| ======================== | |
| START TIME: Wed Jul 3 20:51:52 UTC 2024 | |
| python3 version = Python 3.10.14 | |
| ======================== | |
| The token has not been saved to the git credentials helper. Pass `add_to_git_credential=True` in this function directly or `--add-to-git-credential` if using via `huggingface-cli` if you want to set the git credential as well. | |
| Token is valid (permission: write). | |
| Your token has been saved to /admin/home/ferdinand_mom/.cache/huggingface/token | |
| Login successful | |
| Already on 'bench_cluster' | |
| M examples/config_tiny_llama.py | |
| M examples/config_tiny_llama.yaml | |
| M examples/train_tiny_llama.sh | |
| M src/nanotron/models/llama.py | |
| M src/nanotron/trainer.py | |
| Your branch is up to date with 'origin/bench_cluster'. | |
| Job status: RUNNING | |
| W0703 20:51:54.696000 139757201868608 torch/distributed/run.py:757] | |
| W0703 20:51:54.696000 139757201868608 torch/distributed/run.py:757] ***************************************** | |
| W0703 20:51:54.696000 139757201868608 torch/distributed/run.py:757] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. | |
| W0703 20:51:54.696000 139757201868608 torch/distributed/run.py:757] ***************************************** | |
| [default0]:07/03/2024 20:52:10 [WARNING|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Vocab Size Padding] Padded vocab (size: 50257) with 3 dummy tokens (new size: 50260) | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Config: | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Config(general=GeneralArgs(project='bench_cluster', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: run='%date_%jobid', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: seed=42, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: step=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: consumed_train_samples=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: benchmark_csv_path=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: ignore_sanity_checks=True), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: parallelism=ParallelismArgs(dp=2, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pp=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tp=4, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pp_engine=<nanotron.parallel.pipeline_parallel.engine.OneForwardOneBackwardPipelineEngine object at 0x7f4ac43c0670>, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tp_mode=<TensorParallelLinearMode.REDUCE_SCATTER: 2>, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tp_linear_async_communication=False, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: expert_parallel_size=1), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: eos_token_id=2, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hidden_act='silu', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hidden_size=2048, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: initializer_range=0.02, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: intermediate_size=4096, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: is_llama_config=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: max_position_embeddings=4096, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_attention_heads=32, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_hidden_layers=24, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_key_value_heads=32, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pad_token_id=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pretraining_tp=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rms_norm_eps=1e-05, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rope_scaling=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rope_theta=10000.0, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tie_word_embeddings=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: use_cache=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: vocab_size=50260), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: init_method=RandomInit(std=0.025), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: dtype=torch.bfloat16, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: make_vocab_size_divisible_by=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: ddp_bucket_cap_mb=25), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tokenizer=TokenizerArgs(tokenizer_name_or_path='openai-community/gpt2', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tokenizer_revision=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tokenizer_max_length=None), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: checkpoints=CheckpointsArgs(checkpoints_path=Path('/dev/null'), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: checkpoint_interval=100000, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: save_initial_state=False, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: resume_checkpoint_path=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: checkpoints_path_is_shared_file_system=False), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: logging=LoggingArgs(log_level='info', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: log_level_replica='info', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration_step_info_interval=1), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tokens=TokensArgs(sequence_length=4096, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: train_steps=20, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: micro_batch_size=2, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: batch_accumulation_per_replica=256, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: val_check_interval=-1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: limit_val_batches=0, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: limit_test_batches=0), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: adam_beta1=0.9, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: adam_beta2=0.95, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: torch_adam_is_fused=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: name='adamW'), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: zero_stage=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: weight_decay=0.01, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: clip_grad=1.0, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: accumulate_grad_in_fp32=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0001, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lr_warmup_steps=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lr_warmup_style='linear', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lr_decay_style='linear', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lr_decay_steps=19, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lr_decay_starting_step=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: min_decay_lr=1e-05)), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: data_stages=[DatasetStageArgs(name='Training Stage', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: start_training_step=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: data=DataArgs(dataset=PretrainDatasetsArgs(hf_dataset_or_datasets='roneneldan/TinyStories', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hf_dataset_splits='train', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hf_dataset_config_name=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: dataset_processing_num_proc_per_process=64, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: dataset_overwrite_cache=False, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: text_column_name='text'), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: seed=42, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_loading_workers=0))], | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: profiler=ProfilerArgs(profiler_export_path=Path('/fsx/ferdinandmom/ferdinand-hf/bench_cluster/results/llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-2')), | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: lighteval=None) | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Model Config: | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: LlamaConfig(bos_token_id=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: eos_token_id=2, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hidden_act='silu', | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: hidden_size=2048, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: initializer_range=0.02, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: intermediate_size=4096, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: is_llama_config=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: max_position_embeddings=4096, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_attention_heads=32, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_hidden_layers=24, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: num_key_value_heads=32, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pad_token_id=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: pretraining_tp=1, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rms_norm_eps=1e-05, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rope_scaling=None, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: rope_theta=10000.0, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: tie_word_embeddings=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: use_cache=True, | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: vocab_size=50260) | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Building model.. | |
| [default0]:07/03/2024 20:52:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Setting PP block ranks... | |
| [default4]:07/03/2024 20:52:23 [INFO|DP=1|PP=0|TP=0|ip-26-0-174-36]: No checkpoint path provided. | |
| [default1]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=1|ip-26-0-174-36]: Local number of parameters: 277M (529.27MiB) | |
| [default1]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=1|ip-26-0-174-36]: [After model building] Memory usage: 554.21MiB. Peak allocated: 606.24MiB Peak reserved: 608.00MiB | |
| [default1]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=1|ip-26-0-174-36]: No checkpoint path provided. | |
| [default2]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=2|ip-26-0-174-36]: Local number of parameters: 277M (529.27MiB) | |
| [default2]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=2|ip-26-0-174-36]: [After model building] Memory usage: 554.21MiB. Peak allocated: 606.24MiB Peak reserved: 608.00MiB | |
| [default2]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=2|ip-26-0-174-36]: No checkpoint path provided. | |
| [default0]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Total number of parameters: 1.11G (2117.09MiB) | |
| [default0]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Local number of parameters: 277M (529.27MiB) | |
| [default0]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [After model building] Memory usage: 554.21MiB. Peak allocated: 606.24MiB Peak reserved: 608.00MiB | |
| [default0]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: No checkpoint path provided. | |
| [default0]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Parametrizing model parameters using StandardParametrizator | |
| [default5]:07/03/2024 20:52:23 [INFO|DP=1|PP=0|TP=1|ip-26-0-174-36]: No checkpoint path provided. | |
| [default6]:07/03/2024 20:52:23 [INFO|DP=1|PP=0|TP=2|ip-26-0-174-36]: No checkpoint path provided. | |
| [default3]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=3|ip-26-0-174-36]: Local number of parameters: 277M (529.27MiB) | |
| [default3]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=3|ip-26-0-174-36]: [After model building] Memory usage: 554.21MiB. Peak allocated: 606.24MiB Peak reserved: 608.00MiB | |
| [default3]:07/03/2024 20:52:23 [INFO|DP=0|PP=0|TP=3|ip-26-0-174-36]: No checkpoint path provided. | |
| [default7]:07/03/2024 20:52:23 [INFO|DP=1|PP=0|TP=3|ip-26-0-174-36]: No checkpoint path provided. | |
| [default0]:07/03/2024 20:52:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Optimizer Building] Using LearningRateForSP as learning rate | |
| [default0]:07/03/2024 20:52:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [ZeRO sharding] Size of optimizer params per rank: | |
| [default0]:07/03/2024 20:52:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [ZeRO sharding] DP Rank 0 has 139M out of 277M (50.00%) params' optimizer states | |
| [default0]:07/03/2024 20:52:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [ZeRO sharding] DP Rank 1 has 139M out of 277M (50.00%) params' optimizer states | |
| [default0]:07/03/2024 20:52:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Training Plan] Stage Training Stage has 19 remaining training steps and has consumed 0 samples | |
| [default0]:07/03/2024 20:52:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Using `datasets` library | |
| [default0]:07/03/2024 20:52:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Loading tokenizer from openai-community/gpt2 and transformers/hf_hub versions ('4.41.2', '0.23.4') | |
| [default0]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default0]:07/03/2024 20:52:27 [WARNING|DP=0|PP=0|TP=0|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Training Plan] There are 1 training stages | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Stage Training Stage] start from step 1 | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: [Start training] datetime: 2024-07-03 20:52:27.785097 | mbs: 2 | grad_accum: 256 | global_batch_size: 1024 | sequence_length: 4096 | train_steps: 20 | start_iteration_step: 0 | consumed_train_samples: 0 | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Resuming training from stage Training Stage, it has trained for 0 samples and has 19 remaining train steps | |
| [default0]:07/03/2024 20:52:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 2142.76MiB. Peak allocated 2142.76MiB. Peak reserved: 2198.00MiB | |
| [default1]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default2]:07/03/2024 20:52:27 [WARNING|DP=0|PP=0|TP=2|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default1]:07/03/2024 20:52:27 [WARNING|DP=0|PP=0|TP=1|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default3]:07/03/2024 20:52:27 [WARNING|DP=0|PP=0|TP=3|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default3]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default2]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default5]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default7]:07/03/2024 20:52:27 [WARNING|DP=1|PP=0|TP=3|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default7]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default4]:07/03/2024 20:52:27 [WARNING|DP=1|PP=0|TP=0|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default5]:07/03/2024 20:52:27 [WARNING|DP=1|PP=0|TP=1|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default6]:07/03/2024 20:52:27 [WARNING|DP=1|PP=0|TP=2|ip-26-0-174-36]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default4]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default6]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default1]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default0]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default3]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default3]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default2]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default2]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default5]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default5]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default6]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default6]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default4]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default4]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.) | |
| [default7]: return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default3]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default3]: warnings.warn( | |
| [default0]:07/03/2024 20:53:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 2219.02MiB. Peak allocated 7851.18MiB. Peak reserved: 8530.00MiB | |
| [default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default0]: warnings.warn( | |
| [default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default1]: warnings.warn( | |
| [default2]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default2]: warnings.warn( | |
| [default6]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default6]: warnings.warn( | |
| [default5]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default5]: warnings.warn( | |
| [default4]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default4]: warnings.warn( | |
| [default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions | |
| [default7]: warnings.warn( | |
| [default0]:07/03/2024 20:53:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 1 / 20 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 46K | tokens_per_sec: 91.3K | tokens_per_sec_per_gpu: 11.4K | global_batch_size: 1.02K | lm_loss: 11.4 | lr: 0.0001 | model_tflops_per_gpu: 104 | hardware_tflops_per_gpu: 104 | grad_norm: 20.6 | cuda_memory_allocated: 3.44G | cuda_max_memory_reserved: 9.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 65.8G | hd_free_memory_tb: 246G | |
| [default0]:07/03/2024 20:53:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.64MiB. Peak allocated 4601.55MiB. Peak reserved: 9060.00MiB | |
| [default0]:07/03/2024 20:53:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.77MiB. Peak allocated 8910.84MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 2 / 20 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 35.3K | tokens_per_sec: 119K | tokens_per_sec_per_gpu: 14.8K | global_batch_size: 1.02K | lm_loss: 11.4 | lr: 9.53e-05 | model_tflops_per_gpu: 135 | hardware_tflops_per_gpu: 135 | grad_norm: 20.7 | cuda_memory_allocated: 3.44G | cuda_max_memory_reserved: 9.92G | hd_total_memory_tb: 312G | hd_used_memory_tb: 65.8G | hd_free_memory_tb: 246G | |
| [default0]:07/03/2024 20:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.64MiB. Peak allocated 4601.68MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:54:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.77MiB. Peak allocated 8910.84MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:54:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 3 / 20 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 40.7K | tokens_per_sec: 103K | tokens_per_sec_per_gpu: 12.9K | global_batch_size: 1.02K | lm_loss: 11.6 | lr: 9.05e-05 | model_tflops_per_gpu: 117 | hardware_tflops_per_gpu: 117 | grad_norm: 195 | cuda_memory_allocated: 3.44G | cuda_max_memory_reserved: 9.92G | hd_total_memory_tb: 312G | hd_used_memory_tb: 65.8G | hd_free_memory_tb: 246G | |
| [default0]:07/03/2024 20:54:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.64MiB. Peak allocated 4601.68MiB. Peak reserved: 9456.00MiB | |
| [default0]:STAGE:2024-07-03 20:54:29 215133:215133 ActivityProfilerController.cpp:314] Completed Stage: Warm Up | |
| [default0]:07/03/2024 20:55:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.77MiB. Peak allocated 8910.84MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:55:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 4 / 20 | consumed_tokens: 16.8M | elapsed_time_per_iteration_ms: 54.1K | tokens_per_sec: 77.6K | tokens_per_sec_per_gpu: 9.7K | global_batch_size: 1.02K | lm_loss: 13.6 | lr: 8.58e-05 | model_tflops_per_gpu: 88 | hardware_tflops_per_gpu: 88 | grad_norm: 28.1 | cuda_memory_allocated: 3.44G | cuda_max_memory_reserved: 9.92G | hd_total_memory_tb: 312G | hd_used_memory_tb: 65.8G | hd_free_memory_tb: 246G | |
| [default0]:07/03/2024 20:55:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.64MiB. Peak allocated 4601.68MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:56:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 5 / 20 | consumed_tokens: 21M | elapsed_time_per_iteration_ms: 54.3K | tokens_per_sec: 77.3K | tokens_per_sec_per_gpu: 9.66K | global_batch_size: 1.02K | lm_loss: 12 | lr: 8.11e-05 | model_tflops_per_gpu: 87.6 | hardware_tflops_per_gpu: 87.6 | grad_norm: 48.8 | |
| [default0]:07/03/2024 20:56:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: Memory usage: 3277.64MiB. Peak allocated 8910.84MiB. Peak reserved: 9456.00MiB | |
| [default0]:07/03/2024 20:57:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-174-36]: iteration: 6 / 20 | consumed_tokens: 25.2M | elapsed_time_per_iteration_ms: 54.3K | tokens_per_sec: 77.3K | tokens_per_sec_per_gpu: 9.66K | global_batch_size: 1.02K | lm_loss: 10.9 | lr: 7.63e-05 | model_tflops_per_gpu: 87.6 | hardware_tflops_per_gpu: 87.6 | grad_norm: 19.7 | |
| [default0]:STAGE:2024-07-03 20:59:40 215133:215133 ActivityProfilerController.cpp:320] Completed Stage: Collection | |
| [default0]:STAGE:2024-07-03 20:59:55 215133:215133 ActivityProfilerController.cpp:324] Completed Stage: Post Processing | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:563] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600078 milliseconds before timing out. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:563] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600019 milliseconds before timing out. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:563] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600083 milliseconds before timing out. | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:1537] [PG 2 Rank 1] Timeout at NCCL work: 305728, last enqueued NCCL work: 305841, last completed NCCL work: 305727. | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:577] [Rank 1] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:583] [Rank 1] To avoid data inconsistency, we are taking the entire process down. | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:1414] [PG 2 Rank 1] Process group watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600078 milliseconds before timing out. | |
| [default1]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default1]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f02b2d66897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default1]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7f02b403fc62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7f02b4044a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f02b4045dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #4: <unknown function> + 0xd3e95 (0x7f02ffadee95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default1]:frame #5: <unknown function> + 0x8609 (0x7f0304b25609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default1]:frame #6: clone + 0x43 (0x7f03048f0353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default1]: | |
| [default1]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default1]: what(): [PG 2 Rank 1] Process group watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600078 milliseconds before timing out. | |
| [default1]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default1]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f02b2d66897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default1]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7f02b403fc62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7f02b4044a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f02b4045dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #4: <unknown function> + 0xd3e95 (0x7f02ffadee95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default1]:frame #5: <unknown function> + 0x8609 (0x7f0304b25609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default1]:frame #6: clone + 0x43 (0x7f03048f0353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default1]: | |
| [default1]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1418 (most recent call first): | |
| [default1]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f02b2d66897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default1]:frame #1: <unknown function> + 0xe32119 (0x7f02b3cc9119 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default1]:frame #2: <unknown function> + 0xd3e95 (0x7f02ffadee95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default1]:frame #3: <unknown function> + 0x8609 (0x7f0304b25609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default1]:frame #4: clone + 0x43 (0x7f03048f0353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default1]: | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:1537] [PG 2 Rank 2] Timeout at NCCL work: 305728, last enqueued NCCL work: 305841, last completed NCCL work: 305727. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:577] [Rank 2] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:583] [Rank 2] To avoid data inconsistency, we are taking the entire process down. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:1414] [PG 2 Rank 2] Process group watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600019 milliseconds before timing out. | |
| [default2]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdcf73d7897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default2]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7fdcf86b0c62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7fdcf86b5a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7fdcf86b6dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #4: <unknown function> + 0xd3e95 (0x7fdd4414fe95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #5: <unknown function> + 0x8609 (0x7fdd49196609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #6: clone + 0x43 (0x7fdd48f61353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [default2]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default2]: what(): [PG 2 Rank 2] Process group watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600019 milliseconds before timing out. | |
| [default2]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdcf73d7897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default2]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7fdcf86b0c62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7fdcf86b5a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7fdcf86b6dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #4: <unknown function> + 0xd3e95 (0x7fdd4414fe95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #5: <unknown function> + 0x8609 (0x7fdd49196609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #6: clone + 0x43 (0x7fdd48f61353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [default2]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1418 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdcf73d7897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default2]:frame #1: <unknown function> + 0xe32119 (0x7fdcf833a119 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: <unknown function> + 0xd3e95 (0x7fdd4414fe95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #3: <unknown function> + 0x8609 (0x7fdd49196609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #4: clone + 0x43 (0x7fdd48f61353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:1537] [PG 2 Rank 3] Timeout at NCCL work: 305728, last enqueued NCCL work: 305841, last completed NCCL work: 305727. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:577] [Rank 3] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:583] [Rank 3] To avoid data inconsistency, we are taking the entire process down. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:1414] [PG 2 Rank 3] Process group watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600083 milliseconds before timing out. | |
| [default3]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1ee4da9897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default3]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7f1ee6082c62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7f1ee6087a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f1ee6088dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #4: <unknown function> + 0xd3e95 (0x7f1f31b21e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #5: <unknown function> + 0x8609 (0x7f1f36b68609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #6: clone + 0x43 (0x7f1f36933353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| [default3]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default3]: what(): [PG 2 Rank 3] Process group watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=305728, OpType=_REDUCE_SCATTER_BASE, NumelIn=16777216, NumelOut=4194304, Timeout(ms)=600000) ran for 600083 milliseconds before timing out. | |
| [default3]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:565 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1ee4da9897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default3]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1d2 (0x7f1ee6082c62 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0x1a0 (0x7f1ee6087a80 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f1ee6088dcc in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #4: <unknown function> + 0xd3e95 (0x7f1f31b21e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #5: <unknown function> + 0x8609 (0x7f1f36b68609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #6: clone + 0x43 (0x7f1f36933353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| [default3]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1418 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1ee4da9897 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default3]:frame #1: <unknown function> + 0xe32119 (0x7f1ee5d0c119 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: <unknown function> + 0xd3e95 (0x7f1f31b21e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #3: <unknown function> + 0x8609 (0x7f1f36b68609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #4: clone + 0x43 (0x7f1f36933353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| W0703 21:07:45.706000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:851] Sending process 215133 closing signal SIGTERM | |
| W0703 21:07:45.706000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:851] Sending process 215137 closing signal SIGTERM | |
| W0703 21:07:45.706000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:851] Sending process 215138 closing signal SIGTERM | |
| W0703 21:07:45.707000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:851] Sending process 215139 closing signal SIGTERM | |
| W0703 21:07:45.708000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:851] Sending process 215140 closing signal SIGTERM | |
| E0703 21:07:51.736000 139757201868608 torch/distributed/elastic/multiprocessing/api.py:826] failed (exitcode: -6) local_rank: 1 (pid: 215134) of binary: /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/python3.10 | |
| Traceback (most recent call last): | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/torchrun", line 8, in <module> | |
| sys.exit(main()) | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 347, in wrapper | |
| return f(*args, **kwargs) | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/run.py", line 879, in main | |
| run(args) | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/run.py", line 870, in run | |
| elastic_launch( | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 132, in __call__ | |
| return launch_agent(self._config, self._entrypoint, list(args)) | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 263, in launch_agent | |
| raise ChildFailedError( | |
| torch.distributed.elastic.multiprocessing.errors.ChildFailedError: | |
| ============================================================ | |
| /fsx/ferdinandmom/ferdinand-hf/bench_cluster/nanotron/run_train.py FAILED | |
| ------------------------------------------------------------ | |
| Failures: | |
| [1]: | |
| time : 2024-07-03_21:07:45 | |
| host : ip-26-0-174-36.ec2.internal | |
| rank : 2 (local_rank: 2) | |
| exitcode : -6 (pid: 215135) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 215135 | |
| [2]: | |
| time : 2024-07-03_21:07:45 | |
| host : ip-26-0-174-36.ec2.internal | |
| rank : 3 (local_rank: 3) | |
| exitcode : -6 (pid: 215136) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 215136 | |
| ------------------------------------------------------------ | |
| Root Cause (first observed failure): | |
| [0]: | |
| time : 2024-07-03_21:07:45 | |
| host : ip-26-0-174-36.ec2.internal | |
| rank : 1 (local_rank: 1) | |
| exitcode : -6 (pid: 215134) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 215134 | |
| ============================================================ | |
| srun: error: ip-26-0-174-36: task 0: Exited with exit code 1 | |
| Consider using `hf_transfer` for faster uploads. This solution comes with some limitations. See https://huggingface.co/docs/huggingface_hub/hf_transfer for more details. | |