessed through an API endpoint. `openai` models are accessed through the [OpenAI API](https://platform.openai.com/docs/guides/embeddings).
When more than one embedding models are supplied in `.env.local` file, the first will be used by default, and the others will only be used on LLM's which configured `embeddingModel` to the name of the model.
## Extra parameters
### OpenID connect
The login feature is disabled by default and users are attributed a unique ID based on their browser. But if you want to use OpenID to authenticate your users, you can add the following to your `.env.local` file:
```env
OPENID_CONFIG=`{
PROVIDER_URL: "<your OIDC issuer>",
CLIENT_ID: "<your OIDC client ID>",
CLIENT_SECRET: "<your OIDC client secret>",
SCOPES: "openid profile",
TOLERANCE: // optional
RESOURCE: // optional
}`
```
These variables will enable the openID sign-in modal for users.
### Trusted header authentication
You can set the env variable `TRUSTED_EMAIL_HEADER` to point to the header that contains the user's email address. This will allow you to authenticate users from the header. This setup is usually combined with a proxy that will be in front of chat-ui and will handle the auth and set the header.
> [!WARNING]
> Make sure to only allow requests to chat-ui through your proxy which handles authentication, otherwise users could authenticate as anyone by setting the header manually! Only set this up if you understand the implications and know how to do it correctly.
Here is a list of header names for common auth providers:
- Tailscale Serve: `Tailscale-User-Login`
- Cloudflare Access: `Cf-Access-Authenticated-User-Email`
- oauth2-proxy: `X-Forwarded-Email`
### Theming
You can use a few environment variables to customize the look and feel of chat-ui. These are by default:
```env
PUBLIC_APP_NAME=ChatUI
PUBLIC_APP_ASSETS=chatui
PUBLIC_APP_COLOR=blue
PUBLIC_APP_DESCRIPTION="Making the community's best AI chat models available to everyone."
PUBLIC_APP_DATA_SHARING=
PUBLIC_APP_DISCLAIMER=
```
- `PUBLIC_APP_NAME` The name used as a title throughout the app.
- `PUBLIC_APP_ASSETS` Is used to find logos & favicons in `static/$PUBLIC_APP_ASSETS`, current options are `chatui` and `huggingchat`.
- `PUBLIC_APP_COLOR` Can be any of the [tailwind colors](https://tailwindcss.com/docs/customizing-colors#default-color-palette).
- `PUBLIC_APP_DATA_SHARING` Can be set to 1 to add a toggle in the user settings that lets your users opt-in to data sharing with models creator.
- `PUBLIC_APP_DISCLAIMER` If set to 1, we show a disclaimer about generated outputs on login.
### Web Search config
You can enable the web search through an API by adding `YDC_API_KEY` ([docs.you.com](https://docs.you.com)) or `SERPER_API_KEY` ([serper.dev](https://serper.dev/)) or `SERPAPI_KEY` ([serpapi.com](https://serpapi.com/)) or `SERPSTACK_API_KEY` ([serpstack.com](https://serpstack.com/)) or `SEARCHAPI_KEY` ([searchapi.io](https://www.searchapi.io/)) to your `.env.local`.
You can also simply enable the local google websearch by setting `USE_LOCAL_WEBSEARCH=true` in your `.env.local` or specify a SearXNG instance by adding the query URL to `SEARXNG_QUERY_URL`.
You can enable javascript when parsing webpages to improve compatibility with `WEBSEARCH_JAVASCRIPT=true` at the cost of increased CPU usage. You'll want at least 4 cores when enabling.
### Custom models
You can customize the parameters passed to the model or even use a new model by updating the `MODELS` variable in your `.env.local`. The default one can be found in `.env` and looks like this :
```env
MODELS=`[
{
"name": "mistralai/Mistral-7B-Instruct-v0.2",
"displayName": "mistralai/Mistral-7B-Instruct-v0.2",
"description": "Mistral 7B is a new Apache 2.0 model, released by Mistral AI that outperforms Llama2 13B in benchmarks.",
"websiteUrl": "https://mistral.ai/news/announcing-mistral-7b/",
"preprompt": "",
"chatPromptTemplate" : "<s>{{#each messages}}{{#ifUser}}[INST] {{#if @first}}{{#if @root.preprompt}}{{@root.preprompt}}\n{{/if}}{{/if}}{{content}} [/INST]{{/ifUser}}{{#ifAssistant}}{{content}}</s>{{/ifAssistant}}{{/each}}",
"parameters": {
"temperature": 0.3,
"top_p": 0.95,
"repetition_penalty": 1.2,
"top_k": 50,
"truncate": 3072,
"max_new_tokens": 1024,
"stop": ["</s>"]
},
"promptExamples": [
{
"title": "Write an email from bullet list",
"prompt": "As a restaurant owner, write a professional email to the supplier to get these products every week: \n\n- Wine (x10)\n- Eggs (x24)\n- Bread (x12)"
}, {
"title": "Code a snake game",
"prompt": "Code a basic snake game in python, give explanations for each step."
}, {
"title": "Assist in a task",
"prompt": "How do I make a delicious lemon cheesecake?"
}
]
}
]`
```
You can change things like the parameters, or customize the preprompt to better suit your needs. You can also add more models by adding more objects to the array, with different preprompts for example.
#### chatPromptTemplate
When querying the model for a chat response, the `chatPromptTemplate` template is used. `messages` is an array of chat messages, it has the format `[{ content: string }, ...]`. To identify if a message is a user message or an assistant message the `ifUser` and `ifAssistant` block helpers can be used.
The following is the default `chatPromptTemplate`, although newlines and indentiation have been added for readability. You can find the prompts used in production for HuggingChat [here](https://github.com/huggingface/chat-ui/blob/main/PROMPTS.md).
```prompt
{{preprompt}}
{{#each messages}}
{{#ifUser}}{{@root.userMessageToken}}{{content}}{{@root.userMessageEndToken}}{{/ifUser}}
{{#ifAssistant}}{{@root.assistantMessageToken}}{{content}}{{@root.assistantMessageEndToken}}{{/ifAssistant}}
{{/each}}
{{assistantMessageToken}}
```
#### Multi modal model
We currently support [IDEFICS](https://huggingface.co/blog/idefics) (hosted on TGI), OpenAI and Claude 3 as multimodal models. You can enable it by setting `multimodal: true` in your `MODELS` configuration. For IDEFICS, you must have a [PRO HF Api token](https://huggingface.co/settings/tokens). For OpenAI, see the [OpenAI section](#openai-api-compatible-models). For Anthropic, see the [Anthropic section](#anthropic).
```env
{
"name": "HuggingFaceM4/idefics-80b-instruct",
"multimodal" : true,
"description": "IDEFICS is the new multimodal model by Hugging Face.",
"preprompt": "",
"chatPromptTemplate" : "{{#each messages}}{{#ifUser}}User: {{content}}{{/ifUser}}<end_of_utterance>\nAssistant: {{#ifAssistant}}{{content}}\n{{/ifAssistant}}{{/each}}",
"parameters": {
"temperature": 0.1,
"top_p": 0.95,
"repetition_penalty": 1.2,
"top_k": 12,
"truncate": 1000,
"max_new_tokens": 1024,
"stop": ["<end_of_utterance>", "User:", "\nUser:"]
}
}
```
#### Running your own models using a custom endpoint
If you want to, instead of hitting models on the Hugging Face Inference API, you can run your own models locally.
A good option is to hit a [text-generation-inference](https://github.com/huggingface/text-generation-inference) endpoint. This is what is done in the official [Chat UI Spaces Docker template](https://huggingface.co/new-space?template=huggingchat/chat-ui-template) for instance: both this app and a text-generation-inference server run inside the same container.
To do this, you can add your own endpoints to the `MODELS` variable in `.env.local`, by adding an `"endpoints"` key for each model in `MODELS`.
```env
{
// rest of the model config here
"endpoints": [{
"type" : "tgi",
"url": "https://HOST:PORT",
}]
}
```
If `endpoints` are left unspecified, ChatUI will look for the model on the hosted Hugging Face inference API using the model name.
##### OpenAI API compatible models
Chat UI can be used with any API server that supports OpenAI API compatibility, for example [text-generation-webui](https://github.com/oobabooga/text-generation-webui/tree/main/extensions/openai), [LocalAI](https://github.com/go-skynet/LocalAI), [FastChat](https://github.com/lm-sys/FastChat/blob/main/docs/openai_api.md), [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), and [ialacol](https://github.com/chenhunghan/ialacol) and [vllm](https://docs.vllm.ai/en/latest/serving/openai_compatible_server.html).
The following example config makes Chat UI works with [text-generation-webui](https://github.com/oobabooga/text-generation-webui/tree/main/extensions/openai), the `endpoint.baseUrl` is the url of the OpenAI API compatible server, this overrides the baseUrl to be used by OpenAI instance. The `endpoint.completion` determine which endpoint to be used, default is `chat_completions` which uses `v1/chat/completions`, change to `endpoint.completion` to `completions` to use the `v1/completions` endpoint.
Parameters not supported by OpenAI (e.g., top_k, repetition_penalty, etc.) must be set in the extraBody of endpoints. Be aware that setting them in parameters will cause them to be omitted.
```
MODELS=`[
{
"name": "text-generation-webui",
"id": "text-generation-webui",
"parameters": {
"temperature": 0.9,
"top_p": 0.95,
"max_new_tokens": 1024,
"stop": []
},
"endpoints": [{
"type" : "openai",
"baseURL": "http://localhost:8000/v1",
"extraBody": {
"repetition_penalty": 1.2,
"top_k": 50,
"truncate": 1000
}
}]
}
]`
```
The `openai` type includes official OpenAI models. You can add, for example, GPT4/GPT3.5 as a "openai" model:
```
OPENAI_API_KEY=#your openai api key here
MODELS=`[{
"name": "gpt-4",
"displayName": "GPT 4",
"endpoints" : [{
"type": "openai"
}]
},
{
"name": "gpt-3.5-turbo",
"displayName": "GPT 3.5 Turbo",
"endpoints" : [{
"type": "openai"
}]
}]`
```
You may also consume any model provider that provides compatible OpenAI API endpoint. For example, you may self-host [Portkey](https://github.com/Portkey-AI/gateway) gateway and experiment with Claude or GPTs offered by Azure OpenAI. Example for Claude from Anthropic:
```
MODELS=`[{
"name": "claude-2.1",
"displayName": "Claude 2.1",
"description": "Anthropic has been founded by former OpenAI researchers...",
"parameters": {
"temperature": 0.5,
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "openai",
"baseURL": "https://gateway.example.com/v1",
"defaultHeaders": {
"x-portkey-config": '{"provider":"anthropic","api_key":"sk-ant-abc...xyz"}'
}
}
]
}]`
```
Example for GPT 4 deployed on Azure OpenAI:
```
MODELS=`[{
"id": "gpt-4-1106-preview",
"name": "gpt-4-1106-preview",
"displayName": "gpt-4-1106-preview",
"parameters": {
"temperature": 0.5,
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "openai",
"baseURL": "https://{resource-name}.openai.azure.com/openai/deployments/{deployment-id}",
"defaultHeaders": {
"api-key": "{api-key}"
},
"defaultQuery": {
"api-version": "2023-05-15"
}
}
]
}]`
```
Or try Mistral from [Deepinfra](https://deepinfra.com/mistralai/Mistral-7B-Instruct-v0.1/api?example=openai-http):
> Note, apiKey can either be set custom per endpoint, or globally using `OPENAI_API_KEY` variable.
```
MODELS=`[{
"name": "mistral-7b",
"displayName": "Mistral 7B",
"description": "A 7B dense Transformer, fast-deployed and easily customisable. Small, yet powerful for a variety of use cases. Supports English and code, and a 8k context window.",
"parameters": {
"temperature": 0.5,
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "openai",
"baseURL": "https://api.deepinfra.com/v1/openai",
"apiKey": "abc...xyz"
}
]
}]`
```
##### Llama.cpp API server
chat-ui also supports the llama.cpp API server directly without the need for an adapter. You can do this using the `llamacpp` endpoint type.
If you want to run Chat UI with llama.cpp, you can do the following, using [microsoft/Phi-3-mini-4k-instruct-gguf](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct-gguf) as an example model:
```bash
# install llama.cpp
brew install llama.cpp
# start llama.cpp server
llama-server --hf-repo microsoft/Phi-3-mini-4k-instruct-gguf --hf-file Phi-3-mini-4k-instruct-q4.gguf -c 4096
```
```env
MODELS=`[
{
"name": "Local Zephyr",
"chatPromptTemplate": "<|system|>\n{{preprompt}}</s>\n{{#each messages}}{{#ifUser}}<|user|>\n{{content}}</s>\n<|assistant|>\n{{/ifUser}}{{#ifAssistant}}{{content}}</s>\n{{/ifAssistant}}{{/each}}",
"parameters": {
"temperature": 0.1,
"top_p": 0.95,
"repetition_penalty": 1.2,
"top_k": 50,
"truncate": 1000,
"max_new_tokens": 2048,
"stop": ["</s>"]
},
"endpoints": [
{
"url": "http://127.0.0.1:8080",
"type": "llamacpp"
}
]
}
]`
```
Start chat-ui with `npm run dev` and you should be able to chat with Zephyr locally.
#### Ollama
We also support the Ollama inference server. Spin up a model with
```cli
ollama run mistral
```
Then specify the endpoints like so:
```env
MODELS=`[
{
"name": "Ollama Mistral",
"chatPromptTemplate": "<s>{{#each messages}}{{#ifUser}}[INST] {{#if @first}}{{#if @root.preprompt}}{{@root.preprompt}}\n{{/if}}{{/if}} {{content}} [/INST]{{/ifUser}}{{#ifAssistant}}{{content}}</s> {{/ifAssistant}}{{/each}}",
"parameters": {
"temperature": 0.1,
"top_p": 0.95,
"repetition_penalty": 1.2,
"top_k": 50,
"truncate": 3072,
"max_new_tokens": 1024,
"stop": ["</s>"]
},
"endpoints": [
{
"type": "ollama",
"url" : "http://127.0.0.1:11434",
"ollamaName" : "mistral"
}
]
}
]`
```
#### Anthropic
We also support Anthropic models (including multimodal ones via `multmodal: true`) through the official SDK. You may provide your API key via the `ANTHROPIC_API_KEY` env variable, or alternatively, through the `endpoints.apiKey` as per the following example.
```
MODELS=`[
{
"name": "claude-3-haiku-20240307",
"displayName": "Claude 3 Haiku",
"description": "Fastest and most compact model for near-instant responsiveness",
"multimodal": true,
"parameters": {
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "anthropic",
// optionals
"apiKey": "sk-ant-...",
"baseURL": "https://api.anthropic.com",
"defaultHeaders": {},
"defaultQuery": {}
}
]
},
{
"name": "claude-3-sonnet-20240229",
"displayName": "Claude 3 Sonnet",
"description": "Ideal balance of intelligence and speed",
"multimodal": true,
"parameters": {
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "anthropic",
// optionals
"apiKey": "sk-ant-...",
"baseURL": "https://api.anthropic.com",
"defaultHeaders": {},
"defaultQuery": {}
}
]
},
{
"name": "claude-3-opus-20240229",
"displayName": "Claude 3 Opus",
"description": "Most powerful model for highly complex tasks",
"multimodal": true,
"parameters": {
"max_new_tokens": 4096
},
"endpoints": [
{
"type": "anthropic",
// optionals
"apiKey": "sk-ant-...",
"baseURL": "https://api.anthropic.com",
"defaultHeaders": {},
"defaultQuery": {}
}
]
}
]`
```
We also support using Anthropic models running on Vertex AI. Authentication is done using Google Application Default Credentials. Project ID can be provided through the `endpoints.projectId` as per the following example:
```
MODELS=`[
{
"name": "claude-3-sonnet@20240229",
"displayName": "Claude 3 Sonnet",
"description": "Ideal balance of intelligence and speed",
"multimodal": true,
"parameters": {
"max_new_tokens": 4096,
},
"endpoints": [
{
"type": "anthropic-vertex",
"region": "us-central1",
"projectId": "gcp-project-id",
// optionals
"defaultHeaders": {},
"defaultQuery": {}
}
]
},
{
"name": "claude-3-haiku@20240307",
"displayName": "Claude 3 Haiku",
"description": "Fastest, most compact model for near-instant responsiveness",
"multimodal": true,
"parameters": {
"max_new_tokens": 4096
},
"endpoints": [
{
"type": "anthropic-vertex",
"region": "us-central1",
"projectId": "gcp-project-id",
// optionals
"defaultHeaders": {},
"defaultQuery": {}
}
]
}
]`
```
#### Amazon
You can also specify your Amazon SageMaker instance as an endpoint for chat-ui. The config goes like this:
```env
"endpoints": [
{
"type" : "aws",
"service" : "sagemaker"
"url": "",
"accessKey": "",
"secretKey" : "",
"sessionToken": "",
"region": "",
"weight": 1
}
]
```
You can also set `"service" : "lambda"` to use a lambda instance.
You can get the `accessKey` and `secretKey` from your AWS user, under programmatic access.
#### Cloudflare Workers AI
You can also use Cloudflare Workers AI to run your own models with serverless inference.
You will need to have a Cloudflare account, then get your [account ID](https://developers.cloudflare.com/fundamentals/setup/find-account-and-zone-ids/) as well as your [API token](https://developers.cloudflare.com/workers-ai/get-started/rest-api/#1-get-api-token-and-account-id) for Workers AI.
You can either specify them directly in your `.env.local` using the `CLOUDFLARE_ACCOUNT_ID` and `CLOUDFLARE_API_TOKEN` variables, or you can set them directly in the endpoint config.
You can find the list of models available on Cloudflare [here](https://developers.cloudflare.com/workers-ai/models/#text-generation).
```env
{
"name" : "nousresearch/hermes-2-pro-mistral-7b",
"tokenizer": "nousresearch/hermes-2-pro-mistral-7b",
"parameters": {
"stop": ["<|im_end|>"]
},
"endpoints" : [
{
"type" : "cloudflare"
<!-- optionally specify these
"accountId": "your-account-id",
"authToken": "your-api-token"
-->
}
]
}
```
#### Cohere
You can also use Cohere to run their models directly from chat-ui. You will need to have a Cohere account, then get your [API token](https://dashboard.cohere.com/api-keys). You can either specify it directly in your `.env.local` using the `COHERE_API_TOKEN` variable, or you can set it in the endpoint config.
Here is an example of a Cohere model config. You can set which model you want to use by setting the `id` field to the model name.
```env
{
"name" : "CohereForAI/c4ai-command-r-v01",
"id": "command-r",
"description": "C4AI Command-R is a research release of a 35 billion parameter highly performant generative model",
"endpoints": [
{
"type": "cohere",
<!-- optionally specify these, or use COHERE_API_TOKEN
"apiKey": "your-api-token"
-->
}
]
}
```
##### Google Vertex models
Chat UI can connect to the google Vertex API endpoints ([List of supported models](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models)).
To enable:
1. [Select](https://console.cloud.google.com/project) or [create](https://cloud.google.com/resource-manager/docs/creating-managing-projects#creating_a_project) a Google Cloud project.
1. [Enable billing for your project](https://cloud.google.com/billing/docs/how-to/modify-project).
1. [Enable the Vertex AI API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com).
1. [Set up authentication with a service account](https://cloud.google.com/docs/authentication/getting-started)
so you can access the API from your local workstation.
The service account credentials file can be imported as an environmental variable:
```env
GOOGLE_APPLICATION_CREDENTIALS = clientid.json
```
Make sure your docker container has access to the file and the variable is correctly set.
Afterwards Google Vertex endpoints can be configured as following:
```
MODELS=`[
//...
{
"name": "gemini-1.5-pro",
"displayName": "Vertex Gemini Pro 1.5",
"multimodal": true,
"endpoints" : [{
"type": "vertex",
"project": "abc-xyz",
"location": "europe-west3",
"extraBody": {
"model_version": "gemini-1.5-pro-preview-0409",
},
// Optional
"safetyThreshold": "BLOCK_MEDIUM_AND_ABOVE",
"apiEndpoint": "", // alternative api endpoint url,
"tools": [{
"googleSearchRetrieval": {
"disableAttribution": true
}
}],
"multimodal": {
"image": {
"supportedMimeTypes": ["image/png", "image/jpeg", "image/webp"],
"preferredMimeType": "image/png",
"maxSizeInMB": 5,
"maxWidth": 2000,
"maxHeight": 1000,
}
}
}]
},
]`
```
##### LangServe
LangChain applications that are deployed using LangServe can be called with the following config:
```
MODELS=`[
//...
{
"name": "summarization-chain", //model-name
"endpoints" : [{
"type": "langserve",
"url" : "http://127.0.0.1:8100",
}]
},
]`
```
### Custom endpoint authorization
#### Basic and Bearer
Custom endpoints may require authorization, depending on how you configure them. Authentication will usually be set either with `Basic` or `Bearer`.
For `Basic` we will need to generate a base64 encoding of the username and password.
`echo -n "USER:PASS" | base64`
> VVNFUjpQQVNT
For `Bearer` you can use a token, which can be grabbed from [here](https://huggingface.co/settings/tokens).
You can then add the generated information and the `authorization` parameter to your `.env.local`.
```env
"endpoints": [
{
"url": "https://HOST:PORT",
"authorization": "Basic VVNFUjpQQVNT",
}
]
```
Please note that if `HF_TOKEN` is also set or not empty, it will take precedence.
#### Models hosted on multiple custom endpoints
If the model being hosted will be available on multiple servers/instances add the `weight` parameter to your `.env.local`. The `weight` will be used to determine the probability of requesting a particular endpoint.
```env
"endpoints": [
{
"url": "https://HOST:PORT",
"weight": 1
},
{
"url": "https://HOST:PORT",
"weight": 2
}
...
]
```
#### Client Certificate Authentication (mTLS)
Custom endpoints may require client certificate authentication, depending on how you configure them. To enable mTLS between Chat UI and your custom endpoint, you will need to set the `USE_CLIENT_CERTIFICATE` to `true`, and add the `CERT_PATH` and `KEY_PATH` parameters to your `.env.local`. These parameters should point to the location of the certificate and key files on your local machine. The certificate and key files should be in PEM format. The key file can be encrypted with a passphrase, in which case you will also need to add the `CLIENT_KEY_PASSWORD` parameter to your `.env.local`.
If you're using a certificate signed by a private CA, you will also need to add the `CA_PATH` parameter to your `.env.local`. This parameter should point to the location of the CA certificate file on your local machine.
If you're using a self-signed certificate, e.g. for testing or development purposes, you can set the `REJECT_UNAUTHORIZED` parameter to `false` in your `.env.local`. This will disable certificate validation, and allow Chat UI to connect to your custom endpoint.
#### Specific Embedding Model
A model can use any of the embedding models defined in `.env.local`, (currently used when web searching),
by default it will use the first embedding model, but it can be changed with the field `embeddingModel`:
```env
TEXT_EMBEDDING_MODELS = `[
{
"name": "Xenova/gte-small",
"chunkCharLength": 512,
"endpoints": [
{"type": "transformersjs"}
]
},
{
"name": "intfloat/e5-base-v2",
"chunkCharLength": 768,
"endpoints": [
{"type": "tei", "url": "http://127.0.0.1:8080/", "authorization": "Basic VVNFUjpQQVNT"},
{"type": "tei", "url": "http://127.0.0.1:8081/"}
]
}
]`
MODELS=`[
{
"name": "Ollama Mistral",
"chatPromptTemplate": "...",
"embeddingModel": "intfloat/e5-base-v2"
"parameters": {
...
},
"endpoints": [
...
]
}
]`
```
### Reasoning Models
ChatUI supports specialized reasoning/Chain-of-Thought (CoT) models through the `reasoning` configuration field. When properly configured, this displays a UI widget that allows users to view or collapse the model’s reasoning steps. We support three types of reasoning parsing:
#### Token-Based Delimitations
For models like DeepSeek R1, token-based delimitations can be used to identify reasoning steps. This is done by specifying the `beginToken` and `endToken` fields in the `reasoning` configuration.
Example configuration for DeepSeek R1 (token-based):
```json
{
"name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
// ...
"reasoning": {
"type": "tokens",
"beginToken": "<think>",
"endToken": "</think>"
}
}
```
#### Summarizing the Chain of Thought
For models like QwQ, which return a chain of thought but do not explicitly provide a final answer, the `summarize` type can be used. This automatically summarizes the reasoning steps using the `TASK_MODEL` (or the first model in the configuration if `TASK_MODEL` is not specified) and displays the summary as the final answer.
Example configuration for QwQ (summarize-based):
```json
{
"name": "Qwen/QwQ-32B-Preview",
// ...
"reasoning": {
"type": "summarize"
}
}
```
#### Regex-Based Delimitations
In some cases, the final answer can be extracted from the model output using a regular expression. This is achieved by specifying the `regex` field in the `reasoning` configuration. For example, if your model wraps the final answer in a `\boxed{}` tag, you can use the following configuration:
```json
{
"name": "model/yourmodel",
// ...
"reasoning": {
"type": "regex",
"regex": "\\\\boxed\\{(.+?)\\}"
}
}
```
## Common issues
### 403：You don't have access to this conversation
Most likely you are running chat-ui over HTTP. The recommended option is to setup something like NGINX to handle HTTPS and proxy the requests to chat-ui. If you really need to run over HTTP you can add `COOKIE_SECURE=false` and `COOKIE_SAMESITE=lax` to your `.env.local`.
Make sure to set your `PUBLIC_ORIGIN` in your `.env.local` to the correct URL as well.
## Deploying to a HF Space
Create a `DOTENV_LOCAL` secret to your HF space with the content of your .env.local, and they will be picked up automatically when you run.
## Building
To create a production version of your app:
```bash
npm run build
```
You can preview the production build with `npm run preview`.
> To deploy your app, you may need to install an [adapter](https://kit.svelte.dev/docs/adapters) for your target environment.
## Config changes for HuggingChat
The config file for HuggingChat is stored in the `chart/env/prod.yaml` file. It is the source of truth for the environment variables used for our CI/CD pipeline. For HuggingChat, as we need to customize the app color, as well as the base path, we build a custom docker image. You can find the workflow here.
> [!TIP]
> If you want to make changes to the model config used in production for HuggingChat, you should do so against `chart/env/prod.yaml`.
### Running a copy of HuggingChat locally
If you want to run an exact copy of HuggingChat locally, you will need to do the following first:
1. Create an [OAuth App on the hub](https://huggingface.co/settings/applications/new) with `openid profile email` permissions. Make sure to set the callback URL to something like `http://localhost:5173/chat/login/callback` which matches the right path for your local instance.
2. Create a [HF Token](https://huggingface.co/settings/tokens) with your Hugging Face account. You will need a Pro account to be able to access some of the larger models available through HuggingChat.
3. Create a free account with [serper.dev](https://serper.dev/) (you will get 2500 free search queries)
4. Run an instance of mongoDB, however you want. (Local or remote)
You can then create a new `.env.SECRET_CONFIG` file with the following content
```env
MONGODB_URL=<link to your mongo DB from step 4>
HF_TOKEN=<your HF token from step 2>
OPENID_CONFIG=`{
PROVIDER_URL: "https://huggingface.co",
CLIENT_ID: "<your client ID from step 1>",
CLIENT_SECRET: "<your client secret from step 1>",
}`
SERPER_API_KEY=<your serper API key from step 3>
MESSAGES_BEFORE_LOGIN=<can be any numerical value, or set to 0 to require login>
```
You can then run `npm run updateLocalEnv` in the root of chat-ui. This will create a `.env.local` file which combines the `chart/env/prod.yaml` and the `.env.SECRET_CONFIG` file. You can then run `npm run dev` to start your local instance of HuggingChat.
### Populate database
> [!WARNING]
> The `MONGODB_URL` used for this script will be fetched from `.env.local`. Make sure it's correct! The command runs directly on the database.
You can populate the database using faker data using the `populate` script:
```bash
npm run populate <flags here>
```
At least one flag must be specified, the following flags are available:
- `reset` - resets the database
- `all` - populates all tables
- `users` - populates the users table
- `settings` - populates the settings table for existing users
- `assistants` - populates the assistants table for existing users
- `conversations` - populates the conversations table for existing users
For example, you could use it like so:
```bash
npm run populate reset
```
to clear out the database. Then login in the app to create your user and run the following command:
```bash
npm run populate users settings assistants conversations
```
to populate the database with fake data, including fake conversations and assistants for your user.
### Building the docker images locally
You can build the docker images locally using the following commands:
```bash
docker build -t chat-ui-db:latest --build-arg INCLUDE_DB=true .
docker build -t chat-ui:latest --build-arg INCLUDE_DB=false .
docker build -t huggingchat:latest --build-arg INCLUDE_DB=false --build-arg APP_BASE=/chat --build-arg PUBLIC_APP_COLOR=yellow .
```
| chat-ui/README.md/0 | {
"file_path": "chat-ui/README.md",
"repo_id": "chat-ui",
"token_count": 14260
} |
# Text Embedding Models
By default (for backward compatibility), when `TEXT_EMBEDDING_MODELS` environment variable is not defined, [transformers.js](https://huggingface.co/docs/transformers.js) embedding models will be used for embedding tasks, specifically, the [Xenova/gte-small](https://huggingface.co/Xenova/gte-small) model.
You can customize the embedding model by setting `TEXT_EMBEDDING_MODELS` in your `.env.local` file where the required fields are `name`, `chunkCharLength` and `endpoints`.
Supported text embedding backends are: [`transformers.js`](https://huggingface.co/docs/transformers.js), [`TEI`](https://github.com/huggingface/text-embeddings-inference) and [`OpenAI`](https://platform.openai.com/docs/guides/embeddings). `transformers.js` models run locally as part of `chat-ui`, whereas `TEI` models run in a different environment & accessed through an API endpoint. `openai` models are accessed through the [OpenAI API](https://platform.openai.com/docs/guides/embeddings).
When more than one embedding models are supplied in `.env.local` file, the first will be used by default, and the others will only be used on LLM's which configured `embeddingModel` to the name of the model.
## Transformers.js
The Transformers.js backend uses local CPU for the embedding which can be quite slow. If possible, consider using TEI or OpenAI embeddings instead if you use web search frequently, as performance will improve significantly.
```ini
TEXT_EMBEDDING_MODELS = `[
{
"name": "Xenova/gte-small",
"displayName": "Xenova/gte-small",
"description": "locally running embedding",
"chunkCharLength": 512,
"endpoints": [
{ "type": "transformersjs" }
]
}
]`
```
## Text Embeddings Inference (TEI)
> Text Embeddings Inference (TEI) is a comprehensive toolkit designed for efficient deployment and serving of open source text embeddings models. It enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE, and E5.
Some recommended models at the time of writing (May 2024) are `Snowflake/snowflake-arctic-embed-m` and `BAAI/bge-large-en-v1.5`. You may run TEI locally with GPU support via Docker:
`docker run --gpus all -p 8080:80 -v tei-data:/data --name tei ghcr.io/huggingface/text-embeddings-inference:1.2 --model-id YOUR/HF_MODEL`
You can then hook this up to your Chat UI instance with the following configuration.
```ini
TEXT_EMBEDDING_MODELS=`[
{
"name": "YOUR/HF_MODEL",
"displayName": "YOUR/HF_MODEL",
"preQuery": "Check the model documentation for the preQuery. Not all models have one",
"prePassage": "Check the model documentation for the prePassage. Not all models have one",
"chunkCharLength": 512,
"endpoints": [{
"type": "tei",
"url": "http://127.0.0.1:8080/"
}]
}
]`
```
Examples for `Snowflake/snowflake-arctic-embed-m` and `BAAI/bge-large-en-v1.5`:
```ini
TEXT_EMBEDDING_MODELS=`[
{
"name": "Snowflake/snowflake-arctic-embed-m",
"displayName": "Snowflake/snowflake-arctic-embed-m",
"preQuery": "Represent this sentence for searching relevant passages: ",
"chunkCharLength": 512,
"endpoints": [{
"type": "tei",
"url": "http://127.0.0.1:8080/"
}]
},{
"name": "BAAI/bge-large-en-v1.5",
"displayName": "BAAI/bge-large-en-v1.5",
"chunkCharLength": 512,
"endpoints": [{
"type": "tei",
"url": "http://127.0.0.1:8080/"
}]
}
]`
```
## OpenAI
It's also possible to host your own OpenAI API compatible embedding models. [`Infinity`](https://github.com/michaelfeil/infinity) is one example. You may run it locally with Docker:
`docker run -it --gpus all -v infinity-data:/app/.cache -p 7997:7997 michaelf34/infinity:latest v2 --model-id nomic-ai/nomic-embed-text-v1 --port 7997`
You can then hook this up to your Chat UI instance with the following configuration.
```ini
TEXT_EMBEDDING_MODELS=`[
{
"name": "nomic-ai/nomic-embed-text-v1",
"displayName": "nomic-ai/nomic-embed-text-v1",
"chunkCharLength": 512,
"model": {
"name": "nomic-ai/nomic-embed-text-v1"
},
"endpoints": [
{
"type": "openai",
"url": "https://127.0.0.1:7997/embeddings"
}
]
}
]`
```
| chat-ui/docs/source/configuration/embeddings.md/0 | {
"file_path": "chat-ui/docs/source/configuration/embeddings.md",
"repo_id": "chat-ui",
"token_count": 1568
} |
# Configuration Overview
Chat UI handles configuration with environment variables. The default config for Chat UI is stored in the `.env` file, which you may use as a reference. You will need to override some values to get Chat UI to run locally. This can be done in `.env.local` or via your environment. The bare minimum configuration to get Chat UI running is:
```ini
MONGODB_URL=mongodb://localhost:27017
HF_TOKEN=your_token
```
The following sections detail various sections of the app you may want to configure.
| chat-ui/docs/source/configuration/overview.md/0 | {
"file_path": "chat-ui/docs/source/configuration/overview.md",
"repo_id": "chat-ui",
"token_count": 130
} |
import fs from "fs";
import yaml from "js-yaml";
const file = fs.readFileSync("chart/env/prod.yaml", "utf8");
// have to do a weird stringify/parse because of some node error
const prod = JSON.parse(JSON.stringify(yaml.load(file)));
const vars = prod.envVars as Record<string, string>;
let PUBLIC_CONFIG = "";
Object.entries(vars).forEach(([key, value]) => {
PUBLIC_CONFIG += `${key}=\`${value}\`\n`;
});
const SECRET_CONFIG =
(fs.existsSync(".env.SECRET_CONFIG")
? fs.readFileSync(".env.SECRET_CONFIG", "utf8")
: process.env.SECRET_CONFIG) ?? "";
// Prepend the content of the env variable SECRET_CONFIG
const full_config = `${PUBLIC_CONFIG}\n${SECRET_CONFIG}`;
// Write full_config to .env.local
fs.writeFileSync(".env.local", full_config);
| chat-ui/scripts/updateLocalEnv.ts/0 | {
"file_path": "chat-ui/scripts/updateLocalEnv.ts",
"repo_id": "chat-ui",
"token_count": 288
} |
<script lang="ts">
interface Props {
label?: string;
position?: "top" | "bottom" | "left" | "right";
TooltipClassNames?: string;
children?: import("svelte").Snippet;
}
let { label = "", position = "bottom", TooltipClassNames = "", children }: Props = $props();
const positionClasses = {
top: "bottom-full mb-2",
bottom: "top-full mt-2",
left: "right-full mr-2 top-1/2 -translate-y-1/2",
right: "left-full ml-2 top-1/2 -translate-y-1/2",
};
</script>
<div class="group/tooltip inline-block md:relative">
{@render children?.()}
<div
class="
invisible
absolute
z-10
w-64
whitespace-normal
rounded-md
bg-black
p-2
text-center
text-white
group-hover/tooltip:visible
group-active/tooltip:visible
max-sm:left-1/2
max-sm:-translate-x-1/2
{positionClasses[position]}
{TooltipClassNames}
"
>
{label}
</div>
</div>
| chat-ui/src/lib/components/HoverTooltip.svelte/0 | {
"file_path": "chat-ui/src/lib/components/HoverTooltip.svelte",
"repo_id": "chat-ui",
"token_count": 380
} |
<script lang="ts">
interface Props {
checked: boolean;
name: string;
}
let { checked = $bindable(), name }: Props = $props();
</script>
<input bind:checked type="checkbox" {name} class="peer pointer-events-none absolute opacity-0" />
<div
aria-checked={checked}
aria-roledescription="switch"
aria-label="switch"
role="switch"
tabindex="0"
class="relative inline-flex h-5 w-9 shrink-0 cursor-pointer items-center rounded-full bg-gray-300 p-1 shadow-inner ring-gray-400 transition-all peer-checked:bg-blue-600 peer-focus-visible:ring peer-focus-visible:ring-offset-1 hover:bg-gray-400 dark:bg-gray-600 peer-checked:[&>div]:translate-x-3.5"
>
<div class="h-3.5 w-3.5 rounded-full bg-white shadow-sm transition-all"></div>
</div>
| chat-ui/src/lib/components/Switch.svelte/0 | {
"file_path": "chat-ui/src/lib/components/Switch.svelte",
"repo_id": "chat-ui",
"token_count": 267
} |
<script lang="ts">
import { createBubbler } from "svelte/legacy";
const bubble = createBubbler();
import { useSettingsStore } from "$lib/stores/settings";
import { documentParserToolId } from "$lib/utils/toolIds";
import CarbonImage from "~icons/carbon/image";
interface Props {
// import EosIconsLoading from "~icons/eos-icons/loading";
files: File[];
mimeTypes?: string[];
onDrag?: boolean;
onDragInner?: boolean;
}
let {
files = $bindable(),
mimeTypes = [],
onDrag = $bindable(false),
onDragInner = $bindable(false),
}: Props = $props();
const settings = useSettingsStore();
async function dropHandle(event: DragEvent) {
event.preventDefault();
if (event.dataTransfer && event.dataTransfer.items) {
// Use DataTransferItemList interface to access the file(s)
if (files.length > 0) {
files = [];
}
if (event.dataTransfer.items[0].kind === "file") {
for (let i = 0; i < event.dataTransfer.items.length; i++) {
const file = event.dataTransfer.items[i].getAsFile();
if (file) {
// check if the file matches the mimeTypes
// else abort
if (
!mimeTypes.some((mimeType: string) => {
const [type, subtype] = mimeType.split("/");
const [fileType, fileSubtype] = file.type.split("/");
return (
(type === "*" || type === fileType) &&
(subtype === "*" || subtype === fileSubtype)
);
})
) {
setErrorMsg(
`Some file type not supported. Only allowed: ${mimeTypes.join(
", "
)}. Uploaded document is of type ${file.type}`
);
files = [];
return;
}
// if file is bigger than 10MB abort
if (file.size > 10 * 1024 * 1024) {
setErrorMsg("Some file is too big. (10MB max)");
files = [];
return;
}
// add the file to the files array
files = [...files, file];
settings.instantSet({
tools: [...($settings.tools ?? []), documentParserToolId],
});
}
}
onDrag = false;
}
}
}
function setErrorMsg(errorMsg: string) {
onDrag = false;
alert(errorMsg);
}
</script>
<div
id="dropzone"
role="form"
ondrop={dropHandle}
ondragenter={() => (onDragInner = true)}
ondragleave={() => (onDragInner = false)}
ondragover={(e) => {
e.preventDefault();
bubble("dragover");
}}
class="relative flex h-28 w-full max-w-4xl flex-col items-center justify-center gap-1 rounded-xl border-2 border-dotted {onDragInner
? 'border-blue-200 !bg-blue-500/10 text-blue-600 *:pointer-events-none dark:border-blue-600 dark:bg-blue-500/20 dark:text-blue-500'
: 'bg-gray-100 text-gray-500 dark:border-gray-500 dark:bg-gray-700 dark:text-gray-400'}"
>
<CarbonImage class="text-xl" />
<p>Drop File to add to chat</p>
</div>
| chat-ui/src/lib/components/chat/FileDropzone.svelte/0 | {
"file_path": "chat-ui/src/lib/components/chat/FileDropzone.svelte",
"repo_id": "chat-ui",
"token_count": 1191
} |
import type { Migration } from ".";
import { collections } from "$lib/server/database";
import { ObjectId, type WithId } from "mongodb";
import type { Conversation } from "$lib/types/Conversation";
import {
MessageUpdateType,
MessageWebSearchUpdateType,
type MessageUpdate,
} from "$lib/types/MessageUpdate";
import type { Message } from "$lib/types/Message";
import { logger } from "$lib/server/logger";
// -----------
/** Converts the old message update to the new schema */
function convertMessageUpdate(message: Message, update: MessageUpdate): MessageUpdate | null {
try {
// trim final websearch update, and sources update
if (update.type === "webSearch") {
if (update.subtype === MessageWebSearchUpdateType.Sources) {
return {
type: MessageUpdateType.WebSearch,
subtype: MessageWebSearchUpdateType.Sources,
message: update.message,
sources: update.sources.map(({ link, title }) => ({ link, title })),
};
} else if (update.subtype === MessageWebSearchUpdateType.Finished) {
return {
type: MessageUpdateType.WebSearch,
subtype: MessageWebSearchUpdateType.Finished,
};
}
}
return update;
} catch (error) {
logger.error(error, "Error converting message update during migration. Skipping it..");
return null;
}
}
const trimMessageUpdates: Migration = {
_id: new ObjectId("000000000000000000000006"),
name: "Trim message updates to reduce stored size",
up: async () => {
const allConversations = collections.conversations.find({});
let conversation: WithId<Pick<Conversation, "messages">> | null = null;
while ((conversation = await allConversations.tryNext())) {
const messages = conversation.messages.map((message) => {
// Convert all of the existing updates to the new schema
const updates = message.updates
?.map((update) => convertMessageUpdate(message, update))
.filter((update): update is MessageUpdate => Boolean(update));
return { ...message, updates };
});
// Set the new messages array
await collections.conversations.updateOne({ _id: conversation._id }, { $set: { messages } });
}
return true;
},
runEveryTime: false,
};
export default trimMessageUpdates;
| chat-ui/src/lib/migrations/routines/06-trim-message-updates.ts/0 | {
"file_path": "chat-ui/src/lib/migrations/routines/06-trim-message-updates.ts",
"repo_id": "chat-ui",
"token_count": 703
} |
import { z } from "zod";
import type { Endpoint } from "../endpoints";
import type { TextGenerationStreamOutput } from "@huggingface/inference";
import { createImageProcessorOptionsValidator } from "../images";
import { endpointMessagesToAnthropicMessages } from "./utils";
import type { MessageParam } from "@anthropic-ai/sdk/resources/messages.mjs";
export const endpointAnthropicVertexParametersSchema = z.object({
weight: z.number().int().positive().default(1),
model: z.any(),
type: z.literal("anthropic-vertex"),
region: z.string().default("us-central1"),
projectId: z.string(),
defaultHeaders: z.record(z.string()).optional(),
defaultQuery: z.record(z.string()).optional(),
multimodal: z
.object({
image: createImageProcessorOptionsValidator({
supportedMimeTypes: ["image/png", "image/jpeg", "image/webp"],
preferredMimeType: "image/webp",
// The 4 / 3 compensates for the 33% increase in size when converting to base64
maxSizeInMB: (5 / 4) * 3,
maxWidth: 4096,
maxHeight: 4096,
}),
})
.default({}),
});
export async function endpointAnthropicVertex(
input: z.input<typeof endpointAnthropicVertexParametersSchema>
): Promise<Endpoint> {
const { region, projectId, model, defaultHeaders, defaultQuery, multimodal } =
endpointAnthropicVertexParametersSchema.parse(input);
let AnthropicVertex;
try {
AnthropicVertex = (await import("@anthropic-ai/vertex-sdk")).AnthropicVertex;
} catch (e) {
throw new Error("Failed to import @anthropic-ai/vertex-sdk", { cause: e });
}
const anthropic = new AnthropicVertex({
baseURL: `https://${region}-aiplatform.googleapis.com/v1`,
region,
projectId,
defaultHeaders,
defaultQuery,
});
return async ({ messages, preprompt }) => {
let system = preprompt;
if (messages?.[0]?.from === "system") {
system = messages[0].content;
}
let tokenId = 0;
return (async function* () {
const stream = anthropic.messages.stream({
model: model.id ?? model.name,
messages: (await endpointMessagesToAnthropicMessages(
messages,
multimodal
)) as MessageParam[],
max_tokens: model.parameters?.max_new_tokens,
temperature: model.parameters?.temperature,
top_p: model.parameters?.top_p,
top_k: model.parameters?.top_k,
stop_sequences: model.parameters?.stop,
system,
});
while (true) {
const result = await Promise.race([stream.emitted("text"), stream.emitted("end")]);
// Stream end
if (result === undefined) {
yield {
token: {
id: tokenId++,
text: "",
logprob: 0,
special: true,
},
generated_text: await stream.finalText(),
details: null,
} satisfies TextGenerationStreamOutput;
return;
}
// Text delta
yield {
token: {
id: tokenId++,
text: result as unknown as string,
special: false,
logprob: 0,
},
generated_text: null,
details: null,
} satisfies TextGenerationStreamOutput;
}
})();
};
}
| chat-ui/src/lib/server/endpoints/anthropic/endpointAnthropicVertex.ts/0 | {
"file_path": "chat-ui/src/lib/server/endpoints/anthropic/endpointAnthropicVertex.ts",
"repo_id": "chat-ui",
"token_count": 1193
} |
import type { TextGenerationStreamOutput } from "@huggingface/inference";
import type OpenAI from "openai";
import type { Stream } from "openai/streaming";
/**
* Transform a stream of OpenAI.Completions.Completion into a stream of TextGenerationStreamOutput
*/
export async function* openAICompletionToTextGenerationStream(
completionStream: Stream<OpenAI.Completions.Completion>
) {
let generatedText = "";
let tokenId = 0;
for await (const completion of completionStream) {
const { choices } = completion;
const text = choices[0]?.text ?? "";
const last = choices[0]?.finish_reason === "stop" || choices[0]?.finish_reason === "length";
if (text) {
generatedText = generatedText + text;
}
const output: TextGenerationStreamOutput = {
token: {
id: tokenId++,
text,
logprob: 0,
special: last,
},
generated_text: last ? generatedText : null,
details: null,
};
yield output;
}
}
| chat-ui/src/lib/server/endpoints/openai/openAICompletionToTextGenerationStream.ts/0 | {
"file_path": "chat-ui/src/lib/server/endpoints/openai/openAICompletionToTextGenerationStream.ts",
"repo_id": "chat-ui",
"token_count": 325
} |
import { isURLLocal } from "./isURLLocal";
import { describe, expect, it } from "vitest";
describe("isURLLocal", async () => {
it("should return true for localhost", async () => {
expect(await isURLLocal(new URL("http://localhost"))).toBe(true);
});
it("should return true for 127.0.0.1", async () => {
expect(await isURLLocal(new URL("http://127.0.0.1"))).toBe(true);
});
it("should return true for 127.254.254.254", async () => {
expect(await isURLLocal(new URL("http://127.254.254.254"))).toBe(true);
});
it("should return false for huggingface.co", async () => {
expect(await isURLLocal(new URL("https://huggingface.co/"))).toBe(false);
});
it("should return true for 127.0.0.1.nip.io", async () => {
expect(await isURLLocal(new URL("http://127.0.0.1.nip.io"))).toBe(true);
});
it("should fail on ipv6", async () => {
await expect(isURLLocal(new URL("http://[::1]"))).rejects.toThrow();
});
it("should fail on ipv6 --1.sslip.io", async () => {
await expect(isURLLocal(new URL("http://--1.sslip.io"))).rejects.toThrow();
});
it("should fail on invalid domain names", async () => {
await expect(
isURLLocal(new URL("http://34329487239847329874923948732984.com/"))
).rejects.toThrow();
});
});
| chat-ui/src/lib/server/isURLLocal.spec.ts/0 | {
"file_path": "chat-ui/src/lib/server/isURLLocal.spec.ts",
"repo_id": "chat-ui",
"token_count": 492
} |
import { MessageUpdateType } from "$lib/types/MessageUpdate";
import {
ToolColor,
ToolIcon,
ToolOutputComponents,
type BackendCall,
type BaseTool,
type ConfigTool,
type ToolInput,
} from "$lib/types/Tool";
import type { TextGenerationContext } from "../textGeneration/types";
import { z } from "zod";
import JSON5 from "json5";
import { env } from "$env/dynamic/private";
import jp from "jsonpath";
import calculator from "./calculator";
import directlyAnswer from "./directlyAnswer";
import fetchUrl from "./web/url";
import websearch from "./web/search";
import { callSpace, getIpToken } from "./utils";
import { uploadFile } from "../files/uploadFile";
import type { MessageFile } from "$lib/types/Message";
import { sha256 } from "$lib/utils/sha256";
import { ObjectId } from "mongodb";
import { isValidOutputComponent, ToolOutputPaths } from "./outputs";
import { downloadFile } from "../files/downloadFile";
import { fileTypeFromBlob } from "file-type";
export type BackendToolContext = Pick<
TextGenerationContext,
"conv" | "messages" | "assistant" | "ip" | "username"
> & { preprompt?: string };
const IOType = z.union([z.literal("str"), z.literal("int"), z.literal("float"), z.literal("bool")]);
const toolInputBaseSchema = z.union([
z.object({
name: z.string().min(1).max(80),
description: z.string().max(200).optional(),
paramType: z.literal("required"),
}),
z.object({
name: z.string().min(1).max(80),
description: z.string().max(200).optional(),
paramType: z.literal("optional"),
default: z
.union([z.string().max(300), z.number(), z.boolean(), z.undefined()])
.transform((val) => (val === undefined ? "" : val)),
}),
z.object({
name: z.string().min(1).max(80),
paramType: z.literal("fixed"),
value: z
.union([z.string().max(300), z.number(), z.boolean(), z.undefined()])
.transform((val) => (val === undefined ? "" : val)),
}),
]);
const toolInputSchema = toolInputBaseSchema.and(
z.object({ type: IOType }).or(
z.object({
type: z.literal("file"),
mimeTypes: z.string().min(1),
})
)
);
export const editableToolSchema = z
.object({
name: z
.string()
.regex(/^[a-zA-Z_][a-zA-Z0-9_]*$/) // only allow letters, numbers, and underscores, and start with a letter or underscore
.min(1)
.max(40),
// only allow huggingface spaces either through namespace or direct URLs
baseUrl: z.union([
z.string().regex(/^[^/]+\/[^/]+$/),
z
.string()
.regex(/^https:\/\/huggingface\.co\/spaces\/[a-zA-Z0-9-]+\/[a-zA-Z0-9-]+$/)
.transform((url) => url.split("/").slice(-2).join("/")),
]),
endpoint: z.string().min(1).max(100),
inputs: z.array(toolInputSchema),
outputComponent: z.string().min(1).max(100),
showOutput: z.boolean(),
displayName: z.string().min(1).max(40),
color: ToolColor,
icon: ToolIcon,
description: z.string().min(1).max(100),
})
.transform((tool) => ({
...tool,
outputComponentIdx: parseInt(tool.outputComponent.split(";")[0]),
outputComponent: ToolOutputComponents.parse(tool.outputComponent.split(";")[1]),
}));
export const configTools = z
.array(
z
.object({
name: z.string(),
description: z.string(),
endpoint: z.union([z.string(), z.null()]),
inputs: z.array(toolInputSchema),
outputComponent: ToolOutputComponents.or(z.null()),
outputComponentIdx: z.number().int().default(0),
showOutput: z.boolean(),
_id: z
.string()
.length(24)
.regex(/^[0-9a-fA-F]{24}$/)
.transform((val) => new ObjectId(val)),
baseUrl: z.string().optional(),
displayName: z.string(),
color: ToolColor,
icon: ToolIcon,
isOnByDefault: z.optional(z.literal(true)),
isLocked: z.optional(z.literal(true)),
isHidden: z.optional(z.literal(true)),
})
.transform((val) => ({
type: "config" as const,
...val,
call: getCallMethod(val),
}))
)
// add the extra hardcoded tools
.transform((val) => [...val, calculator, directlyAnswer, fetchUrl, websearch]);
export function getCallMethod(tool: Omit<BaseTool, "call">): BackendCall {
return async function* (params, ctx, uuid) {
if (
tool.endpoint === null ||
!tool.baseUrl ||
!tool.outputComponent ||
tool.outputComponentIdx === null
) {
throw new Error(`Tool function ${tool.name} has no endpoint`);
}
const ipToken = await getIpToken(ctx.ip, ctx.username);
function coerceInput(value: unknown, type: ToolInput["type"]) {
const valueStr = String(value);
switch (type) {
case "str":
return valueStr;
case "int":
return parseInt(valueStr);
case "float":
return parseFloat(valueStr);
case "bool":
return valueStr === "true";
default:
throw new Error(`Unsupported type ${type}`);
}
}
const inputs = tool.inputs.map(async (input) => {
if (input.type === "file" && input.paramType !== "required") {
throw new Error("File inputs are always required and cannot be optional or fixed");
}
if (input.paramType === "fixed") {
return coerceInput(input.value, input.type);
} else if (input.paramType === "optional") {
return coerceInput(params[input.name] ?? input.default, input.type);
} else if (input.paramType === "required") {
if (params[input.name] === undefined) {
throw new Error(`Missing required input ${input.name}`);
}
if (input.type === "file") {
// todo: parse file here !
// structure is {input|output}-{msgIdx}-{fileIdx}-{filename}
const filename = params[input.name];
if (!filename || typeof filename !== "string") {
throw new Error(`Filename is not a string`);
}
const messages = ctx.messages;
const msgIdx = parseInt(filename.split("_")[1]);
const fileIdx = parseInt(filename.split("_")[2]);
if (Number.isNaN(msgIdx) || Number.isNaN(fileIdx)) {
throw Error(`Message index or file index is missing`);
}
if (msgIdx >= messages.length) {
throw Error(`Message index ${msgIdx} is out of bounds`);
}
const file = messages[msgIdx].files?.[fileIdx];
if (!file) {
throw Error(`File index ${fileIdx} is out of bounds`);
}
const blob = await downloadFile(file.value, ctx.conv._id)
.then((file) => fetch(`data:${file.mime};base64,${file.value}`))
.then((res) => res.blob())
.catch((err) => {
throw Error("Failed to download file", { cause: err });
});
return blob;
} else {
return coerceInput(params[input.name], input.type);
}
}
});
const outputs = yield* callSpace(
tool.baseUrl,
tool.endpoint,
await Promise.all(inputs),
ipToken,
uuid
);
if (!isValidOutputComponent(tool.outputComponent)) {
throw new Error(`Tool output component is not defined`);
}
const { type, path } = ToolOutputPaths[tool.outputComponent];
if (!path || !type) {
throw new Error(`Tool output type ${tool.outputComponent} is not supported`);
}
const files: MessageFile[] = [];
const toolOutputs: Array<Record<string, string>> = [];
if (outputs.length <= tool.outputComponentIdx) {
throw new Error(`Tool output component index is out of bounds`);
}
// if its not an object, return directly
if (
outputs[tool.outputComponentIdx] !== undefined &&
typeof outputs[tool.outputComponentIdx] !== "object"
) {
return {
outputs: [{ [tool.name + "-0"]: outputs[tool.outputComponentIdx] }],
display: tool.showOutput,
};
}
await Promise.all(
jp
.query(outputs[tool.outputComponentIdx], path)
.map(async (output: string | string[], idx) => {
const arrayedOutput = Array.isArray(output) ? output : [output];
if (type === "file") {
// output files are actually URLs
await Promise.all(
arrayedOutput.map(async (output, idx) => {
await fetch(output)
.then((res) => res.blob())
.then(async (blob) => {
const { ext, mime } = (await fileTypeFromBlob(blob)) ?? { ext: "octet-stream" };
return new File(
[blob],
`${idx}-${await sha256(JSON.stringify(params))}.${ext}`,
{
type: mime,
}
);
})
.then((file) => uploadFile(file, ctx.conv))
.then((file) => files.push(file));
})
);
toolOutputs.push({
[tool.name + "-" + idx.toString()]:
`Only and always answer: 'I used the tool ${tool.displayName}, here is the result.' Don't add anything else.`,
});
} else {
for (const output of arrayedOutput) {
toolOutputs.push({
[tool.name + "-" + idx.toString()]: output,
});
}
}
})
);
for (const file of files) {
yield {
type: MessageUpdateType.File,
name: file.name,
sha: file.value,
mime: file.mime,
};
}
return { outputs: toolOutputs, display: tool.showOutput };
};
}
export const toolFromConfigs = configTools.parse(JSON5.parse(env.TOOLS)) satisfies ConfigTool[];
| chat-ui/src/lib/server/tools/index.ts/0 | {
"file_path": "chat-ui/src/lib/server/tools/index.ts",
"repo_id": "chat-ui",
"token_count": 3696
} |
import type { SerializedHTMLElement } from "./types";
interface DBSCANOptions<T> {
dataset: T[];
epsilon?: number;
epsilonCompare?: (distance: number, epsilon: number) => boolean;
minimumPoints?: number;
distanceFunction: (a: T, b: T) => number;
}
export function spatialParser() {
/**
* Implementation for dbscan, inlined and migrated to typescript from https://github.com/cdxOo/dbscan (MIT License)
*/
const DBSCAN = <T>({
dataset,
epsilon = 1,
epsilonCompare = (dist, e) => dist < e,
minimumPoints = 2,
distanceFunction,
}: DBSCANOptions<T>) => {
const visitedIndices: Record<number, boolean> = {};
const isVisited = (i: number) => visitedIndices[i];
const markVisited = (i: number) => {
visitedIndices[i] = true;
};
const clusteredIndices: Record<number, boolean> = {};
const isClustered = (i: number) => clusteredIndices[i];
const markClustered = (i: number) => {
clusteredIndices[i] = true;
};
const uniqueMerge = <U>(targetArray: U[], sourceArray: U[]) => {
for (let i = 0; i < sourceArray.length; i += 1) {
const item = sourceArray[i];
if (targetArray.indexOf(item) < 0) {
targetArray.push(item);
}
}
};
const findNeighbors = (index: number) => {
const neighbors = [];
for (let other = 0; other < dataset.length; other += 1) {
const distance = distanceFunction(dataset[index], dataset[other]);
if (epsilonCompare(distance, epsilon)) {
neighbors.push(other);
}
}
return neighbors;
};
const noise: number[] = [];
const addNoise = (i: number) => noise.push(i);
const clusters: number[][] = [];
const createCluster = () => clusters.push([]) - 1;
const addIndexToCluster = (c: number, i: number) => {
clusters[c].push(i);
markClustered(i);
};
const expandCluster = (c: number, neighbors: number[]) => {
for (let i = 0; i < neighbors.length; i += 1) {
const neighborIndex = neighbors[i];
if (!isVisited(neighborIndex)) {
markVisited(neighborIndex);
const secondaryNeighbors = findNeighbors(neighborIndex);
if (secondaryNeighbors.length >= minimumPoints) {
uniqueMerge(neighbors, secondaryNeighbors);
}
}
if (!isClustered(neighborIndex)) {
addIndexToCluster(c, neighborIndex);
}
}
};
dataset.forEach((_, index) => {
if (!isVisited(index)) {
markVisited(index);
const neighbors = findNeighbors(index);
if (neighbors.length < minimumPoints) {
addNoise(index);
} else {
const clusterIndex = createCluster();
addIndexToCluster(clusterIndex, index);
expandCluster(clusterIndex, neighbors);
}
}
});
return { clusters, noise };
};
// -----------
// Scraping implementation
const IgnoredTagsList = [
"footer",
"nav",
"aside",
"script",
"style",
"noscript",
"form",
"button",
];
const InlineTags = [
"a",
"abbrv",
"span",
"address",
"time",
"acronym",
"strong",
"b",
"br",
"sub",
"sup",
"tt",
"var",
"em",
"i",
];
type ReadableNode = HTMLElement;
type NodeWithRect = {
node: ReadableNode;
rect: DOMRect;
};
const isOnlyChild = (node: Node) => {
if (!node.parentElement) return true;
if (node.parentElement.nodeName === "body") return false;
if (node.parentElement.childNodes.length === 1) return true;
return false;
};
const hasValidInlineParent = (node: Node) => {
return node.parentElement && !node.parentElement.matches("div, section, article, main, body ");
};
const hasValidParent = (node: Node) => {
return node.parentElement && !node.parentElement.isSameNode(document.body);
};
const possibleCodeParents = Array.from(document.querySelectorAll("pre, p"));
const possibleTableParents = Array.from(document.querySelectorAll("table"));
const possibleListParents = Array.from(document.querySelectorAll("ul, ol"));
/**
* We want to find the highest parent of text node in the cluster.
* For example in this case: <p><span>Text here</span></p>
* the P tag is highest parent.
*/
const findHighestDirectParentOfReadableNode = (node: Node): HTMLElement => {
// go up the tree until the parent is no longer an only child
let parent = node.parentElement;
// if the parent is an inline tag, then go up one more level
while (
parent &&
hasValidInlineParent(parent) &&
InlineTags.includes(parent?.tagName.toLowerCase())
) {
parent = parent.parentElement;
}
while (parent && isOnlyChild(parent)) {
if (!hasValidParent(parent)) break;
parent = parent.parentElement;
}
if (!parent) {
throw new Error(
"disconnected node found, this should not really be possible when traversing through the dom"
);
}
// if the parent is a span, code or div tag check if there is a pre tag or p tag above it
if (["span", "code", "div"].includes(parent.nodeName.toLowerCase())) {
const hasParent = possibleCodeParents.find((tag) => tag.contains(parent)) as HTMLElement;
if (hasParent) {
parent = hasParent;
}
}
// if the parent is a li tag check if there is a ul or ol tag above it
if (parent.nodeName.toLowerCase() === "li") {
const hasParent = possibleListParents.find((tag) => tag.contains(parent)) as HTMLElement;
if (hasParent) {
parent = hasParent;
}
}
// if the parent is a td, th, tr tag check if there is a table tag above it
if (["td", "th", "tr"].includes(parent.nodeName.toLowerCase())) {
const hasParent = possibleTableParents.find((tag) => tag.contains(parent)) as HTMLElement;
if (hasParent) {
parent = hasParent;
}
}
return parent;
};
const barredNodes = Array.from(document.querySelectorAll(IgnoredTagsList.join(",")));
const doesNodePassHeuristics = (node: Node) => {
if ((node.textContent ?? "").trim().length < 10) {
return false;
}
const parentNode = findHighestDirectParentOfReadableNode(node);
if (parentNode && parentNode instanceof Element) {
if (
!parentNode.checkVisibility({
checkOpacity: true,
checkVisibilityCSS: true,
})
)
return false;
const rect = parentNode.getBoundingClientRect();
// elements that are readable usually don't have really small height or width
if (rect.width < 4 || rect.height < 4) {
return false;
}
}
if (parentNode && parentNode instanceof Element) {
if (barredNodes.some((barredNode) => barredNode.contains(parentNode))) {
return false;
}
}
return true;
};
const getAllReadableNodes = (): NodeWithRect[] => {
if (!document.body) throw new Error("Page failed to load");
const treeWalker = document.createTreeWalker(document.body, NodeFilter.SHOW_TEXT, {
acceptNode(node) {
if (doesNodePassHeuristics(node)) {
return NodeFilter.FILTER_ACCEPT;
} else {
return NodeFilter.FILTER_SKIP;
}
},
});
const readableNodes = [];
while (treeWalker.nextNode()) {
readableNodes.push(treeWalker.currentNode as ReadableNode);
}
/*
* <table><p>hello</p><p>world</p></table>
* table is already included in the parent of the first p tag
*/
const parentsForReadableNodes = readableNodes.map(findHighestDirectParentOfReadableNode);
const listWithOnlyParents: HTMLElement[] = [];
// find unique nodes in the parent list, a unique node is a node that is not a child of any other node in the list
for (let i = 0; i < parentsForReadableNodes.length; i++) {
const node = parentsForReadableNodes[i];
const hasParentInList = parentsForReadableNodes.find((otherNode, idx) => {
if (i === idx) return false;
return otherNode.contains(node);
});
listWithOnlyParents.push(hasParentInList ? hasParentInList : node);
}
const uniqueParents = Array.from(new Set(listWithOnlyParents));
return uniqueParents.map((node) => {
return {
node,
rect: node.getBoundingClientRect(),
};
});
};
const distanceFunction = (a: NodeWithRect, b: NodeWithRect) => {
// we make two assumptions here which are fine to make for rects returned from getBoundingClientRect
// 1. rects are upright and not rotated
// 2. If two rects intersect, we assume distance to be 0
let dx = 0;
let dy = 0;
const rect1 = a.rect;
const rect2 = b.rect;
// Calculate the horizontal distance
if (rect1.x + rect1.width < rect2.x) {
dx = rect2.x - (rect1.x + rect1.width);
} else if (rect2.x + rect2.width < rect1.x) {
dx = rect1.x - (rect2.x + rect2.width);
}
// Calculate the vertical distance
if (rect1.y + rect1.height < rect2.y) {
dy = rect2.y - (rect1.y + rect1.height);
} else if (rect2.y + rect2.height < rect1.y) {
dy = rect1.y - (rect2.y + rect2.height);
}
const distance = Math.sqrt(dx * dx + dy * dy);
// Return the Euclidean distance
return distance;
};
/**
* Clusters nodes using dbscan
*/
const clusterReadableNodes = (nodes: NodeWithRect[]) => {
const { clusters } = DBSCAN({
dataset: nodes,
epsilon: 28,
minimumPoints: 1,
distanceFunction,
});
return clusters;
};
const totalTextLength = (cluster: number[]) => {
return cluster
.map((t) => readableNodes[t].node.innerText?.replaceAll(/ {2}|\r\n|\n|\r/gm, ""))
.join("").length;
};
const approximatelyEqual = (a: number, b: number, epsilon = 1) => {
return Math.abs(a - b) < epsilon;
};
const getClusterBounds = (cluster: number[]) => {
const leftMostPoint = Math.min(...cluster.map((c) => readableNodes[c].rect.x));
const topMostPoint = Math.min(...cluster.map((c) => readableNodes[c].rect.y));
const rightMostPoint = Math.max(
...cluster.map((c) => readableNodes[c].rect.x + readableNodes[c].rect.width)
);
const bottomMostPoint = Math.max(
...cluster.map((c) => readableNodes[c].rect.y + readableNodes[c].rect.height)
);
return {
// left most element
x: leftMostPoint,
y: topMostPoint,
width: rightMostPoint - leftMostPoint,
height: bottomMostPoint - topMostPoint,
};
};
const round = (num: number, decimalPlaces = 2) => {
const factor = Math.pow(10, decimalPlaces);
return Math.round(num * factor) / factor;
};
/** minimum distance to center of the screen */
const clusterCentrality = (cluster: number[]) => {
const bounds = getClusterBounds(cluster);
const centerOfScreen = window.innerWidth / 2;
// the cluster contains the center of the screen
if (bounds.x < centerOfScreen && bounds.x + bounds.width > centerOfScreen) {
return 0;
}
// the cluster is to the left of the screen
if (bounds.x + bounds.width < centerOfScreen) {
return centerOfScreen - (bounds.x + bounds.width);
}
// the cluster is to the right of the screen
return bounds.x - centerOfScreen;
};
/** measure of text share that belong to the cluster */
const percentageTextShare = (cluster: number[], totalLength: number) => {
// apply an exponentially increasing penalty for centrality per 100 pixels distance from center
return round((totalTextLength(cluster) / totalLength) * 100);
};
const shouldMergeClusters = (clusterA: number[], clusterB: number[]) => {
const clusterABounds = getClusterBounds(clusterA);
const clusterBBounds = getClusterBounds(clusterB);
// A cluster is horizontally aligned if the x and width are roughly equal
const isHorizontallyAligned =
approximatelyEqual(clusterABounds.x, clusterBBounds.x, 40) &&
approximatelyEqual(clusterABounds.width, clusterBBounds.width, 40);
if (!isHorizontallyAligned) return false;
// check the y gap between the clusters
const higherCluster = clusterABounds.y < clusterBBounds.y ? clusterABounds : clusterBBounds;
const lowerCluster = clusterABounds.y < clusterBBounds.y ? clusterBBounds : clusterABounds;
const yGap = lowerCluster.y - (higherCluster.y + higherCluster.height);
if (approximatelyEqual(yGap, 0, 100)) return true;
};
const findCriticalClusters = (clusters: number[][]) => {
// merge the clusters that have similar widths and x position
let i = 0;
while (i < clusters.length) {
const cluster = clusters[i];
for (let j = i + 1; j < clusters.length; j++) {
const otherCluster = clusters[j];
if (shouldMergeClusters(cluster, otherCluster)) {
cluster.push(...otherCluster);
clusters.splice(j, 1);
j -= 1;
}
}
i++;
}
const totalText = totalTextLength(clusters.flat());
// sort in descending order of text share
const clusterWithMetrics = clusters.map((cluster) => {
const centrality = clusterCentrality(cluster);
return {
cluster,
centrality,
percentageTextShare: percentageTextShare(cluster, totalText),
};
});
// if there is a dominant cluster with more than 60% text share, return that
const dominantCluster = clusterWithMetrics[0]?.percentageTextShare > 60;
if (dominantCluster) return [clusterWithMetrics[0].cluster];
// clusters are sorted by text share after applying a penalty for centrality
const sortedClusters = clusterWithMetrics.sort((a, b) => {
const penaltyForA = Math.pow(0.9, a.centrality / 100);
const penaltyForB = Math.pow(0.9, b.centrality / 100);
const adjustedTextShareA = a.percentageTextShare * penaltyForA;
const adjustedTextShareB = b.percentageTextShare * penaltyForB;
return adjustedTextShareB - adjustedTextShareA;
});
// find all clusters that are similar to the largest cluster in terms of text share
// and see if they are enough to cover at least 60% of the text share
const largeTextShareClusters = sortedClusters.filter((c) =>
approximatelyEqual(c.percentageTextShare, sortedClusters[0]?.percentageTextShare, 10)
);
const totalTextShareOfLargeClusters = largeTextShareClusters.reduce(
(acc, cluster) => acc + cluster.percentageTextShare,
0
);
if (totalTextShareOfLargeClusters > 60) {
return largeTextShareClusters.map((c) => c.cluster);
}
// choose clusters till the text share is greater than 60%
let totalTextShare = 0;
const criticalClusters = [];
for (const cluster of sortedClusters) {
/** Ignore clusters with less than 2%*/
if (cluster.percentageTextShare < 2) continue;
if (totalTextShare > 60) break;
criticalClusters.push(cluster.cluster);
totalTextShare += cluster.percentageTextShare;
}
// if the total text share is less than 60% then return an empty array
// as this website should not be particularly useful for the web search anyways
// this should almost never happen on structured website with a lot of text
if (totalTextShare < 60) {
return [];
}
return criticalClusters;
};
const allowListedAttributes = ["href", "src", "alt", "title", "class", "id"];
function serializeHTMLElement(node: Element): SerializedHTMLElement {
return {
tagName: node.tagName.toLowerCase(),
attributes: allowListedAttributes.reduce(
(acc, attr) => {
const value = node.getAttribute(attr);
if (value) {
acc[attr] = value;
}
return acc;
},
{} as Record<string, string>
),
content: Array.from(node.childNodes).map(serializeNode).filter(Boolean),
};
}
function serializeNode(node: Node): SerializedHTMLElement | string {
if (node.nodeType === 1) return serializeHTMLElement(node as Element);
else if (node.nodeType === 3) return node.textContent ?? "";
else return "";
}
function getPageMetadata(): {
title: string;
siteName?: string;
author?: string;
description?: string;
createdAt?: string;
updatedAt?: string;
} {
const title = document.title ?? "";
const siteName =
document.querySelector("meta[property='og:site_name']")?.getAttribute("content") ?? undefined;
const author =
document.querySelector("meta[name='author']")?.getAttribute("content") ?? undefined;
const description =
document.querySelector("meta[name='description']")?.getAttribute("content") ??
document.querySelector("meta[property='og:description']")?.getAttribute("content") ??
undefined;
const createdAt =
document.querySelector("meta[property='article:published_time']")?.getAttribute("content") ??
document.querySelector("meta[name='date']")?.getAttribute("content") ??
undefined;
const updatedAt =
document.querySelector("meta[property='article:modified_time']")?.getAttribute("content") ??
undefined;
return { title, siteName, author, description, createdAt, updatedAt };
}
const readableNodes = getAllReadableNodes();
const clusters = clusterReadableNodes(readableNodes);
const criticalClusters = findCriticalClusters(clusters);
// filter readable nodes using the above information as well as heuristics
const filteredNodes = readableNodes.filter((_, idx) => {
return criticalClusters.some((cluster) => {
return cluster.includes(idx);
});
});
const elements = filteredNodes
.filter(
(node, idx, nodes) => !nodes.slice(idx + 1).some((otherNode) => node.node === otherNode.node)
)
.map<SerializedHTMLElement>(({ node }) => serializeHTMLElement(node));
const metadata = getPageMetadata();
return { ...metadata, elements };
}
| chat-ui/src/lib/server/websearch/scrape/parser.ts/0 | {
"file_path": "chat-ui/src/lib/server/websearch/scrape/parser.ts",
"repo_id": "chat-ui",
"token_count": 6106
} |
import { base } from "$app/paths";
import { ERROR_MESSAGES, error } from "$lib/stores/errors";
import { share } from "./utils/share";
import { page } from "$app/stores";
import { get } from "svelte/store";
import { getShareUrl } from "./utils/getShareUrl";
export async function shareConversation(id: string, title: string) {
try {
if (id.length === 7) {
const url = get(page).url;
await share(getShareUrl(url, id), title, true);
} else {
const res = await fetch(`${base}/conversation/${id}/share`, {
method: "POST",
headers: {
"Content-Type": "application/json",
},
});
if (!res.ok) {
error.set("Error while sharing conversation, try again.");
console.error("Error while sharing conversation: " + (await res.text()));
return;
}
const { url } = await res.json();
await share(url, title, true);
}
} catch (err) {
error.set(ERROR_MESSAGES.default);
console.error(err);
}
}
| chat-ui/src/lib/shareConversation.ts/0 | {
"file_path": "chat-ui/src/lib/shareConversation.ts",
"repo_id": "chat-ui",
"token_count": 363
} |
import type { ObjectId } from "mongodb";
import type { Conversation } from "./Conversation";
import type { Timestamps } from "./Timestamps";
import type { HeaderElement } from "$lib/server/websearch/markdown/types";
export interface WebSearch extends Timestamps {
_id?: ObjectId;
convId?: Conversation["_id"];
prompt: string;
searchQuery: string;
results: WebSearchSource[];
contextSources: WebSearchUsedSource[];
}
export interface WebSearchSource {
title?: string;
link: string;
}
export interface WebSearchScrapedSource extends WebSearchSource {
page: WebSearchPage;
}
export interface WebSearchPage {
title: string;
siteName?: string;
author?: string;
description?: string;
createdAt?: string;
modifiedAt?: string;
markdownTree: HeaderElement;
}
export interface WebSearchUsedSource extends WebSearchScrapedSource {
context: string;
}
export type WebSearchMessageSources = {
type: "sources";
sources: WebSearchSource[];
};
// eslint-disable-next-line no-shadow
export enum WebSearchProvider {
GOOGLE = "Google",
YOU = "You.com",
SEARXNG = "SearXNG",
BING = "Bing",
}
| chat-ui/src/lib/types/WebSearch.ts/0 | {
"file_path": "chat-ui/src/lib/types/WebSearch.ts",
"repo_id": "chat-ui",
"token_count": 348
} |
type Gen<T, TReturn> = AsyncGenerator<T, TReturn, undefined>;
type GenPromiseMap<T, TReturn> = Map<
Gen<T, TReturn>,
Promise<{ gen: Gen<T, TReturn> } & IteratorResult<T, TReturn>>
>;
/** Merges multiple async generators into a single async generator that yields values from all of them in parallel. */
export async function* mergeAsyncGenerators<T, TReturn>(
generators: Gen<T, TReturn>[]
): Gen<T, TReturn[]> {
const promises: GenPromiseMap<T, TReturn> = new Map();
const results: Map<Gen<T, TReturn>, TReturn> = new Map();
for (const gen of generators) {
promises.set(
gen,
gen.next().then((result) => ({ gen, ...result }))
);
}
while (promises.size) {
const { gen, value, done } = await Promise.race(promises.values());
if (done) {
results.set(gen, value as TReturn);
promises.delete(gen);
} else {
promises.set(
gen,
gen.next().then((result) => ({ gen, ...result }))
);
yield value as T;
}
}
const orderedResults = generators.map((gen) => results.get(gen) as TReturn);
return orderedResults;
}
| chat-ui/src/lib/utils/mergeAsyncGenerators.ts/0 | {
"file_path": "chat-ui/src/lib/utils/mergeAsyncGenerators.ts",
"repo_id": "chat-ui",
"token_count": 407
} |
import type { Conversation } from "$lib/types/Conversation";
import type { Message } from "$lib/types/Message";
import { v4 } from "uuid";
export function addChildren(
conv: Pick<Conversation, "messages" | "rootMessageId">,
message: Omit<Message, "id">,
parentId?: Message["id"]
): Message["id"] {
// if this is the first message we just push it
if (conv.messages.length === 0) {
const messageId = v4();
conv.rootMessageId = messageId;
conv.messages.push({
...message,
ancestors: [],
id: messageId,
});
return messageId;
}
if (!parentId) {
throw new Error("You need to specify a parentId if this is not the first message");
}
const messageId = v4();
if (!conv.rootMessageId) {
// if there is no parentId we just push the message
if (!!parentId && parentId !== conv.messages[conv.messages.length - 1].id) {
throw new Error("This is a legacy conversation, you can only append to the last message");
}
conv.messages.push({ ...message, id: messageId });
return messageId;
}
const ancestors = [...(conv.messages.find((m) => m.id === parentId)?.ancestors ?? []), parentId];
conv.messages.push({
...message,
ancestors,
id: messageId,
children: [],
});
const parent = conv.messages.find((m) => m.id === parentId);
if (parent) {
if (parent.children) {
parent.children.push(messageId);
} else parent.children = [messageId];
}
return messageId;
}
| chat-ui/src/lib/utils/tree/addChildren.ts/0 | {
"file_path": "chat-ui/src/lib/utils/tree/addChildren.ts",
"repo_id": "chat-ui",
"token_count": 501
} |
import { collections } from "$lib/server/database";
import { ObjectId } from "mongodb";
export async function GET({ params }) {
const id = params.id;
const assistantId = new ObjectId(id);
const assistant = await collections.assistants.findOne({
_id: assistantId,
});
if (assistant) {
return Response.json(assistant);
} else {
return Response.json({ message: "Assistant not found" }, { status: 404 });
}
}
| chat-ui/src/routes/api/assistant/[id]/+server.ts/0 | {
"file_path": "chat-ui/src/routes/api/assistant/[id]/+server.ts",
"repo_id": "chat-ui",
"token_count": 133
} |
import type { RequestHandler } from "./$types";
import { collections } from "$lib/server/database";
import { ObjectId } from "mongodb";
import { error, redirect } from "@sveltejs/kit";
import { base } from "$app/paths";
import { z } from "zod";
import type { Message } from "$lib/types/Message";
import { models, validateModel } from "$lib/server/models";
import { defaultEmbeddingModel } from "$lib/server/embeddingModels";
import { v4 } from "uuid";
import { authCondition } from "$lib/server/auth";
import { usageLimits } from "$lib/server/usageLimits";
import { MetricsServer } from "$lib/server/metrics";
export const POST: RequestHandler = async ({ locals, request }) => {
const body = await request.text();
let title = "";
const parsedBody = z
.object({
fromShare: z.string().optional(),
model: validateModel(models),
assistantId: z.string().optional(),
preprompt: z.string().optional(),
})
.safeParse(JSON.parse(body));
if (!parsedBody.success) {
error(400, "Invalid request");
}
const values = parsedBody.data;
const convCount = await collections.conversations.countDocuments(authCondition(locals));
if (usageLimits?.conversations && convCount > usageLimits?.conversations) {
error(429, "You have reached the maximum number of conversations. Delete some to continue.");
}
const model = models.find((m) => (m.id || m.name) === values.model);
if (!model) {
error(400, "Invalid model");
}
let messages: Message[] = [
{
id: v4(),
from: "system",
content: values.preprompt ?? "",
createdAt: new Date(),
updatedAt: new Date(),
children: [],
ancestors: [],
},
];
let rootMessageId: Message["id"] = messages[0].id;
let embeddingModel: string;
if (values.fromShare) {
const conversation = await collections.sharedConversations.findOne({
_id: values.fromShare,
});
if (!conversation) {
error(404, "Conversation not found");
}
title = conversation.title;
messages = conversation.messages;
rootMessageId = conversation.rootMessageId ?? rootMessageId;
values.model = conversation.model;
values.preprompt = conversation.preprompt;
values.assistantId = conversation.assistantId?.toString();
embeddingModel = conversation.embeddingModel;
}
embeddingModel ??= model.embeddingModel ?? defaultEmbeddingModel.name;
if (model.unlisted) {
error(400, "Can't start a conversation with an unlisted model");
}
// get preprompt from assistant if it exists
const assistant = await collections.assistants.findOne({
_id: new ObjectId(values.assistantId),
});
if (assistant) {
values.preprompt = assistant.preprompt;
} else {
values.preprompt ??= model?.preprompt ?? "";
}
if (messages && messages.length > 0 && messages[0].from === "system") {
messages[0].content = values.preprompt;
}
const res = await collections.conversations.insertOne({
_id: new ObjectId(),
title: title || "New Chat",
rootMessageId,
messages,
model: values.model,
preprompt: values.preprompt,
assistantId: values.assistantId ? new ObjectId(values.assistantId) : undefined,
createdAt: new Date(),
updatedAt: new Date(),
userAgent: request.headers.get("User-Agent") ?? undefined,
embeddingModel,
...(locals.user ? { userId: locals.user._id } : { sessionId: locals.sessionId }),
...(values.fromShare ? { meta: { fromShareId: values.fromShare } } : {}),
});
MetricsServer.getMetrics().model.conversationsTotal.inc({ model: values.model });
return new Response(
JSON.stringify({
conversationId: res.insertedId.toString(),
}),
{ headers: { "Content-Type": "application/json" } }
);
};
export const GET: RequestHandler = async () => {
redirect(302, `${base}/`);
};
| chat-ui/src/routes/conversation/+server.ts/0 | {
"file_path": "chat-ui/src/routes/conversation/+server.ts",
"repo_id": "chat-ui",
"token_count": 1252
} |
<script lang="ts">
import type { PageData } from "./$types";
import { env as envPublic } from "$env/dynamic/public";
import { isHuggingChat } from "$lib/utils/isHuggingChat";
import { base } from "$app/paths";
import { page } from "$app/state";
import CarbonHelpFilled from "~icons/carbon/help-filled";
import CarbonTools from "~icons/carbon/tools";
import CarbonImage from "~icons/carbon/image";
import { useSettingsStore } from "$lib/stores/settings";
interface Props {
data: PageData;
}
let { data }: Props = $props();
const settings = useSettingsStore();
</script>
<svelte:head>
{#if isHuggingChat}
<title>HuggingChat - Models</title>
<meta property="og:title" content="HuggingChat - Models" />
<meta property="og:type" content="link" />
<meta property="og:description" content="Browse HuggingChat available models" />
<meta property="og:url" content={page.url.href} />
{/if}
</svelte:head>
<div class="scrollbar-custom h-full overflow-y-auto py-12 max-sm:pt-8 md:py-24">
<div class="pt-42 mx-auto flex flex-col px-5 xl:w-[60rem] 2xl:w-[64rem]">
<div class="flex items-center">
<h1 class="text-2xl font-bold">Models</h1>
{#if isHuggingChat}
<a
href="https://huggingface.co/spaces/huggingchat/chat-ui/discussions/372"
class="ml-auto dark:text-gray-400 dark:hover:text-gray-300"
target="_blank"
aria-label="Hub discussion about models"
>
<CarbonHelpFilled />
</a>
{/if}
</div>
<h2 class="text-gray-500">All models available on {envPublic.PUBLIC_APP_NAME}</h2>
<div class="mt-8 grid grid-cols-1 gap-3 sm:gap-5 xl:grid-cols-2">
{#each data.models.filter((el) => !el.unlisted) as model, index (model.id)}
<div
aria-label="Model card"
role="region"
class="relative flex flex-col gap-2 overflow-hidden rounded-xl border bg-gray-50/50 px-6 py-5 shadow hover:bg-gray-50 hover:shadow-inner dark:border-gray-800/70 dark:bg-gray-950/20 dark:hover:bg-gray-950/40"
class:active-model={model.id === $settings.activeModel}
>
<a
href="{base}/models/{model.id}"
class="absolute inset-0 z-10"
aria-label="View details for {model.displayName}"
></a>
<div class="flex items-center justify-between gap-1">
{#if model.logoUrl}
<img
class="overflown aspect-square size-6 rounded border dark:border-gray-700"
src={model.logoUrl}
alt="{model.displayName} logo"
/>
{:else}
<div
class="size-6 rounded border border-transparent bg-gray-300 dark:bg-gray-800"
aria-hidden="true"
></div>
{/if}
{#if model.tools}
<span
title="This model supports tools."
class="ml-auto grid size-[21px] place-items-center rounded-lg border border-purple-300 dark:border-purple-700"
aria-label="Model supports tools"
role="img"
>
<CarbonTools class="text-xxs text-purple-700 dark:text-purple-500" />
</span>
{/if}
{#if model.multimodal}
<span
title="This model is multimodal and supports image inputs natively."
class="ml-auto flex size-[21px] items-center justify-center rounded-lg border border-blue-700 dark:border-blue-500"
aria-label="Model is multimodal"
role="img"
>
<CarbonImage class="text-xxs text-blue-700 dark:text-blue-500" />
</span>
{/if}
{#if model.reasoning}
<span
title="This model supports reasoning."
class="ml-auto grid size-[21px] place-items-center rounded-lg border border-purple-300 dark:border-purple-700"
aria-label="Model supports reasoning"
role="img"
>
<svg xmlns="http://www.w3.org/2000/svg" width="14" height="14" viewBox="0 0 32 32">
<path
class="stroke-purple-700"
style="stroke-width: 2; fill: none; stroke-linecap: round; stroke-linejoin: round; stroke-dasharray: 50;"
d="M16 6v3.33M16 6c0-2.65 3.25-4.3 5.4-2.62 1.2.95 1.6 2.65.95 4.04a3.63 3.63 0 0 1 4.61.16 3.45 3.45 0 0 1 .46 4.37 5.32 5.32 0 0 1 1.87 4.75c-.22 1.66-1.39 3.6-3.07 4.14M16 6c0-2.65-3.25-4.3-5.4-2.62a3.37 3.37 0 0 0-.95 4.04 3.65 3.65 0 0 0-4.6.16 3.37 3.37 0 0 0-.49 4.27 5.57 5.57 0 0 0-1.85 4.85 5.3 5.3 0 0 0 3.07 4.15M16 9.33v17.34m0-17.34c0 2.18 1.82 4 4 4m6.22 7.5c.67 1.3.56 2.91-.27 4.11a4.05 4.05 0 0 1-4.62 1.5c0 1.53-1.05 2.9-2.66 2.9A2.7 2.7 0 0 1 16 26.66m10.22-5.83a4.05 4.05 0 0 0-3.55-2.17m-16.9 2.18a4.05 4.05 0 0 0 .28 4.1c1 1.44 2.92 2.09 4.59 1.5 0 1.52 1.12 2.88 2.7 2.88A2.7 2.7 0 0 0 16 26.67M5.78 20.85a4.04 4.04 0 0 1 3.55-2.18"
/>
</svg>
</span>
{/if}
{#if model.id === $settings.activeModel}
<span
class="rounded-full border border-blue-500 bg-blue-500/5 px-2 py-0.5 text-xs text-blue-500 dark:border-blue-500 dark:bg-blue-500/10"
>
Active
</span>
{:else if index === 0}
<span
class="rounded-full border border-gray-300 px-2 py-0.5 text-xs text-gray-500 dark:border-gray-500 dark:text-gray-400"
>
Default
</span>
{/if}
</div>
<span class="flex items-center gap-2 font-semibold">
{model.displayName}
</span>
<span class="whitespace-pre-wrap text-sm text-gray-500 dark:text-gray-400">
{model.description || "-"}
</span>
</div>
{/each}
</div>
</div>
</div>
| chat-ui/src/routes/models/+page.svelte/0 | {
"file_path": "chat-ui/src/routes/models/+page.svelte",
"repo_id": "chat-ui",
"token_count": 2588
} |
import { collections } from "$lib/server/database";
import { error, type RequestHandler } from "@sveltejs/kit";
import { ObjectId } from "mongodb";
export const GET: RequestHandler = async ({ params }) => {
const assistant = await collections.assistants.findOne({
_id: new ObjectId(params.assistantId),
});
if (!assistant) {
error(404, "No assistant found");
}
if (!assistant.avatar) {
error(404, "No avatar found");
}
const fileId = collections.bucket.find({ filename: assistant._id.toString() });
const content = await fileId.next().then(async (file) => {
if (!file?._id) {
error(404, "Avatar not found");
}
const fileStream = collections.bucket.openDownloadStream(file?._id);
const fileBuffer = await new Promise<Buffer>((resolve, reject) => {
const chunks: Uint8Array[] = [];
fileStream.on("data", (chunk) => chunks.push(chunk));
fileStream.on("error", reject);
fileStream.on("end", () => resolve(Buffer.concat(chunks)));
});
return fileBuffer;
});
return new Response(content, {
headers: {
"Content-Type": "image/jpeg",
"Content-Security-Policy":
"default-src 'none'; script-src 'none'; style-src 'none'; sandbox;",
},
});
};
| chat-ui/src/routes/settings/(nav)/assistants/[assistantId]/avatar.jpg/+server.ts/0 | {
"file_path": "chat-ui/src/routes/settings/(nav)/assistants/[assistantId]/avatar.jpg/+server.ts",
"repo_id": "chat-ui",
"token_count": 420
} |
import { base } from "$app/paths";
import { requiresUser } from "$lib/server/auth.js";
import { collections } from "$lib/server/database.js";
import { editableToolSchema } from "$lib/server/tools/index.js";
import { generateSearchTokens } from "$lib/utils/searchTokens.js";
import { error, fail, redirect } from "@sveltejs/kit";
import { ObjectId } from "mongodb";
export const actions = {
default: async ({ request, params, locals }) => {
const tool = await collections.tools.findOne({
_id: new ObjectId(params.toolId),
});
if (!tool) {
throw Error("Tool not found");
}
if (tool.createdById.toString() !== (locals.user?._id ?? locals.sessionId).toString()) {
throw Error("You are not the creator of this tool");
}
// can only create tools when logged in, IF login is setup
if (!locals.user && requiresUser) {
const errors = [{ field: "description", message: "Must be logged in. Unauthorized" }];
return fail(400, { error: true, errors });
}
const body = await request.formData();
const toolStringified = body.get("tool");
if (!toolStringified || typeof toolStringified !== "string") {
error(400, "Tool is required");
}
const parse = editableToolSchema.safeParse(JSON.parse(toolStringified));
if (!parse.success) {
// Loop through the errors array and create a custom errors array
const errors = parse.error.errors.map((error) => {
return {
field: error.path[0],
message: error.message,
};
});
return fail(400, { error: true, errors });
}
// modify the tool
await collections.tools.updateOne(
{ _id: tool._id },
{
$set: {
...parse.data,
updatedAt: new Date(),
searchTokens: generateSearchTokens(parse.data.displayName),
},
}
);
redirect(302, `${base}/tools/${tool._id.toString()}`);
},
};
| chat-ui/src/routes/tools/[toolId]/edit/+page.server.ts/0 | {
"file_path": "chat-ui/src/routes/tools/[toolId]/edit/+page.server.ts",
"repo_id": "chat-ui",
"token_count": 650
} |
import json
import os
from dataclasses import dataclass
import numpy as np
import pyarrow as pa
import datasets
from utils import get_duration
SPEED_TEST_N_EXAMPLES = 100_000_000_000
SPEED_TEST_CHUNK_SIZE = 10_000
RESULTS_BASEPATH, RESULTS_FILENAME = os.path.split(__file__)
RESULTS_FILE_PATH = os.path.join(RESULTS_BASEPATH, "results", RESULTS_FILENAME.replace(".py", ".json"))
def generate_100B_dataset(num_examples: int, chunk_size: int) -> datasets.Dataset:
table = pa.Table.from_pydict({"col": [0] * chunk_size})
table = pa.concat_tables([table] * (num_examples // chunk_size))
return datasets.Dataset(table, fingerprint="table_100B")
@dataclass
class RandIter:
low: int
high: int
size: int
seed: int
def __post_init__(self):
rng = np.random.default_rng(self.seed)
self._sampled_values = rng.integers(low=self.low, high=self.high, size=self.size).tolist()
def __iter__(self):
return iter(self._sampled_values)
def __len__(self):
return self.size
@get_duration
def get_first_row(dataset: datasets.Dataset):
_ = dataset[0]
@get_duration
def get_last_row(dataset: datasets.Dataset):
_ = dataset[-1]
@get_duration
def get_batch_of_1024_rows(dataset: datasets.Dataset):
_ = dataset[range(len(dataset) // 2, len(dataset) // 2 + 1024)]
@get_duration
def get_batch_of_1024_random_rows(dataset: datasets.Dataset):
_ = dataset[RandIter(0, len(dataset), 1024, seed=42)]
def benchmark_table_100B():
times = {"num examples": SPEED_TEST_N_EXAMPLES}
functions = (get_first_row, get_last_row, get_batch_of_1024_rows, get_batch_of_1024_random_rows)
print("generating dataset")
dataset = generate_100B_dataset(num_examples=SPEED_TEST_N_EXAMPLES, chunk_size=SPEED_TEST_CHUNK_SIZE)
print("Functions")
for func in functions:
print(func.__name__)
times[func.__name__] = func(dataset)
with open(RESULTS_FILE_PATH, "wb") as f:
f.write(json.dumps(times).encode("utf-8"))
if __name__ == "__main__": # useful to run the profiler
benchmark_table_100B()
| datasets/benchmarks/benchmark_getitem_100B.py/0 | {
"file_path": "datasets/benchmarks/benchmark_getitem_100B.py",
"repo_id": "datasets",
"token_count": 867
} |
# Datasets 🤝 Arrow
## What is Arrow?
[Arrow](https://arrow.apache.org/) enables large amounts of data to be processed and moved quickly. It is a specific data format that stores data in a columnar memory layout. This provides several significant advantages:
* Arrow's standard format allows [zero-copy reads](https://en.wikipedia.org/wiki/Zero-copy) which removes virtually all serialization overhead.
* Arrow is language-agnostic so it supports different programming languages.
* Arrow is column-oriented so it is faster at querying and processing slices or columns of data.
* Arrow allows for copy-free hand-offs to standard machine learning tools such as NumPy, Pandas, PyTorch, and TensorFlow.
* Arrow supports many, possibly nested, column types.
## Memory-mapping
🤗 Datasets uses Arrow for its local caching system. It allows datasets to be backed by an on-disk cache, which is memory-mapped for fast lookup.
This architecture allows for large datasets to be used on machines with relatively small device memory.
For example, loading the full English Wikipedia dataset only takes a few MB of RAM:
```python
>>> import os; import psutil; import timeit
>>> from datasets import load_dataset
# Process.memory_info is expressed in bytes, so convert to megabytes
>>> mem_before = psutil.Process(os.getpid()).memory_info().rss / (1024 * 1024)
>>> wiki = load_dataset("wikipedia", "20220301.en", split="train")
>>> mem_after = psutil.Process(os.getpid()).memory_info().rss / (1024 * 1024)
>>> print(f"RAM memory used: {(mem_after - mem_before)} MB")
RAM memory used: 50 MB
```
This is possible because the Arrow data is actually memory-mapped from disk, and not loaded in memory.
Memory-mapping allows access to data on disk, and leverages virtual memory capabilities for fast lookups.
## Performance
Iterating over a memory-mapped dataset using Arrow is fast. Iterating over Wikipedia on a laptop gives you speeds of 1-3 Gbit/s:
```python
>>> s = """batch_size = 1000
... for batch in wiki.iter(batch_size):
... ...
... """
>>> elapsed_time = timeit.timeit(stmt=s, number=1, globals=globals())
>>> print(f"Time to iterate over the {wiki.dataset_size >> 30} GB dataset: {elapsed_time:.1f} sec, "
... f"ie. {float(wiki.dataset_size >> 27)/elapsed_time:.1f} Gb/s")
Time to iterate over the 18 GB dataset: 31.8 sec, ie. 4.8 Gb/s
```
| datasets/docs/source/about_arrow.md/0 | {
"file_path": "datasets/docs/source/about_arrow.md",
"repo_id": "datasets",
"token_count": 682
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.