Upload folder using huggingface_hub
Browse files- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/config.yaml +96 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/hydra.yaml +175 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/overrides.yaml +2 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/cli.log +17 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/error.log +50 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/experiment_config.json +110 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/config.yaml +96 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/hydra.yaml +175 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/overrides.yaml +2 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/cli.log +17 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/error.log +50 -0
- runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/experiment_config.json +110 -0
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/config.yaml
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
backend:
|
| 2 |
+
name: pytorch
|
| 3 |
+
version: 2.4.0
|
| 4 |
+
_target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
|
| 5 |
+
task: text-generation
|
| 6 |
+
model: meta-llama/Llama-3.1-8B-Instruct
|
| 7 |
+
processor: meta-llama/Llama-3.1-8B-Instruct
|
| 8 |
+
library: null
|
| 9 |
+
device: cuda
|
| 10 |
+
device_ids: '0'
|
| 11 |
+
seed: 42
|
| 12 |
+
inter_op_num_threads: null
|
| 13 |
+
intra_op_num_threads: null
|
| 14 |
+
hub_kwargs: {}
|
| 15 |
+
no_weights: true
|
| 16 |
+
device_map: null
|
| 17 |
+
torch_dtype: null
|
| 18 |
+
amp_autocast: false
|
| 19 |
+
amp_dtype: null
|
| 20 |
+
eval_mode: true
|
| 21 |
+
to_bettertransformer: false
|
| 22 |
+
low_cpu_mem_usage: null
|
| 23 |
+
attn_implementation: null
|
| 24 |
+
cache_implementation: null
|
| 25 |
+
torch_compile: false
|
| 26 |
+
torch_compile_config: {}
|
| 27 |
+
quantization_scheme: null
|
| 28 |
+
quantization_config: {}
|
| 29 |
+
deepspeed_inference: false
|
| 30 |
+
deepspeed_inference_config: {}
|
| 31 |
+
peft_type: null
|
| 32 |
+
peft_config: {}
|
| 33 |
+
launcher:
|
| 34 |
+
name: process
|
| 35 |
+
_target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
|
| 36 |
+
device_isolation: false
|
| 37 |
+
device_isolation_action: warn
|
| 38 |
+
start_method: spawn
|
| 39 |
+
benchmark:
|
| 40 |
+
name: energy_star
|
| 41 |
+
_target_: optimum_benchmark.benchmarks.energy_star.benchmark.EnergyStarBenchmark
|
| 42 |
+
dataset_name: EnergyStarAI/text_generation
|
| 43 |
+
dataset_config: ''
|
| 44 |
+
dataset_split: train
|
| 45 |
+
num_samples: 1000
|
| 46 |
+
input_shapes:
|
| 47 |
+
batch_size: 1
|
| 48 |
+
text_column_name: text
|
| 49 |
+
truncation: true
|
| 50 |
+
max_length: -1
|
| 51 |
+
dataset_prefix1: ''
|
| 52 |
+
dataset_prefix2: ''
|
| 53 |
+
t5_task: ''
|
| 54 |
+
image_column_name: image
|
| 55 |
+
resize: false
|
| 56 |
+
question_column_name: question
|
| 57 |
+
context_column_name: context
|
| 58 |
+
sentence1_column_name: sentence1
|
| 59 |
+
sentence2_column_name: sentence2
|
| 60 |
+
audio_column_name: audio
|
| 61 |
+
iterations: 10
|
| 62 |
+
warmup_runs: 10
|
| 63 |
+
energy: true
|
| 64 |
+
forward_kwargs: {}
|
| 65 |
+
generate_kwargs:
|
| 66 |
+
max_new_tokens: 10
|
| 67 |
+
min_new_tokens: 10
|
| 68 |
+
call_kwargs: {}
|
| 69 |
+
experiment_name: text_generation
|
| 70 |
+
environment:
|
| 71 |
+
cpu: ' AMD EPYC 7R32'
|
| 72 |
+
cpu_count: 48
|
| 73 |
+
cpu_ram_mb: 200472.73984
|
| 74 |
+
system: Linux
|
| 75 |
+
machine: x86_64
|
| 76 |
+
platform: Linux-5.10.192-183.736.amzn2.x86_64-x86_64-with-glibc2.35
|
| 77 |
+
processor: x86_64
|
| 78 |
+
python_version: 3.9.20
|
| 79 |
+
gpu:
|
| 80 |
+
- NVIDIA A10G
|
| 81 |
+
gpu_count: 1
|
| 82 |
+
gpu_vram_mb: 24146608128
|
| 83 |
+
optimum_benchmark_version: 0.2.0
|
| 84 |
+
optimum_benchmark_commit: null
|
| 85 |
+
transformers_version: 4.44.0
|
| 86 |
+
transformers_commit: null
|
| 87 |
+
accelerate_version: 0.33.0
|
| 88 |
+
accelerate_commit: null
|
| 89 |
+
diffusers_version: 0.30.0
|
| 90 |
+
diffusers_commit: null
|
| 91 |
+
optimum_version: null
|
| 92 |
+
optimum_commit: null
|
| 93 |
+
timm_version: null
|
| 94 |
+
timm_commit: null
|
| 95 |
+
peft_version: null
|
| 96 |
+
peft_commit: null
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/hydra.yaml
ADDED
|
@@ -0,0 +1,175 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
hydra:
|
| 2 |
+
run:
|
| 3 |
+
dir: ./runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54
|
| 4 |
+
sweep:
|
| 5 |
+
dir: sweeps/${experiment_name}/${backend.model}/${now:%Y-%m-%d-%H-%M-%S}
|
| 6 |
+
subdir: ${hydra.job.num}
|
| 7 |
+
launcher:
|
| 8 |
+
_target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
|
| 9 |
+
sweeper:
|
| 10 |
+
_target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
|
| 11 |
+
max_batch_size: null
|
| 12 |
+
params: null
|
| 13 |
+
help:
|
| 14 |
+
app_name: ${hydra.job.name}
|
| 15 |
+
header: '${hydra.help.app_name} is powered by Hydra.
|
| 16 |
+
|
| 17 |
+
'
|
| 18 |
+
footer: 'Powered by Hydra (https://hydra.cc)
|
| 19 |
+
|
| 20 |
+
Use --hydra-help to view Hydra specific help
|
| 21 |
+
|
| 22 |
+
'
|
| 23 |
+
template: '${hydra.help.header}
|
| 24 |
+
|
| 25 |
+
== Configuration groups ==
|
| 26 |
+
|
| 27 |
+
Compose your configuration from those groups (group=option)
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
$APP_CONFIG_GROUPS
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
== Config ==
|
| 34 |
+
|
| 35 |
+
Override anything in the config (foo.bar=value)
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
$CONFIG
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
${hydra.help.footer}
|
| 42 |
+
|
| 43 |
+
'
|
| 44 |
+
hydra_help:
|
| 45 |
+
template: 'Hydra (${hydra.runtime.version})
|
| 46 |
+
|
| 47 |
+
See https://hydra.cc for more info.
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
== Flags ==
|
| 51 |
+
|
| 52 |
+
$FLAGS_HELP
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
== Configuration groups ==
|
| 56 |
+
|
| 57 |
+
Compose your configuration from those groups (For example, append hydra/job_logging=disabled
|
| 58 |
+
to command line)
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
$HYDRA_CONFIG_GROUPS
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
Use ''--cfg hydra'' to Show the Hydra config.
|
| 65 |
+
|
| 66 |
+
'
|
| 67 |
+
hydra_help: ???
|
| 68 |
+
hydra_logging:
|
| 69 |
+
version: 1
|
| 70 |
+
formatters:
|
| 71 |
+
colorlog:
|
| 72 |
+
(): colorlog.ColoredFormatter
|
| 73 |
+
format: '[%(cyan)s%(asctime)s%(reset)s][%(purple)sHYDRA%(reset)s] %(message)s'
|
| 74 |
+
handlers:
|
| 75 |
+
console:
|
| 76 |
+
class: logging.StreamHandler
|
| 77 |
+
formatter: colorlog
|
| 78 |
+
stream: ext://sys.stdout
|
| 79 |
+
root:
|
| 80 |
+
level: INFO
|
| 81 |
+
handlers:
|
| 82 |
+
- console
|
| 83 |
+
disable_existing_loggers: false
|
| 84 |
+
job_logging:
|
| 85 |
+
version: 1
|
| 86 |
+
formatters:
|
| 87 |
+
simple:
|
| 88 |
+
format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
|
| 89 |
+
colorlog:
|
| 90 |
+
(): colorlog.ColoredFormatter
|
| 91 |
+
format: '[%(cyan)s%(asctime)s%(reset)s][%(blue)s%(name)s%(reset)s][%(log_color)s%(levelname)s%(reset)s]
|
| 92 |
+
- %(message)s'
|
| 93 |
+
log_colors:
|
| 94 |
+
DEBUG: purple
|
| 95 |
+
INFO: green
|
| 96 |
+
WARNING: yellow
|
| 97 |
+
ERROR: red
|
| 98 |
+
CRITICAL: red
|
| 99 |
+
handlers:
|
| 100 |
+
console:
|
| 101 |
+
class: logging.StreamHandler
|
| 102 |
+
formatter: colorlog
|
| 103 |
+
stream: ext://sys.stdout
|
| 104 |
+
file:
|
| 105 |
+
class: logging.FileHandler
|
| 106 |
+
formatter: simple
|
| 107 |
+
filename: ${hydra.job.name}.log
|
| 108 |
+
root:
|
| 109 |
+
level: INFO
|
| 110 |
+
handlers:
|
| 111 |
+
- console
|
| 112 |
+
- file
|
| 113 |
+
disable_existing_loggers: false
|
| 114 |
+
env: {}
|
| 115 |
+
mode: RUN
|
| 116 |
+
searchpath: []
|
| 117 |
+
callbacks: {}
|
| 118 |
+
output_subdir: .hydra
|
| 119 |
+
overrides:
|
| 120 |
+
hydra:
|
| 121 |
+
- hydra.run.dir=./runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54
|
| 122 |
+
- hydra.mode=RUN
|
| 123 |
+
task:
|
| 124 |
+
- backend.model=meta-llama/Llama-3.1-8B-Instruct
|
| 125 |
+
- backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
| 126 |
+
job:
|
| 127 |
+
name: cli
|
| 128 |
+
chdir: true
|
| 129 |
+
override_dirname: backend.model=meta-llama/Llama-3.1-8B-Instruct,backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
| 130 |
+
id: ???
|
| 131 |
+
num: ???
|
| 132 |
+
config_name: text_generation
|
| 133 |
+
env_set:
|
| 134 |
+
OVERRIDE_BENCHMARKS: '1'
|
| 135 |
+
env_copy: []
|
| 136 |
+
config:
|
| 137 |
+
override_dirname:
|
| 138 |
+
kv_sep: '='
|
| 139 |
+
item_sep: ','
|
| 140 |
+
exclude_keys: []
|
| 141 |
+
runtime:
|
| 142 |
+
version: 1.3.2
|
| 143 |
+
version_base: '1.3'
|
| 144 |
+
cwd: /
|
| 145 |
+
config_sources:
|
| 146 |
+
- path: hydra.conf
|
| 147 |
+
schema: pkg
|
| 148 |
+
provider: hydra
|
| 149 |
+
- path: optimum_benchmark
|
| 150 |
+
schema: pkg
|
| 151 |
+
provider: main
|
| 152 |
+
- path: hydra_plugins.hydra_colorlog.conf
|
| 153 |
+
schema: pkg
|
| 154 |
+
provider: hydra-colorlog
|
| 155 |
+
- path: /optimum-benchmark/examples/energy_star
|
| 156 |
+
schema: file
|
| 157 |
+
provider: command-line
|
| 158 |
+
- path: ''
|
| 159 |
+
schema: structured
|
| 160 |
+
provider: schema
|
| 161 |
+
output_dir: /runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54
|
| 162 |
+
choices:
|
| 163 |
+
benchmark: energy_star
|
| 164 |
+
launcher: process
|
| 165 |
+
backend: pytorch
|
| 166 |
+
hydra/env: default
|
| 167 |
+
hydra/callbacks: null
|
| 168 |
+
hydra/job_logging: colorlog
|
| 169 |
+
hydra/hydra_logging: colorlog
|
| 170 |
+
hydra/hydra_help: default
|
| 171 |
+
hydra/help: default
|
| 172 |
+
hydra/sweeper: basic
|
| 173 |
+
hydra/launcher: basic
|
| 174 |
+
hydra/output: default
|
| 175 |
+
verbose: false
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/.hydra/overrides.yaml
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
- backend.model=meta-llama/Llama-3.1-8B-Instruct
|
| 2 |
+
- backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/cli.log
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[2024-10-23 20:29:57,163][launcher][INFO] - ََAllocating process launcher
|
| 2 |
+
[2024-10-23 20:29:57,163][process][INFO] - + Setting multiprocessing start method to spawn.
|
| 3 |
+
[2024-10-23 20:29:57,173][process][INFO] - + Launched benchmark in isolated process 200.
|
| 4 |
+
[PROC-0][2024-10-23 20:29:59,693][datasets][INFO] - PyTorch version 2.4.0 available.
|
| 5 |
+
[PROC-0][2024-10-23 20:30:00,603][backend][INFO] - َAllocating pytorch backend
|
| 6 |
+
[PROC-0][2024-10-23 20:30:00,603][backend][INFO] - + Setting random seed to 42
|
| 7 |
+
[PROC-0][2024-10-23 20:30:01,832][pytorch][INFO] - + Using AutoModel class AutoModelForCausalLM
|
| 8 |
+
[PROC-0][2024-10-23 20:30:01,832][pytorch][INFO] - + Creating backend temporary directory
|
| 9 |
+
[PROC-0][2024-10-23 20:30:01,832][pytorch][INFO] - + Loading model with random weights
|
| 10 |
+
[PROC-0][2024-10-23 20:30:01,832][pytorch][INFO] - + Creating no weights model
|
| 11 |
+
[PROC-0][2024-10-23 20:30:01,833][pytorch][INFO] - + Creating no weights model directory
|
| 12 |
+
[PROC-0][2024-10-23 20:30:01,833][pytorch][INFO] - + Creating no weights model state dict
|
| 13 |
+
[PROC-0][2024-10-23 20:30:01,852][pytorch][INFO] - + Saving no weights model safetensors
|
| 14 |
+
[PROC-0][2024-10-23 20:30:01,853][pytorch][INFO] - + Saving no weights model pretrained config
|
| 15 |
+
[PROC-0][2024-10-23 20:30:01,853][pytorch][INFO] - + Loading no weights AutoModel
|
| 16 |
+
[PROC-0][2024-10-23 20:30:01,854][pytorch][INFO] - + Loading model directly on device: cuda
|
| 17 |
+
[2024-10-23 20:30:02,845][experiment][ERROR] - Error during experiment
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/error.log
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Error executing job with overrides: ['backend.model=meta-llama/Llama-3.1-8B-Instruct', 'backend.processor=meta-llama/Llama-3.1-8B-Instruct']
|
| 2 |
+
Traceback (most recent call last):
|
| 3 |
+
File "/optimum-benchmark/optimum_benchmark/cli.py", line 65, in benchmark_cli
|
| 4 |
+
benchmark_report: BenchmarkReport = launch(experiment_config=experiment_config)
|
| 5 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 102, in launch
|
| 6 |
+
raise error
|
| 7 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 90, in launch
|
| 8 |
+
report = launcher.launch(run, experiment_config.benchmark, experiment_config.backend)
|
| 9 |
+
File "/optimum-benchmark/optimum_benchmark/launchers/process/launcher.py", line 47, in launch
|
| 10 |
+
while not process_context.join():
|
| 11 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 189, in join
|
| 12 |
+
raise ProcessRaisedException(msg, error_index, failed_process.pid)
|
| 13 |
+
torch.multiprocessing.spawn.ProcessRaisedException:
|
| 14 |
+
|
| 15 |
+
-- Process 0 terminated with the following error:
|
| 16 |
+
Traceback (most recent call last):
|
| 17 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 76, in _wrap
|
| 18 |
+
fn(i, *args)
|
| 19 |
+
File "/optimum-benchmark/optimum_benchmark/launchers/process/launcher.py", line 63, in entrypoint
|
| 20 |
+
worker_output = worker(*worker_args)
|
| 21 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 55, in run
|
| 22 |
+
backend: Backend = backend_factory(backend_config)
|
| 23 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 81, in __init__
|
| 24 |
+
self.load_model_with_no_weights()
|
| 25 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 246, in load_model_with_no_weights
|
| 26 |
+
self.load_model_from_pretrained()
|
| 27 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 204, in load_model_from_pretrained
|
| 28 |
+
self.pretrained_model = self.automodel_class.from_pretrained(
|
| 29 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/auto/auto_factory.py", line 564, in from_pretrained
|
| 30 |
+
return model_class.from_pretrained(
|
| 31 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/modeling_utils.py", line 3810, in from_pretrained
|
| 32 |
+
model = cls(config, *model_args, **model_kwargs)
|
| 33 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 1116, in __init__
|
| 34 |
+
self.model = LlamaModel(config)
|
| 35 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 902, in __init__
|
| 36 |
+
[LlamaDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
|
| 37 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 902, in <listcomp>
|
| 38 |
+
[LlamaDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
|
| 39 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 691, in __init__
|
| 40 |
+
self.mlp = LlamaMLP(config)
|
| 41 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 287, in __init__
|
| 42 |
+
self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=config.mlp_bias)
|
| 43 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/nn/modules/linear.py", line 99, in __init__
|
| 44 |
+
self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs))
|
| 45 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/utils/_device.py", line 79, in __torch_function__
|
| 46 |
+
return func(*args, **kwargs)
|
| 47 |
+
torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 224.00 MiB. GPU 0 has a total capacity of 22.19 GiB of which 69.50 MiB is free. Process 28476 has 22.12 GiB memory in use. Of the allocated memory 21.83 GiB is allocated by PyTorch, and 1.24 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-29-54/experiment_config.json
ADDED
|
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_name": "text_generation",
|
| 3 |
+
"backend": {
|
| 4 |
+
"name": "pytorch",
|
| 5 |
+
"version": "2.4.0",
|
| 6 |
+
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
| 7 |
+
"task": "text-generation",
|
| 8 |
+
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
| 9 |
+
"processor": "meta-llama/Llama-3.1-8B-Instruct",
|
| 10 |
+
"library": "transformers",
|
| 11 |
+
"device": "cuda",
|
| 12 |
+
"device_ids": "0",
|
| 13 |
+
"seed": 42,
|
| 14 |
+
"inter_op_num_threads": null,
|
| 15 |
+
"intra_op_num_threads": null,
|
| 16 |
+
"hub_kwargs": {
|
| 17 |
+
"revision": "main",
|
| 18 |
+
"force_download": false,
|
| 19 |
+
"local_files_only": false,
|
| 20 |
+
"trust_remote_code": true
|
| 21 |
+
},
|
| 22 |
+
"no_weights": true,
|
| 23 |
+
"device_map": null,
|
| 24 |
+
"torch_dtype": null,
|
| 25 |
+
"amp_autocast": false,
|
| 26 |
+
"amp_dtype": null,
|
| 27 |
+
"eval_mode": true,
|
| 28 |
+
"to_bettertransformer": false,
|
| 29 |
+
"low_cpu_mem_usage": null,
|
| 30 |
+
"attn_implementation": null,
|
| 31 |
+
"cache_implementation": null,
|
| 32 |
+
"torch_compile": false,
|
| 33 |
+
"torch_compile_config": {},
|
| 34 |
+
"quantization_scheme": null,
|
| 35 |
+
"quantization_config": {},
|
| 36 |
+
"deepspeed_inference": false,
|
| 37 |
+
"deepspeed_inference_config": {},
|
| 38 |
+
"peft_type": null,
|
| 39 |
+
"peft_config": {}
|
| 40 |
+
},
|
| 41 |
+
"launcher": {
|
| 42 |
+
"name": "process",
|
| 43 |
+
"_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
|
| 44 |
+
"device_isolation": false,
|
| 45 |
+
"device_isolation_action": "warn",
|
| 46 |
+
"start_method": "spawn"
|
| 47 |
+
},
|
| 48 |
+
"benchmark": {
|
| 49 |
+
"name": "energy_star",
|
| 50 |
+
"_target_": "optimum_benchmark.benchmarks.energy_star.benchmark.EnergyStarBenchmark",
|
| 51 |
+
"dataset_name": "EnergyStarAI/text_generation",
|
| 52 |
+
"dataset_config": "",
|
| 53 |
+
"dataset_split": "train",
|
| 54 |
+
"num_samples": 1000,
|
| 55 |
+
"input_shapes": {
|
| 56 |
+
"batch_size": 1
|
| 57 |
+
},
|
| 58 |
+
"text_column_name": "text",
|
| 59 |
+
"truncation": true,
|
| 60 |
+
"max_length": -1,
|
| 61 |
+
"dataset_prefix1": "",
|
| 62 |
+
"dataset_prefix2": "",
|
| 63 |
+
"t5_task": "",
|
| 64 |
+
"image_column_name": "image",
|
| 65 |
+
"resize": false,
|
| 66 |
+
"question_column_name": "question",
|
| 67 |
+
"context_column_name": "context",
|
| 68 |
+
"sentence1_column_name": "sentence1",
|
| 69 |
+
"sentence2_column_name": "sentence2",
|
| 70 |
+
"audio_column_name": "audio",
|
| 71 |
+
"iterations": 10,
|
| 72 |
+
"warmup_runs": 10,
|
| 73 |
+
"energy": true,
|
| 74 |
+
"forward_kwargs": {},
|
| 75 |
+
"generate_kwargs": {
|
| 76 |
+
"max_new_tokens": 10,
|
| 77 |
+
"min_new_tokens": 10
|
| 78 |
+
},
|
| 79 |
+
"call_kwargs": {}
|
| 80 |
+
},
|
| 81 |
+
"environment": {
|
| 82 |
+
"cpu": " AMD EPYC 7R32",
|
| 83 |
+
"cpu_count": 48,
|
| 84 |
+
"cpu_ram_mb": 200472.73984,
|
| 85 |
+
"system": "Linux",
|
| 86 |
+
"machine": "x86_64",
|
| 87 |
+
"platform": "Linux-5.10.192-183.736.amzn2.x86_64-x86_64-with-glibc2.35",
|
| 88 |
+
"processor": "x86_64",
|
| 89 |
+
"python_version": "3.9.20",
|
| 90 |
+
"gpu": [
|
| 91 |
+
"NVIDIA A10G"
|
| 92 |
+
],
|
| 93 |
+
"gpu_count": 1,
|
| 94 |
+
"gpu_vram_mb": 24146608128,
|
| 95 |
+
"optimum_benchmark_version": "0.2.0",
|
| 96 |
+
"optimum_benchmark_commit": null,
|
| 97 |
+
"transformers_version": "4.44.0",
|
| 98 |
+
"transformers_commit": null,
|
| 99 |
+
"accelerate_version": "0.33.0",
|
| 100 |
+
"accelerate_commit": null,
|
| 101 |
+
"diffusers_version": "0.30.0",
|
| 102 |
+
"diffusers_commit": null,
|
| 103 |
+
"optimum_version": null,
|
| 104 |
+
"optimum_commit": null,
|
| 105 |
+
"timm_version": null,
|
| 106 |
+
"timm_commit": null,
|
| 107 |
+
"peft_version": null,
|
| 108 |
+
"peft_commit": null
|
| 109 |
+
}
|
| 110 |
+
}
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/config.yaml
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
backend:
|
| 2 |
+
name: pytorch
|
| 3 |
+
version: 2.4.0
|
| 4 |
+
_target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
|
| 5 |
+
task: text-generation
|
| 6 |
+
model: meta-llama/Llama-3.1-8B-Instruct
|
| 7 |
+
processor: meta-llama/Llama-3.1-8B-Instruct
|
| 8 |
+
library: null
|
| 9 |
+
device: cuda
|
| 10 |
+
device_ids: '0'
|
| 11 |
+
seed: 42
|
| 12 |
+
inter_op_num_threads: null
|
| 13 |
+
intra_op_num_threads: null
|
| 14 |
+
hub_kwargs: {}
|
| 15 |
+
no_weights: true
|
| 16 |
+
device_map: null
|
| 17 |
+
torch_dtype: null
|
| 18 |
+
amp_autocast: false
|
| 19 |
+
amp_dtype: null
|
| 20 |
+
eval_mode: true
|
| 21 |
+
to_bettertransformer: false
|
| 22 |
+
low_cpu_mem_usage: null
|
| 23 |
+
attn_implementation: null
|
| 24 |
+
cache_implementation: null
|
| 25 |
+
torch_compile: false
|
| 26 |
+
torch_compile_config: {}
|
| 27 |
+
quantization_scheme: null
|
| 28 |
+
quantization_config: {}
|
| 29 |
+
deepspeed_inference: false
|
| 30 |
+
deepspeed_inference_config: {}
|
| 31 |
+
peft_type: null
|
| 32 |
+
peft_config: {}
|
| 33 |
+
launcher:
|
| 34 |
+
name: process
|
| 35 |
+
_target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
|
| 36 |
+
device_isolation: false
|
| 37 |
+
device_isolation_action: warn
|
| 38 |
+
start_method: spawn
|
| 39 |
+
benchmark:
|
| 40 |
+
name: energy_star
|
| 41 |
+
_target_: optimum_benchmark.benchmarks.energy_star.benchmark.EnergyStarBenchmark
|
| 42 |
+
dataset_name: EnergyStarAI/text_generation
|
| 43 |
+
dataset_config: ''
|
| 44 |
+
dataset_split: train
|
| 45 |
+
num_samples: 1000
|
| 46 |
+
input_shapes:
|
| 47 |
+
batch_size: 1
|
| 48 |
+
text_column_name: text
|
| 49 |
+
truncation: true
|
| 50 |
+
max_length: -1
|
| 51 |
+
dataset_prefix1: ''
|
| 52 |
+
dataset_prefix2: ''
|
| 53 |
+
t5_task: ''
|
| 54 |
+
image_column_name: image
|
| 55 |
+
resize: false
|
| 56 |
+
question_column_name: question
|
| 57 |
+
context_column_name: context
|
| 58 |
+
sentence1_column_name: sentence1
|
| 59 |
+
sentence2_column_name: sentence2
|
| 60 |
+
audio_column_name: audio
|
| 61 |
+
iterations: 10
|
| 62 |
+
warmup_runs: 10
|
| 63 |
+
energy: true
|
| 64 |
+
forward_kwargs: {}
|
| 65 |
+
generate_kwargs:
|
| 66 |
+
max_new_tokens: 10
|
| 67 |
+
min_new_tokens: 10
|
| 68 |
+
call_kwargs: {}
|
| 69 |
+
experiment_name: text_generation
|
| 70 |
+
environment:
|
| 71 |
+
cpu: ' AMD EPYC 7R32'
|
| 72 |
+
cpu_count: 48
|
| 73 |
+
cpu_ram_mb: 200472.73984
|
| 74 |
+
system: Linux
|
| 75 |
+
machine: x86_64
|
| 76 |
+
platform: Linux-5.10.192-183.736.amzn2.x86_64-x86_64-with-glibc2.35
|
| 77 |
+
processor: x86_64
|
| 78 |
+
python_version: 3.9.20
|
| 79 |
+
gpu:
|
| 80 |
+
- NVIDIA A10G
|
| 81 |
+
gpu_count: 1
|
| 82 |
+
gpu_vram_mb: 24146608128
|
| 83 |
+
optimum_benchmark_version: 0.2.0
|
| 84 |
+
optimum_benchmark_commit: null
|
| 85 |
+
transformers_version: 4.44.0
|
| 86 |
+
transformers_commit: null
|
| 87 |
+
accelerate_version: 0.33.0
|
| 88 |
+
accelerate_commit: null
|
| 89 |
+
diffusers_version: 0.30.0
|
| 90 |
+
diffusers_commit: null
|
| 91 |
+
optimum_version: null
|
| 92 |
+
optimum_commit: null
|
| 93 |
+
timm_version: null
|
| 94 |
+
timm_commit: null
|
| 95 |
+
peft_version: null
|
| 96 |
+
peft_commit: null
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/hydra.yaml
ADDED
|
@@ -0,0 +1,175 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
hydra:
|
| 2 |
+
run:
|
| 3 |
+
dir: ./runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04
|
| 4 |
+
sweep:
|
| 5 |
+
dir: sweeps/${experiment_name}/${backend.model}/${now:%Y-%m-%d-%H-%M-%S}
|
| 6 |
+
subdir: ${hydra.job.num}
|
| 7 |
+
launcher:
|
| 8 |
+
_target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
|
| 9 |
+
sweeper:
|
| 10 |
+
_target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
|
| 11 |
+
max_batch_size: null
|
| 12 |
+
params: null
|
| 13 |
+
help:
|
| 14 |
+
app_name: ${hydra.job.name}
|
| 15 |
+
header: '${hydra.help.app_name} is powered by Hydra.
|
| 16 |
+
|
| 17 |
+
'
|
| 18 |
+
footer: 'Powered by Hydra (https://hydra.cc)
|
| 19 |
+
|
| 20 |
+
Use --hydra-help to view Hydra specific help
|
| 21 |
+
|
| 22 |
+
'
|
| 23 |
+
template: '${hydra.help.header}
|
| 24 |
+
|
| 25 |
+
== Configuration groups ==
|
| 26 |
+
|
| 27 |
+
Compose your configuration from those groups (group=option)
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
$APP_CONFIG_GROUPS
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
== Config ==
|
| 34 |
+
|
| 35 |
+
Override anything in the config (foo.bar=value)
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
$CONFIG
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
${hydra.help.footer}
|
| 42 |
+
|
| 43 |
+
'
|
| 44 |
+
hydra_help:
|
| 45 |
+
template: 'Hydra (${hydra.runtime.version})
|
| 46 |
+
|
| 47 |
+
See https://hydra.cc for more info.
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
== Flags ==
|
| 51 |
+
|
| 52 |
+
$FLAGS_HELP
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
== Configuration groups ==
|
| 56 |
+
|
| 57 |
+
Compose your configuration from those groups (For example, append hydra/job_logging=disabled
|
| 58 |
+
to command line)
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
$HYDRA_CONFIG_GROUPS
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
Use ''--cfg hydra'' to Show the Hydra config.
|
| 65 |
+
|
| 66 |
+
'
|
| 67 |
+
hydra_help: ???
|
| 68 |
+
hydra_logging:
|
| 69 |
+
version: 1
|
| 70 |
+
formatters:
|
| 71 |
+
colorlog:
|
| 72 |
+
(): colorlog.ColoredFormatter
|
| 73 |
+
format: '[%(cyan)s%(asctime)s%(reset)s][%(purple)sHYDRA%(reset)s] %(message)s'
|
| 74 |
+
handlers:
|
| 75 |
+
console:
|
| 76 |
+
class: logging.StreamHandler
|
| 77 |
+
formatter: colorlog
|
| 78 |
+
stream: ext://sys.stdout
|
| 79 |
+
root:
|
| 80 |
+
level: INFO
|
| 81 |
+
handlers:
|
| 82 |
+
- console
|
| 83 |
+
disable_existing_loggers: false
|
| 84 |
+
job_logging:
|
| 85 |
+
version: 1
|
| 86 |
+
formatters:
|
| 87 |
+
simple:
|
| 88 |
+
format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
|
| 89 |
+
colorlog:
|
| 90 |
+
(): colorlog.ColoredFormatter
|
| 91 |
+
format: '[%(cyan)s%(asctime)s%(reset)s][%(blue)s%(name)s%(reset)s][%(log_color)s%(levelname)s%(reset)s]
|
| 92 |
+
- %(message)s'
|
| 93 |
+
log_colors:
|
| 94 |
+
DEBUG: purple
|
| 95 |
+
INFO: green
|
| 96 |
+
WARNING: yellow
|
| 97 |
+
ERROR: red
|
| 98 |
+
CRITICAL: red
|
| 99 |
+
handlers:
|
| 100 |
+
console:
|
| 101 |
+
class: logging.StreamHandler
|
| 102 |
+
formatter: colorlog
|
| 103 |
+
stream: ext://sys.stdout
|
| 104 |
+
file:
|
| 105 |
+
class: logging.FileHandler
|
| 106 |
+
formatter: simple
|
| 107 |
+
filename: ${hydra.job.name}.log
|
| 108 |
+
root:
|
| 109 |
+
level: INFO
|
| 110 |
+
handlers:
|
| 111 |
+
- console
|
| 112 |
+
- file
|
| 113 |
+
disable_existing_loggers: false
|
| 114 |
+
env: {}
|
| 115 |
+
mode: RUN
|
| 116 |
+
searchpath: []
|
| 117 |
+
callbacks: {}
|
| 118 |
+
output_subdir: .hydra
|
| 119 |
+
overrides:
|
| 120 |
+
hydra:
|
| 121 |
+
- hydra.run.dir=./runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04
|
| 122 |
+
- hydra.mode=RUN
|
| 123 |
+
task:
|
| 124 |
+
- backend.model=meta-llama/Llama-3.1-8B-Instruct
|
| 125 |
+
- backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
| 126 |
+
job:
|
| 127 |
+
name: cli
|
| 128 |
+
chdir: true
|
| 129 |
+
override_dirname: backend.model=meta-llama/Llama-3.1-8B-Instruct,backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
| 130 |
+
id: ???
|
| 131 |
+
num: ???
|
| 132 |
+
config_name: text_generation
|
| 133 |
+
env_set:
|
| 134 |
+
OVERRIDE_BENCHMARKS: '1'
|
| 135 |
+
env_copy: []
|
| 136 |
+
config:
|
| 137 |
+
override_dirname:
|
| 138 |
+
kv_sep: '='
|
| 139 |
+
item_sep: ','
|
| 140 |
+
exclude_keys: []
|
| 141 |
+
runtime:
|
| 142 |
+
version: 1.3.2
|
| 143 |
+
version_base: '1.3'
|
| 144 |
+
cwd: /
|
| 145 |
+
config_sources:
|
| 146 |
+
- path: hydra.conf
|
| 147 |
+
schema: pkg
|
| 148 |
+
provider: hydra
|
| 149 |
+
- path: optimum_benchmark
|
| 150 |
+
schema: pkg
|
| 151 |
+
provider: main
|
| 152 |
+
- path: hydra_plugins.hydra_colorlog.conf
|
| 153 |
+
schema: pkg
|
| 154 |
+
provider: hydra-colorlog
|
| 155 |
+
- path: /optimum-benchmark/examples/energy_star
|
| 156 |
+
schema: file
|
| 157 |
+
provider: command-line
|
| 158 |
+
- path: ''
|
| 159 |
+
schema: structured
|
| 160 |
+
provider: schema
|
| 161 |
+
output_dir: /runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04
|
| 162 |
+
choices:
|
| 163 |
+
benchmark: energy_star
|
| 164 |
+
launcher: process
|
| 165 |
+
backend: pytorch
|
| 166 |
+
hydra/env: default
|
| 167 |
+
hydra/callbacks: null
|
| 168 |
+
hydra/job_logging: colorlog
|
| 169 |
+
hydra/hydra_logging: colorlog
|
| 170 |
+
hydra/hydra_help: default
|
| 171 |
+
hydra/help: default
|
| 172 |
+
hydra/sweeper: basic
|
| 173 |
+
hydra/launcher: basic
|
| 174 |
+
hydra/output: default
|
| 175 |
+
verbose: false
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/.hydra/overrides.yaml
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
- backend.model=meta-llama/Llama-3.1-8B-Instruct
|
| 2 |
+
- backend.processor=meta-llama/Llama-3.1-8B-Instruct
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/cli.log
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[2024-10-23 20:30:07,112][launcher][INFO] - ََAllocating process launcher
|
| 2 |
+
[2024-10-23 20:30:07,113][process][INFO] - + Setting multiprocessing start method to spawn.
|
| 3 |
+
[2024-10-23 20:30:07,123][process][INFO] - + Launched benchmark in isolated process 421.
|
| 4 |
+
[PROC-0][2024-10-23 20:30:09,639][datasets][INFO] - PyTorch version 2.4.0 available.
|
| 5 |
+
[PROC-0][2024-10-23 20:30:10,573][backend][INFO] - َAllocating pytorch backend
|
| 6 |
+
[PROC-0][2024-10-23 20:30:10,573][backend][INFO] - + Setting random seed to 42
|
| 7 |
+
[PROC-0][2024-10-23 20:30:11,062][pytorch][INFO] - + Using AutoModel class AutoModelForCausalLM
|
| 8 |
+
[PROC-0][2024-10-23 20:30:11,062][pytorch][INFO] - + Creating backend temporary directory
|
| 9 |
+
[PROC-0][2024-10-23 20:30:11,062][pytorch][INFO] - + Loading model with random weights
|
| 10 |
+
[PROC-0][2024-10-23 20:30:11,062][pytorch][INFO] - + Creating no weights model
|
| 11 |
+
[PROC-0][2024-10-23 20:30:11,063][pytorch][INFO] - + Creating no weights model directory
|
| 12 |
+
[PROC-0][2024-10-23 20:30:11,063][pytorch][INFO] - + Creating no weights model state dict
|
| 13 |
+
[PROC-0][2024-10-23 20:30:11,084][pytorch][INFO] - + Saving no weights model safetensors
|
| 14 |
+
[PROC-0][2024-10-23 20:30:11,085][pytorch][INFO] - + Saving no weights model pretrained config
|
| 15 |
+
[PROC-0][2024-10-23 20:30:11,085][pytorch][INFO] - + Loading no weights AutoModel
|
| 16 |
+
[PROC-0][2024-10-23 20:30:11,086][pytorch][INFO] - + Loading model directly on device: cuda
|
| 17 |
+
[2024-10-23 20:30:12,070][experiment][ERROR] - Error during experiment
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/error.log
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Error executing job with overrides: ['backend.model=meta-llama/Llama-3.1-8B-Instruct', 'backend.processor=meta-llama/Llama-3.1-8B-Instruct']
|
| 2 |
+
Traceback (most recent call last):
|
| 3 |
+
File "/optimum-benchmark/optimum_benchmark/cli.py", line 65, in benchmark_cli
|
| 4 |
+
benchmark_report: BenchmarkReport = launch(experiment_config=experiment_config)
|
| 5 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 102, in launch
|
| 6 |
+
raise error
|
| 7 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 90, in launch
|
| 8 |
+
report = launcher.launch(run, experiment_config.benchmark, experiment_config.backend)
|
| 9 |
+
File "/optimum-benchmark/optimum_benchmark/launchers/process/launcher.py", line 47, in launch
|
| 10 |
+
while not process_context.join():
|
| 11 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 189, in join
|
| 12 |
+
raise ProcessRaisedException(msg, error_index, failed_process.pid)
|
| 13 |
+
torch.multiprocessing.spawn.ProcessRaisedException:
|
| 14 |
+
|
| 15 |
+
-- Process 0 terminated with the following error:
|
| 16 |
+
Traceback (most recent call last):
|
| 17 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 76, in _wrap
|
| 18 |
+
fn(i, *args)
|
| 19 |
+
File "/optimum-benchmark/optimum_benchmark/launchers/process/launcher.py", line 63, in entrypoint
|
| 20 |
+
worker_output = worker(*worker_args)
|
| 21 |
+
File "/optimum-benchmark/optimum_benchmark/experiment.py", line 55, in run
|
| 22 |
+
backend: Backend = backend_factory(backend_config)
|
| 23 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 81, in __init__
|
| 24 |
+
self.load_model_with_no_weights()
|
| 25 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 246, in load_model_with_no_weights
|
| 26 |
+
self.load_model_from_pretrained()
|
| 27 |
+
File "/optimum-benchmark/optimum_benchmark/backends/pytorch/backend.py", line 204, in load_model_from_pretrained
|
| 28 |
+
self.pretrained_model = self.automodel_class.from_pretrained(
|
| 29 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/auto/auto_factory.py", line 564, in from_pretrained
|
| 30 |
+
return model_class.from_pretrained(
|
| 31 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/modeling_utils.py", line 3810, in from_pretrained
|
| 32 |
+
model = cls(config, *model_args, **model_kwargs)
|
| 33 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 1116, in __init__
|
| 34 |
+
self.model = LlamaModel(config)
|
| 35 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 902, in __init__
|
| 36 |
+
[LlamaDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
|
| 37 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 902, in <listcomp>
|
| 38 |
+
[LlamaDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
|
| 39 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 691, in __init__
|
| 40 |
+
self.mlp = LlamaMLP(config)
|
| 41 |
+
File "/opt/conda/lib/python3.9/site-packages/transformers/models/llama/modeling_llama.py", line 287, in __init__
|
| 42 |
+
self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=config.mlp_bias)
|
| 43 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/nn/modules/linear.py", line 99, in __init__
|
| 44 |
+
self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs))
|
| 45 |
+
File "/opt/conda/lib/python3.9/site-packages/torch/utils/_device.py", line 79, in __torch_function__
|
| 46 |
+
return func(*args, **kwargs)
|
| 47 |
+
torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 224.00 MiB. GPU 0 has a total capacity of 22.19 GiB of which 69.50 MiB is free. Process 28976 has 22.12 GiB memory in use. Of the allocated memory 21.83 GiB is allocated by PyTorch, and 1.24 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.
|
runs/text_generation/meta-llama/Llama-3.1-8B-Instruct/2024-10-23-20-30-04/experiment_config.json
ADDED
|
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_name": "text_generation",
|
| 3 |
+
"backend": {
|
| 4 |
+
"name": "pytorch",
|
| 5 |
+
"version": "2.4.0",
|
| 6 |
+
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
| 7 |
+
"task": "text-generation",
|
| 8 |
+
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
| 9 |
+
"processor": "meta-llama/Llama-3.1-8B-Instruct",
|
| 10 |
+
"library": "transformers",
|
| 11 |
+
"device": "cuda",
|
| 12 |
+
"device_ids": "0",
|
| 13 |
+
"seed": 42,
|
| 14 |
+
"inter_op_num_threads": null,
|
| 15 |
+
"intra_op_num_threads": null,
|
| 16 |
+
"hub_kwargs": {
|
| 17 |
+
"revision": "main",
|
| 18 |
+
"force_download": false,
|
| 19 |
+
"local_files_only": false,
|
| 20 |
+
"trust_remote_code": true
|
| 21 |
+
},
|
| 22 |
+
"no_weights": true,
|
| 23 |
+
"device_map": null,
|
| 24 |
+
"torch_dtype": null,
|
| 25 |
+
"amp_autocast": false,
|
| 26 |
+
"amp_dtype": null,
|
| 27 |
+
"eval_mode": true,
|
| 28 |
+
"to_bettertransformer": false,
|
| 29 |
+
"low_cpu_mem_usage": null,
|
| 30 |
+
"attn_implementation": null,
|
| 31 |
+
"cache_implementation": null,
|
| 32 |
+
"torch_compile": false,
|
| 33 |
+
"torch_compile_config": {},
|
| 34 |
+
"quantization_scheme": null,
|
| 35 |
+
"quantization_config": {},
|
| 36 |
+
"deepspeed_inference": false,
|
| 37 |
+
"deepspeed_inference_config": {},
|
| 38 |
+
"peft_type": null,
|
| 39 |
+
"peft_config": {}
|
| 40 |
+
},
|
| 41 |
+
"launcher": {
|
| 42 |
+
"name": "process",
|
| 43 |
+
"_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
|
| 44 |
+
"device_isolation": false,
|
| 45 |
+
"device_isolation_action": "warn",
|
| 46 |
+
"start_method": "spawn"
|
| 47 |
+
},
|
| 48 |
+
"benchmark": {
|
| 49 |
+
"name": "energy_star",
|
| 50 |
+
"_target_": "optimum_benchmark.benchmarks.energy_star.benchmark.EnergyStarBenchmark",
|
| 51 |
+
"dataset_name": "EnergyStarAI/text_generation",
|
| 52 |
+
"dataset_config": "",
|
| 53 |
+
"dataset_split": "train",
|
| 54 |
+
"num_samples": 1000,
|
| 55 |
+
"input_shapes": {
|
| 56 |
+
"batch_size": 1
|
| 57 |
+
},
|
| 58 |
+
"text_column_name": "text",
|
| 59 |
+
"truncation": true,
|
| 60 |
+
"max_length": -1,
|
| 61 |
+
"dataset_prefix1": "",
|
| 62 |
+
"dataset_prefix2": "",
|
| 63 |
+
"t5_task": "",
|
| 64 |
+
"image_column_name": "image",
|
| 65 |
+
"resize": false,
|
| 66 |
+
"question_column_name": "question",
|
| 67 |
+
"context_column_name": "context",
|
| 68 |
+
"sentence1_column_name": "sentence1",
|
| 69 |
+
"sentence2_column_name": "sentence2",
|
| 70 |
+
"audio_column_name": "audio",
|
| 71 |
+
"iterations": 10,
|
| 72 |
+
"warmup_runs": 10,
|
| 73 |
+
"energy": true,
|
| 74 |
+
"forward_kwargs": {},
|
| 75 |
+
"generate_kwargs": {
|
| 76 |
+
"max_new_tokens": 10,
|
| 77 |
+
"min_new_tokens": 10
|
| 78 |
+
},
|
| 79 |
+
"call_kwargs": {}
|
| 80 |
+
},
|
| 81 |
+
"environment": {
|
| 82 |
+
"cpu": " AMD EPYC 7R32",
|
| 83 |
+
"cpu_count": 48,
|
| 84 |
+
"cpu_ram_mb": 200472.73984,
|
| 85 |
+
"system": "Linux",
|
| 86 |
+
"machine": "x86_64",
|
| 87 |
+
"platform": "Linux-5.10.192-183.736.amzn2.x86_64-x86_64-with-glibc2.35",
|
| 88 |
+
"processor": "x86_64",
|
| 89 |
+
"python_version": "3.9.20",
|
| 90 |
+
"gpu": [
|
| 91 |
+
"NVIDIA A10G"
|
| 92 |
+
],
|
| 93 |
+
"gpu_count": 1,
|
| 94 |
+
"gpu_vram_mb": 24146608128,
|
| 95 |
+
"optimum_benchmark_version": "0.2.0",
|
| 96 |
+
"optimum_benchmark_commit": null,
|
| 97 |
+
"transformers_version": "4.44.0",
|
| 98 |
+
"transformers_commit": null,
|
| 99 |
+
"accelerate_version": "0.33.0",
|
| 100 |
+
"accelerate_commit": null,
|
| 101 |
+
"diffusers_version": "0.30.0",
|
| 102 |
+
"diffusers_commit": null,
|
| 103 |
+
"optimum_version": null,
|
| 104 |
+
"optimum_commit": null,
|
| 105 |
+
"timm_version": null,
|
| 106 |
+
"timm_commit": null,
|
| 107 |
+
"peft_version": null,
|
| 108 |
+
"peft_commit": null
|
| 109 |
+
}
|
| 110 |
+
}
|