| ======================== | |
| START TIME: Fri Jul 5 15:14:30 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 | |
| Your branch is up to date with 'origin/bench_cluster'. | |
| Job status: RUNNING | |
| [2024-07-05 15:14:38,465] torch.distributed.run: [WARNING] | |
| [2024-07-05 15:14:38,465] torch.distributed.run: [WARNING] ***************************************** | |
| [2024-07-05 15:14:38,465] torch.distributed.run: [WARNING] 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. | |
| [2024-07-05 15:14:38,465] torch.distributed.run: [WARNING] ***************************************** | |
| [default0]:07/05/2024 15:14:56 [WARNING|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Vocab Size Padding] Padded vocab (size: 50257) with 3 dummy tokens (new size: 50260) | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config: | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config(general=GeneralArgs(project='bench_cluster', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: run='%date_%jobid', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: step=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: consumed_train_samples=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: benchmark_csv_path=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ignore_sanity_checks=True), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: parallelism=ParallelismArgs(dp=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp=2, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp=4, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp_engine=<nanotron.parallel.pipeline_parallel.engine.OneForwardOneBackwardPipelineEngine object at 0x7ff8bd3ec8b0>, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_mode=<TensorParallelLinearMode.REDUCE_SCATTER: 2>, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_linear_async_communication=False, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: expert_parallel_size=1), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=2, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=4096, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=4096, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=24, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=32, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=50260), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: init_method=RandomInit(std=0.025), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dtype=torch.bfloat16, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: make_vocab_size_divisible_by=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ddp_bucket_cap_mb=25), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer=TokenizerArgs(tokenizer_name_or_path='openai-community/gpt2', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_revision=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_max_length=None), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('/dev/null'), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoint_interval=100000, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_initial_state=False, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: resume_checkpoint_path=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints_path_is_shared_file_system=False), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: logging=LoggingArgs(log_level='info', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: log_level_replica='info', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration_step_info_interval=1), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokens=TokensArgs(sequence_length=4096, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: train_steps=20, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: micro_batch_size=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: batch_accumulation_per_replica=1024, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: val_check_interval=-1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_val_batches=0, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_test_batches=0), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta1=0.9, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta2=0.95, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: torch_adam_is_fused=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: name='adamW'), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: zero_stage=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: weight_decay=0.01, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: clip_grad=1.0, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: accumulate_grad_in_fp32=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0001, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_steps=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_style='linear', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_style='linear', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_steps=19, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_starting_step=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: min_decay_lr=1e-05)), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data_stages=[DatasetStageArgs(name='Training Stage', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: start_training_step=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data=DataArgs(dataset=PretrainDatasetsArgs(hf_dataset_or_datasets='roneneldan/TinyStories', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hf_dataset_splits='train', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hf_dataset_config_name=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dataset_processing_num_proc_per_process=64, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dataset_overwrite_cache=False, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: text_column_name='text'), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_loading_workers=0))], | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: profiler=ProfilerArgs(profiler_export_path=PosixPath('/fsx/ferdinandmom/ferdinand-hf/bench_cluster/results/llama-1B/8_GPUS/dp-1_tp-4_pp-2_mbz-1')), | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lighteval=None) | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Model Config: | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: LlamaConfig(bos_token_id=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=2, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=4096, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=4096, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=24, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=32, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=50260) | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Building model.. | |
| [default0]:07/05/2024 15:14:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Setting PP block ranks... | |
| [default4]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: Local number of parameters: 131M (249.16MiB) | |
| [default3]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: Local number of parameters: 173M (329.19MiB) | |
| [default4]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 260.10MiB. Peak allocated: 262.13MiB Peak reserved: 264.00MiB | |
| [default4]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: No checkpoint path provided. | |
| [default3]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: [After model building] Memory usage: 344.13MiB. Peak allocated: 346.16MiB Peak reserved: 348.00MiB | |
| [default3]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: No checkpoint path provided. | |
| [default0]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Total number of parameters: 1.21G (2313.42MiB) | |
| [default0]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Local number of parameters: 173M (329.19MiB) | |
| [default0]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 344.13MiB. Peak allocated: 346.16MiB Peak reserved: 348.00MiB | |
| [default1]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Local number of parameters: 173M (329.19MiB) | |
| [default1]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: [After model building] Memory usage: 344.13MiB. Peak allocated: 346.16MiB Peak reserved: 348.00MiB | |
| [default1]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: No checkpoint path provided. | |
| [default2]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: Local number of parameters: 173M (329.19MiB) | |
| [default2]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: [After model building] Memory usage: 344.13MiB. Peak allocated: 346.16MiB Peak reserved: 348.00MiB | |
| [default2]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: No checkpoint path provided. | |
| [default0]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided. | |
| [default0]:07/05/2024 15:15:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Parametrizing model parameters using StandardParametrizator | |
| [default6]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=2|ip-26-0-160-225]: Local number of parameters: 131M (249.16MiB) | |
| [default6]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=2|ip-26-0-160-225]: [After model building] Memory usage: 260.10MiB. Peak allocated: 262.13MiB Peak reserved: 264.00MiB | |
| [default6]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=2|ip-26-0-160-225]: No checkpoint path provided. | |
| [default5]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=1|ip-26-0-160-225]: Local number of parameters: 131M (249.16MiB) | |
| [default5]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=1|ip-26-0-160-225]: [After model building] Memory usage: 260.10MiB. Peak allocated: 262.13MiB Peak reserved: 264.00MiB | |
| [default7]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=3|ip-26-0-160-225]: Local number of parameters: 131M (249.16MiB) | |
| [default7]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=3|ip-26-0-160-225]: [After model building] Memory usage: 260.10MiB. Peak allocated: 262.13MiB Peak reserved: 264.00MiB | |
| [default5]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=1|ip-26-0-160-225]: No checkpoint path provided. | |
| [default7]:07/05/2024 15:15:11 [INFO|DP=0|PP=1|TP=3|ip-26-0-160-225]: No checkpoint path provided. | |
| [default0]:07/05/2024 15:15:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Optimizer Building] Using LearningRateForSP as learning rate | |
| [default0]:07/05/2024 15:15:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] Size of optimizer params per rank: | |
| [default0]:07/05/2024 15:15:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 0 has 173M out of 173M (100.00%) params' optimizer states | |
| [default0]:07/05/2024 15:15:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] Stage Training Stage has 19 remaining training steps and has consumed 0 samples | |
| [default0]:07/05/2024 15:15:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Using `datasets` library | |
| [default0]:07/05/2024 15:15:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Loading tokenizer from openai-community/gpt2 and transformers/hf_hub versions ('4.41.2', '0.23.4') | |
| [default0]:07/05/2024 15:15:14 [WARNING|DP=0|PP=0|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default0]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default0]:07/05/2024 15:15:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] There are 1 training stages | |
| [default0]:07/05/2024 15:15:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Stage Training Stage] start from step 1 | |
| [default0]:07/05/2024 15:15:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | |
| [default0]:07/05/2024 15:15:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Start training] datetime: 2024-07-05 15:15:16.488343 | mbs: 1 | grad_accum: 1024 | global_batch_size: 1024 | sequence_length: 4096 | train_steps: 20 | start_iteration_step: 0 | consumed_train_samples: 0 | |
| [default2]:07/05/2024 15:15:16 [WARNING|DP=0|PP=0|TP=2|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default1]:07/05/2024 15:15:16 [WARNING|DP=0|PP=0|TP=1|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default2]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default1]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default5]:07/05/2024 15:15:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default5]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default6]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default3]:07/05/2024 15:15:16 [WARNING|DP=0|PP=0|TP=3|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default7]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default4]:07/05/2024 15:15:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default6]:07/05/2024 15:15:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default4]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default3]:Repo card metadata block was not found. Setting CardData to empty. | |
| [default7]:07/05/2024 15:15:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty. | |
| [default0]:07/05/2024 15:15:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Resuming training from stage Training Stage, it has trained for 0 samples and has 19 remaining train steps | |
| [default0]:07/05/2024 15:15:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 1660.89MiB. Peak allocated 1660.89MiB. Peak reserved: 1668.00MiB | |
| [default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/__init__.py:266: 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]: 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/__init__.py:266: 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]: 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/__init__.py:266: 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]: 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/__init__.py:266: 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]: 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/__init__.py:266: 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]: 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/__init__.py:266: 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]: 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/__init__.py:266: 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]: Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/__init__.py:266: 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]: Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
| [default0]:07/05/2024 15:16:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 1731.93MiB. Peak allocated 4645.52MiB. Peak reserved: 4720.00MiB | |
| [default0]:07/05/2024 15:17:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 3048.73MiB. Peak reserved: 4720.00MiB | |
| [default4]:07/05/2024 15:17:01 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 1 / 20 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 104K | tokens_per_sec: 40.4K | tokens_per_sec_per_gpu: 5.05K | global_batch_size: 1.02K | lm_loss: 11.1 | lr: 0.0001 | model_tflops_per_gpu: 45.8 | hardware_tflops_per_gpu: 45.8 | grad_norm: 15 | cuda_memory_allocated: 2.44G | cuda_max_memory_reserved: 3.18G | hd_total_memory_tb: 312G | hd_used_memory_tb: 66.6G | hd_free_memory_tb: 246G | |
| [default4]:07/05/2024 15:18:26 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 2 / 20 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 84.8K | tokens_per_sec: 49.5K | tokens_per_sec_per_gpu: 6.18K | global_batch_size: 1.02K | lm_loss: 11.1 | lr: 9.53e-05 | model_tflops_per_gpu: 56.1 | hardware_tflops_per_gpu: 56.1 | grad_norm: 15.1 | cuda_memory_allocated: 2.44G | cuda_max_memory_reserved: 4.03G | hd_total_memory_tb: 312G | hd_used_memory_tb: 66.6G | hd_free_memory_tb: 246G | |
| [default0]:07/05/2024 15:18:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 5772.57MiB. Peak reserved: 5918.00MiB | |
| [default0]:07/05/2024 15:18:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 3048.78MiB. Peak reserved: 5918.00MiB | |
| [default0]:07/05/2024 15:20:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 5772.57MiB. Peak reserved: 5918.00MiB | |
| [default4]:07/05/2024 15:20:02 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 3 / 20 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 96.1K | tokens_per_sec: 43.6K | tokens_per_sec_per_gpu: 5.46K | global_batch_size: 1.02K | lm_loss: 11.4 | lr: 9.05e-05 | model_tflops_per_gpu: 49.5 | hardware_tflops_per_gpu: 49.5 | grad_norm: 106 | cuda_memory_allocated: 2.44G | cuda_max_memory_reserved: 4.03G | hd_total_memory_tb: 312G | hd_used_memory_tb: 66.6G | hd_free_memory_tb: 246G | |
| [default0]:STAGE:2024-07-05 15:20:02 512161:512161 ActivityProfilerController.cpp:314] Completed Stage: Warm Up | |
| [default0]:07/05/2024 15:20:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 3048.78MiB. Peak reserved: 5918.00MiB | |
| [default4]:07/05/2024 15:22:11 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 4 / 20 | consumed_tokens: 16.8M | elapsed_time_per_iteration_ms: 129K | tokens_per_sec: 32.6K | tokens_per_sec_per_gpu: 4.07K | global_batch_size: 1.02K | lm_loss: 11.7 | lr: 8.58e-05 | model_tflops_per_gpu: 37 | hardware_tflops_per_gpu: 37 | grad_norm: 24.5 | cuda_memory_allocated: 2.44G | cuda_max_memory_reserved: 4.03G | hd_total_memory_tb: 312G | hd_used_memory_tb: 66.6G | hd_free_memory_tb: 246G | |
| [default0]:07/05/2024 15:22:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 5772.57MiB. Peak reserved: 5918.00MiB | |
| [default0]:07/05/2024 15:22:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 3048.78MiB. Peak reserved: 5918.00MiB | |
| [default0]:07/05/2024 15:24:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3048.73MiB. Peak allocated 5772.57MiB. Peak reserved: 5918.00MiB | |
| [default4]:07/05/2024 15:24:19 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 5 / 20 | consumed_tokens: 21M | elapsed_time_per_iteration_ms: 129K | tokens_per_sec: 32.6K | tokens_per_sec_per_gpu: 4.07K | global_batch_size: 1.02K | lm_loss: 10 | lr: 8.11e-05 | model_tflops_per_gpu: 36.9 | hardware_tflops_per_gpu: 36.9 | grad_norm: 11 | |
| [default4]:07/05/2024 15:26:29 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: iteration: 6 / 20 | consumed_tokens: 25.2M | elapsed_time_per_iteration_ms: 130K | tokens_per_sec: 32.4K | tokens_per_sec_per_gpu: 4.05K | global_batch_size: 1.02K | lm_loss: 9.46 | lr: 7.63e-05 | model_tflops_per_gpu: 36.7 | hardware_tflops_per_gpu: 36.7 | grad_norm: 7.21 | |
| [default0]:STAGE:2024-07-05 15:32:33 512161:512161 ActivityProfilerController.cpp:320] Completed Stage: Collection | |
| [default0]:STAGE:2024-07-05 15:33:08 512161:512161 ActivityProfilerController.cpp:324] Completed Stage: Post Processing | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800138 milliseconds before timing out. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:523] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800149 milliseconds before timing out. | |
| [default7]:[rank7]:[E ProcessGroupNCCL.cpp:523] [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800224 milliseconds before timing out. | |
| [default4]:[rank4]:[E ProcessGroupNCCL.cpp:523] [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800228 milliseconds before timing out. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:523] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800167 milliseconds before timing out. | |
| [default6]:[rank6]:[E ProcessGroupNCCL.cpp:523] [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800240 milliseconds before timing out. | |
| [default5]:[rank5]:[E ProcessGroupNCCL.cpp:523] [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800238 milliseconds before timing out. | |
| [default5]:[rank5]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=55306, OpType=RECV, NumelIn=7, NumelOut=7, Timeout(ms)=600000) ran for 600282 milliseconds before timing out. | |
| [default5]:[rank5]:[E ProcessGroupNCCL.cpp:537] 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. | |
| [default5]:[rank5]:[E ProcessGroupNCCL.cpp:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default5]:[rank5]:[E ProcessGroupNCCL.cpp:1182] [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800238 milliseconds before timing out. | |
| [default5]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default5]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1415193d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default5]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7f141633b6e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f141633ec3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f141633f839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #4: <unknown function> + 0xd3e95 (0x7f1460043e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default5]:frame #5: <unknown function> + 0x8609 (0x7f146514b609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default5]:frame #6: clone + 0x43 (0x7f1464f16353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default5]: | |
| [default5]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default5]: what(): [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800238 milliseconds before timing out. | |
| [default5]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default5]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1415193d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default5]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7f141633b6e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f141633ec3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f141633f839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #4: <unknown function> + 0xd3e95 (0x7f1460043e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default5]:frame #5: <unknown function> + 0x8609 (0x7f146514b609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default5]:frame #6: clone + 0x43 (0x7f1464f16353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default5]: | |
| [default5]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default5]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f1415193d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default5]:frame #1: <unknown function> + 0xdf6b11 (0x7f1416095b11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default5]:frame #2: <unknown function> + 0xd3e95 (0x7f1460043e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default5]:frame #3: <unknown function> + 0x8609 (0x7f146514b609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default5]:frame #4: clone + 0x43 (0x7f1464f16353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default5]: | |
| [default6]:[rank6]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=55306, OpType=RECV, NumelIn=7, NumelOut=7, Timeout(ms)=600000) ran for 600310 milliseconds before timing out. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:523] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=700453, OpType=_REDUCE_SCATTER_BASE, NumelIn=8388608, NumelOut=2097152, Timeout(ms)=600000) ran for 600220 milliseconds before timing out. | |
| [default4]:[rank4]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=55306, OpType=RECV, NumelIn=7, NumelOut=7, Timeout(ms)=600000) ran for 600270 milliseconds before timing out. | |
| [default4]:[rank4]:[E ProcessGroupNCCL.cpp:537] 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. | |
| [default4]:[rank4]:[E ProcessGroupNCCL.cpp:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default4]:[rank4]:[E ProcessGroupNCCL.cpp:1182] [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800228 milliseconds before timing out. | |
| [default4]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default4]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7ff36a0b0d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default4]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7ff36b2586e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7ff36b25bc3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7ff36b25c839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #4: <unknown function> + 0xd3e95 (0x7ff3b4f60e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default4]:frame #5: <unknown function> + 0x8609 (0x7ff3ba068609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default4]:frame #6: clone + 0x43 (0x7ff3b9e33353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default4]: | |
| [default4]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default4]: what(): [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800228 milliseconds before timing out. | |
| [default4]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default4]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7ff36a0b0d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default4]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7ff36b2586e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7ff36b25bc3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7ff36b25c839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #4: <unknown function> + 0xd3e95 (0x7ff3b4f60e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default4]:frame #5: <unknown function> + 0x8609 (0x7ff3ba068609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default4]:frame #6: clone + 0x43 (0x7ff3b9e33353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default4]: | |
| [default4]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default4]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7ff36a0b0d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default4]:frame #1: <unknown function> + 0xdf6b11 (0x7ff36afb2b11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default4]:frame #2: <unknown function> + 0xd3e95 (0x7ff3b4f60e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default4]:frame #3: <unknown function> + 0x8609 (0x7ff3ba068609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default4]:frame #4: clone + 0x43 (0x7ff3b9e33353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default4]: | |
| [default7]:[rank7]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=55306, OpType=RECV, NumelIn=7, NumelOut=7, Timeout(ms)=600000) ran for 600375 milliseconds before timing out. | |
| [default6]:[rank6]:[E ProcessGroupNCCL.cpp:537] 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. | |
| [default6]:[rank6]:[E ProcessGroupNCCL.cpp:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default6]:[rank6]:[E ProcessGroupNCCL.cpp:1182] [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800240 milliseconds before timing out. | |
| [default6]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default6]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fb5dda2cd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default6]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7fb5debd46e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7fb5debd7c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7fb5debd8839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #4: <unknown function> + 0xd3e95 (0x7fb6288dce95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default6]:frame #5: <unknown function> + 0x8609 (0x7fb62d9e4609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default6]:frame #6: clone + 0x43 (0x7fb62d7af353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default6]: | |
| [default6]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default6]: what(): [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800240 milliseconds before timing out. | |
| [default6]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default6]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fb5dda2cd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default6]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7fb5debd46e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7fb5debd7c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7fb5debd8839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #4: <unknown function> + 0xd3e95 (0x7fb6288dce95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default6]:frame #5: <unknown function> + 0x8609 (0x7fb62d9e4609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default6]:frame #6: clone + 0x43 (0x7fb62d7af353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default6]: | |
| [default6]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default6]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fb5dda2cd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default6]:frame #1: <unknown function> + 0xdf6b11 (0x7fb5de92eb11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default6]:frame #2: <unknown function> + 0xd3e95 (0x7fb6288dce95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default6]:frame #3: <unknown function> + 0x8609 (0x7fb62d9e4609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default6]:frame #4: clone + 0x43 (0x7fb62d7af353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default6]: | |
| [default1]:[rank1]:[E ProcessGroupNCCL.cpp:523] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=700453, OpType=_REDUCE_SCATTER_BASE, NumelIn=8388608, NumelOut=2097152, Timeout(ms)=600000) ran for 600440 milliseconds before timing out. | |
| [default7]:[rank7]:[E ProcessGroupNCCL.cpp:537] 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. | |
| [default7]:[rank7]:[E ProcessGroupNCCL.cpp:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default7]:[rank7]:[E ProcessGroupNCCL.cpp:1182] [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800224 milliseconds before timing out. | |
| [default7]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default7]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f67217dbd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default7]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7f67229836e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f6722986c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f6722987839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #4: <unknown function> + 0xd3e95 (0x7f676c68be95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default7]:frame #5: <unknown function> + 0x8609 (0x7f6771793609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default7]:frame #6: clone + 0x43 (0x7f677155e353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default7]: | |
| [default7]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default7]: what(): [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800224 milliseconds before timing out. | |
| [default7]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default7]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f67217dbd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default7]:frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x1e6 (0x7f67229836e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f6722986c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f6722987839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #4: <unknown function> + 0xd3e95 (0x7f676c68be95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default7]:frame #5: <unknown function> + 0x8609 (0x7f6771793609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default7]:frame #6: clone + 0x43 (0x7f677155e353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default7]: | |
| [default7]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default7]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f67217dbd87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default7]:frame #1: <unknown function> + 0xdf6b11 (0x7f67226ddb11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default7]:frame #2: <unknown function> + 0xd3e95 (0x7f676c68be95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default7]:frame #3: <unknown function> + 0x8609 (0x7f6771793609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default7]:frame #4: clone + 0x43 (0x7f677155e353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default7]: | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:523] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=700453, OpType=_REDUCE_SCATTER_BASE, NumelIn=8388608, NumelOut=2097152, Timeout(ms)=600000) ran for 600790 milliseconds before timing out. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:537] 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:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default3]:[rank3]:[E ProcessGroupNCCL.cpp:1182] [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800167 milliseconds before timing out. | |
| [default3]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f2d41b88d87 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> > >) + 0x1e6 (0x7f2d42d306e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f2d42d33c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f2d42d34839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #4: <unknown function> + 0xd3e95 (0x7f2d8ca38e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #5: <unknown function> + 0x8609 (0x7f2d91b40609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #6: clone + 0x43 (0x7f2d9190b353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| [default3]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default3]: what(): [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800167 milliseconds before timing out. | |
| [default3]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f2d41b88d87 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> > >) + 0x1e6 (0x7f2d42d306e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7f2d42d33c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7f2d42d34839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #4: <unknown function> + 0xd3e95 (0x7f2d8ca38e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #5: <unknown function> + 0x8609 (0x7f2d91b40609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #6: clone + 0x43 (0x7f2d9190b353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| [default3]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default3]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f2d41b88d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default3]:frame #1: <unknown function> + 0xdf6b11 (0x7f2d42a8ab11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default3]:frame #2: <unknown function> + 0xd3e95 (0x7f2d8ca38e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default3]:frame #3: <unknown function> + 0x8609 (0x7f2d91b40609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default3]:frame #4: clone + 0x43 (0x7f2d9190b353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default3]: | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:537] 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:543] To avoid data inconsistency, we are taking the entire process down. | |
| [default2]:[rank2]:[E ProcessGroupNCCL.cpp:1182] [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800149 milliseconds before timing out. | |
| [default2]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdb4dc59d87 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> > >) + 0x1e6 (0x7fdb4ee016e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7fdb4ee04c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7fdb4ee05839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #4: <unknown function> + 0xd3e95 (0x7fdb98b09e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #5: <unknown function> + 0x8609 (0x7fdb9dc11609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #6: clone + 0x43 (0x7fdb9d9dc353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [default2]:terminate called after throwing an instance of 'c10::DistBackendError' | |
| [default2]: what(): [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=26, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=1800000) ran for 1800149 milliseconds before timing out. | |
| [default2]:Exception raised from checkTimeout at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:525 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdb4dc59d87 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> > >) + 0x1e6 (0x7fdb4ee016e6 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x19d (0x7fdb4ee04c3d in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x119 (0x7fdb4ee05839 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #4: <unknown function> + 0xd3e95 (0x7fdb98b09e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #5: <unknown function> + 0x8609 (0x7fdb9dc11609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #6: clone + 0x43 (0x7fdb9d9dc353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [default2]:Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1186 (most recent call first): | |
| [default2]:frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fdb4dc59d87 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libc10.so) | |
| [default2]:frame #1: <unknown function> + 0xdf6b11 (0x7fdb4eb5bb11 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) | |
| [default2]:frame #2: <unknown function> + 0xd3e95 (0x7fdb98b09e95 in /fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/bin/../lib/libstdc++.so.6) | |
| [default2]:frame #3: <unknown function> + 0x8609 (0x7fdb9dc11609 in /lib/x86_64-linux-gnu/libpthread.so.0) | |
| [default2]:frame #4: clone + 0x43 (0x7fdb9d9dc353 in /lib/x86_64-linux-gnu/libc.so.6) | |
| [default2]: | |
| [2024-07-05 16:06:31,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 512161 closing signal SIGTERM | |
| [2024-07-05 16:06:31,741] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 512162 closing signal SIGTERM | |
| [2024-07-05 16:06:31,741] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 512163 closing signal SIGTERM | |
| [2024-07-05 16:06:31,742] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 512164 closing signal SIGTERM | |
| [2024-07-05 16:06:45,961] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: -6) local_rank: 4 (pid: 512165) 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 812, in main | |
| run(args) | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/run.py", line 803, in run | |
| elastic_launch( | |
| File "/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 135, 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 268, 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 : 2024-07-05_16:06:31 | |
| host : ip-26-0-160-225.ec2.internal | |
| rank : 5 (local_rank: 5) | |
| exitcode : -6 (pid: 512166) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 512166 | |
| [2]: | |
| time : 2024-07-05_16:06:31 | |
| host : ip-26-0-160-225.ec2.internal | |
| rank : 6 (local_rank: 6) | |
| exitcode : -6 (pid: 512167) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 512167 | |
| [3]: | |
| time : 2024-07-05_16:06:31 | |
| host : ip-26-0-160-225.ec2.internal | |
| rank : 7 (local_rank: 7) | |
| exitcode : -6 (pid: 512168) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 512168 | |
| ------------------------------------------------------------ | |
| Root Cause (first observed failure): | |
| [0]: | |
| time : 2024-07-05_16:06:31 | |
| host : ip-26-0-160-225.ec2.internal | |
| rank : 4 (local_rank: 4) | |
| exitcode : -6 (pid: 512165) | |
| error_file: <N/A> | |
| traceback : Signal 6 (SIGABRT) received by PID 512165 | |
| ============================================================ | |
| srun: error: ip-26-0-160-225: 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. | |