======================== START TIME: Tue Jul 2 18:44:38 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 W0702 18:44:44.501000 140679052117824 torch/distributed/run.py:757] W0702 18:44:44.501000 140679052117824 torch/distributed/run.py:757] ***************************************** W0702 18:44:44.501000 140679052117824 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. W0702 18:44:44.501000 140679052117824 torch/distributed/run.py:757] ***************************************** W0702 18:44:49.409000 140253572294464 torch/distributed/run.py:757] W0702 18:44:49.409000 140253572294464 torch/distributed/run.py:757] ***************************************** W0702 18:44:49.409000 140253572294464 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. W0702 18:44:49.409000 140253572294464 torch/distributed/run.py:757] ***************************************** [default0]:07/02/2024 18:45:11 [WARNING|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Vocab Size Padding] Padded vocab (size: 50257) with 7 dummy tokens (new size: 50264) [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Config: [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Config(general=GeneralArgs(project='bench_cluster', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: run='%date_%jobid', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: seed=42, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: step=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: consumed_train_samples=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: benchmark_csv_path=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: ignore_sanity_checks=True), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: parallelism=ParallelismArgs(dp=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pp=2, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tp=8, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pp_engine=, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tp_mode=, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tp_linear_async_communication=False, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: expert_parallel_size=1), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: eos_token_id=2, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hidden_act='silu', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hidden_size=2048, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: initializer_range=0.02, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: intermediate_size=4096, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: is_llama_config=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: max_position_embeddings=4096, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_attention_heads=32, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_hidden_layers=24, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_key_value_heads=32, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pad_token_id=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pretraining_tp=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rms_norm_eps=1e-05, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rope_scaling=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rope_theta=10000.0, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tie_word_embeddings=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: use_cache=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: vocab_size=50264), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: init_method=RandomInit(std=0.025), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: dtype=torch.bfloat16, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: make_vocab_size_divisible_by=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: ddp_bucket_cap_mb=25), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tokenizer=TokenizerArgs(tokenizer_name_or_path='openai-community/gpt2', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tokenizer_revision=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tokenizer_max_length=None), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: checkpoints=CheckpointsArgs(checkpoints_path=Path('/dev/null'), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: checkpoint_interval=100000, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: save_initial_state=False, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: resume_checkpoint_path=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: checkpoints_path_is_shared_file_system=False), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: logging=LoggingArgs(log_level='info', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: log_level_replica='info', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: iteration_step_info_interval=1), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tokens=TokensArgs(sequence_length=4096, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: train_steps=20, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: micro_batch_size=16, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: batch_accumulation_per_replica=64, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: val_check_interval=-1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: limit_val_batches=0, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: limit_test_batches=0), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: adam_beta1=0.9, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: adam_beta2=0.95, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: torch_adam_is_fused=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: name='adamW'), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: zero_stage=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: weight_decay=0.01, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: clip_grad=1.0, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: accumulate_grad_in_fp32=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0001, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lr_warmup_steps=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lr_warmup_style='linear', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lr_decay_style='linear', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lr_decay_steps=19, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lr_decay_starting_step=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: min_decay_lr=1e-05)), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: data_stages=[DatasetStageArgs(name='Training Stage', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: start_training_step=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: data=DataArgs(dataset=PretrainDatasetsArgs(hf_dataset_or_datasets='roneneldan/TinyStories', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hf_dataset_splits='train', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hf_dataset_config_name=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: dataset_processing_num_proc_per_process=64, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: dataset_overwrite_cache=False, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: text_column_name='text'), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: seed=42, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_loading_workers=32))], [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: profiler=ProfilerArgs(profiler_export_path=Path('/fsx/ferdinandmom/ferdinand-hf/bench_cluster/results/llama-1B/16_GPUS/dp-1_tp-8_pp-2_mbz-16')), [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: lighteval=None) [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Model Config: [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: LlamaConfig(bos_token_id=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: eos_token_id=2, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hidden_act='silu', [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: hidden_size=2048, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: initializer_range=0.02, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: intermediate_size=4096, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: is_llama_config=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: max_position_embeddings=4096, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_attention_heads=32, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_hidden_layers=24, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: num_key_value_heads=32, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pad_token_id=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: pretraining_tp=1, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rms_norm_eps=1e-05, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rope_scaling=None, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: rope_theta=10000.0, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: tie_word_embeddings=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: use_cache=True, [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: vocab_size=50264) [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Building model.. [default0]:07/02/2024 18:45:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Setting PP block ranks... [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=5|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=5|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=5|ip-26-0-170-160]: No checkpoint path provided. [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: No checkpoint path provided. [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=3|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=3|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=3|ip-26-0-170-160]: No checkpoint path provided. [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-160]: No checkpoint path provided. [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=7|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=7|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=7|ip-26-0-170-160]: No checkpoint path provided. [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=4|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=4|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=4|ip-26-0-165-24]: No checkpoint path provided. [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=2|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=2|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=2|ip-26-0-170-160]: No checkpoint path provided. [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=6|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=6|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=6|ip-26-0-170-160]: No checkpoint path provided. [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=4|ip-26-0-170-160]: Local number of parameters: 65.3M (124.62MiB) [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=4|ip-26-0-170-160]: [After model building] Memory usage: 135.64MiB. Peak allocated: 137.67MiB Peak reserved: 150.00MiB [default4]:07/02/2024 18:45:28 [INFO|DP=0|PP=1|TP=4|ip-26-0-170-160]: No checkpoint path provided. [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=6|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=6|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default6]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=6|ip-26-0-165-24]: No checkpoint path provided. [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default2]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-24]: No checkpoint path provided. [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=7|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=7|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default1]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-24]: No checkpoint path provided. [default7]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=7|ip-26-0-165-24]: No checkpoint path provided. [default3]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-24]: No checkpoint path provided. [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Total number of parameters: 1.21G (2314.22MiB) [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: No checkpoint path provided. [default0]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Parametrizing model parameters using StandardParametrizator [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=5|ip-26-0-165-24]: Local number of parameters: 86.3M (164.65MiB) [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=5|ip-26-0-165-24]: [After model building] Memory usage: 179.67MiB. Peak allocated: 181.70MiB Peak reserved: 198.00MiB [default5]:07/02/2024 18:45:28 [INFO|DP=0|PP=0|TP=5|ip-26-0-165-24]: No checkpoint path provided. [default0]:07/02/2024 18:45:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Optimizer Building] Using LearningRateForSP as learning rate [default0]:07/02/2024 18:45:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [ZeRO sharding] Size of optimizer params per rank: [default0]:07/02/2024 18:45:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [ZeRO sharding] DP Rank 0 has 86.3M out of 86.3M (100.00%) params' optimizer states [default0]:07/02/2024 18:45:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Training Plan] Stage Training Stage has 19 remaining training steps and has consumed 0 samples [default0]:07/02/2024 18:45:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Using `datasets` library [default0]:07/02/2024 18:45:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: 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]:07/02/2024 18:45:31 [WARNING|DP=0|PP=0|TP=0|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Training Plan] There are 1 training stages [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Stage Training Stage] start from step 1 [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: [Start training] datetime: 2024-07-02 18:45:33.417001 | mbs: 16 | grad_accum: 64 | global_batch_size: 1024 | sequence_length: 4096 | train_steps: 20 | start_iteration_step: 0 | consumed_train_samples: 0 [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Resuming training from stage Training Stage, it has trained for 0 samples and has 19 remaining train steps [default0]:07/02/2024 18:45:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 839.67MiB. Peak allocated 839.67MiB. Peak reserved: 858.00MiB [default5]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=5|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default0]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default3]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=3|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default1]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=1|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default2]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=2|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default4]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=4|ip-26-0-165-24]: 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. [default0]:Repo card metadata block was not found. Setting CardData to empty. [default3]:Repo card metadata block was not found. Setting CardData to empty. [default1]:Repo card metadata block was not found. Setting CardData to empty. [default5]:Repo card metadata block was not found. Setting CardData to empty. [default2]:Repo card metadata block was not found. Setting CardData to empty. [default3]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=3|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default2]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=2|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default7]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=7|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default1]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=1|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default7]:Repo card metadata block was not found. Setting CardData to empty. [default5]:Repo card metadata block was not found. Setting CardData to empty. [default3]:Repo card metadata block was not found. Setting CardData to empty. [default4]:Repo card metadata block was not found. Setting CardData to empty. [default5]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=5|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default7]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=7|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default6]:Repo card metadata block was not found. Setting CardData to empty. [default6]:Repo card metadata block was not found. Setting CardData to empty. [default4]:Repo card metadata block was not found. Setting CardData to empty. [default6]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=6|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default4]:07/02/2024 18:45:33 [WARNING|DP=0|PP=1|TP=4|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty. [default6]:07/02/2024 18:45:33 [WARNING|DP=0|PP=0|TP=6|ip-26-0-165-24]: Repo card metadata block was not found. Setting CardData to empty. [default7]:Repo card metadata block was not found. Setting CardData to empty. [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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 [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]:/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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [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( [default0]:07/02/2024 18:46:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 911.23MiB. Peak allocated 30864.21MiB. Peak reserved: 31232.00MiB [default0]:07/02/2024 18:46:12 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 1 / 20 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 36.4K | tokens_per_sec: 115K | tokens_per_sec_per_gpu: 7.2K | global_batch_size: 1.02K | lm_loss: 11.2 | lr: 0.0001 | model_tflops_per_gpu: 65.3 | hardware_tflops_per_gpu: 65.3 | grad_norm: 12.1 | cuda_memory_allocated: 1.26G | cuda_max_memory_reserved: 16.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G [default0]:07/02/2024 18:46:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 1572.71MiB. Peak reserved: 31232.00MiB [default0]:07/02/2024 18:46:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:46:30 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 2 / 20 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18K | tokens_per_sec: 233K | tokens_per_sec_per_gpu: 14.5K | global_batch_size: 1.02K | lm_loss: 11.2 | lr: 9.53e-05 | model_tflops_per_gpu: 132 | hardware_tflops_per_gpu: 132 | grad_norm: 12.2 | cuda_memory_allocated: 1.26G | cuda_max_memory_reserved: 16.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G [default0]:07/02/2024 18:46:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 1572.75MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:46:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:46:48 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 3 / 20 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 18.1K | tokens_per_sec: 232K | tokens_per_sec_per_gpu: 14.5K | global_batch_size: 1.02K | lm_loss: 10 | lr: 9.05e-05 | model_tflops_per_gpu: 132 | hardware_tflops_per_gpu: 132 | grad_norm: 51.6 | cuda_memory_allocated: 1.26G | cuda_max_memory_reserved: 16.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G [default0]:STAGE:2024-07-02 18:46:48 655682:655682 ActivityProfilerController.cpp:314] Completed Stage: Warm Up [default0]:07/02/2024 18:46:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 1572.75MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:47:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:47:05 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 4 / 20 | consumed_tokens: 16.8M | elapsed_time_per_iteration_ms: 17.4K | tokens_per_sec: 241K | tokens_per_sec_per_gpu: 15K | global_batch_size: 1.02K | lm_loss: 11.7 | lr: 8.58e-05 | model_tflops_per_gpu: 136 | hardware_tflops_per_gpu: 136 | grad_norm: 18.2 | cuda_memory_allocated: 1.26G | cuda_max_memory_reserved: 16.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G [default0]:07/02/2024 18:47:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 1572.75MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:47:23 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 5 / 20 | consumed_tokens: 21M | elapsed_time_per_iteration_ms: 17.5K | tokens_per_sec: 239K | tokens_per_sec_per_gpu: 14.9K | global_batch_size: 1.02K | lm_loss: 10.4 | lr: 8.11e-05 | model_tflops_per_gpu: 136 | hardware_tflops_per_gpu: 136 | grad_norm: 16 [default0]:07/02/2024 18:47:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:47:40 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 6 / 20 | consumed_tokens: 25.2M | elapsed_time_per_iteration_ms: 17.5K | tokens_per_sec: 240K | tokens_per_sec_per_gpu: 15K | global_batch_size: 1.02K | lm_loss: 9.9 | lr: 7.63e-05 | model_tflops_per_gpu: 136 | hardware_tflops_per_gpu: 136 | grad_norm: 9.07 [default0]:STAGE:2024-07-02 18:48:02 655682:655682 ActivityProfilerController.cpp:320] Completed Stage: Collection [default0]:STAGE:2024-07-02 18:48:05 655682:655682 ActivityProfilerController.cpp:324] Completed Stage: Post Processing [default0]:07/02/2024 18:50:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:51:16 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 7 / 20 | consumed_tokens: 29.4M | elapsed_time_per_iteration_ms: 216K | tokens_per_sec: 19.4K | tokens_per_sec_per_gpu: 1.21K | global_batch_size: 1.02K | lm_loss: 9.37 | lr: 7.16e-05 | model_tflops_per_gpu: 11 | hardware_tflops_per_gpu: 11 | grad_norm: 6.23 [default0]:07/02/2024 18:51:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:51:34 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 8 / 20 | consumed_tokens: 33.6M | elapsed_time_per_iteration_ms: 17.4K | tokens_per_sec: 241K | tokens_per_sec_per_gpu: 15.1K | global_batch_size: 1.02K | lm_loss: 8.89 | lr: 6.68e-05 | model_tflops_per_gpu: 137 | hardware_tflops_per_gpu: 137 | grad_norm: 5.76 [default0]:07/02/2024 18:51:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:51:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:51:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 9 / 20 | consumed_tokens: 37.7M | elapsed_time_per_iteration_ms: 18.3K | tokens_per_sec: 229K | tokens_per_sec_per_gpu: 14.3K | global_batch_size: 1.02K | lm_loss: 8.8 | lr: 6.21e-05 | model_tflops_per_gpu: 130 | hardware_tflops_per_gpu: 130 | grad_norm: 11.2 [default0]:07/02/2024 18:52:09 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 10 / 20 | consumed_tokens: 41.9M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 243K | tokens_per_sec_per_gpu: 15.2K | global_batch_size: 1.02K | lm_loss: 8.33 | lr: 5.74e-05 | model_tflops_per_gpu: 138 | hardware_tflops_per_gpu: 138 | grad_norm: 5.72 [default0]:07/02/2024 18:52:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:52:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:52:26 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 11 / 20 | consumed_tokens: 46.1M | elapsed_time_per_iteration_ms: 17.2K | tokens_per_sec: 243K | tokens_per_sec_per_gpu: 15.2K | global_batch_size: 1.02K | lm_loss: 8.06 | lr: 5.26e-05 | model_tflops_per_gpu: 138 | hardware_tflops_per_gpu: 138 | grad_norm: 4.91 [default0]:07/02/2024 18:52:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:52:43 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 12 / 20 | consumed_tokens: 50.3M | elapsed_time_per_iteration_ms: 17K | tokens_per_sec: 247K | tokens_per_sec_per_gpu: 15.4K | global_batch_size: 1.02K | lm_loss: 7.9 | lr: 4.79e-05 | model_tflops_per_gpu: 140 | hardware_tflops_per_gpu: 140 | grad_norm: 4.86 [default0]:07/02/2024 18:53:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:53:01 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 13 / 20 | consumed_tokens: 54.5M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 242K | tokens_per_sec_per_gpu: 15.1K | global_batch_size: 1.02K | lm_loss: 7.75 | lr: 4.32e-05 | model_tflops_per_gpu: 137 | hardware_tflops_per_gpu: 137 | grad_norm: 4.69 [default0]:07/02/2024 18:53:18 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 14 / 20 | consumed_tokens: 58.7M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 243K | tokens_per_sec_per_gpu: 15.2K | global_batch_size: 1.02K | lm_loss: 7.62 | lr: 3.84e-05 | model_tflops_per_gpu: 138 | hardware_tflops_per_gpu: 138 | grad_norm: 4.69 [default0]:07/02/2024 18:53:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:53:35 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 15 / 20 | consumed_tokens: 62.9M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 242K | tokens_per_sec_per_gpu: 15.1K | global_batch_size: 1.02K | lm_loss: 7.48 | lr: 3.37e-05 | model_tflops_per_gpu: 137 | hardware_tflops_per_gpu: 137 | grad_norm: 4.49 [default0]:07/02/2024 18:53:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:53:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 16 / 20 | consumed_tokens: 67.1M | elapsed_time_per_iteration_ms: 17K | tokens_per_sec: 247K | tokens_per_sec_per_gpu: 15.4K | global_batch_size: 1.02K | lm_loss: 7.34 | lr: 2.89e-05 | model_tflops_per_gpu: 140 | hardware_tflops_per_gpu: 140 | grad_norm: 3.99 [default0]:07/02/2024 18:53:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:54:09 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 17 / 20 | consumed_tokens: 71.3M | elapsed_time_per_iteration_ms: 17.1K | tokens_per_sec: 245K | tokens_per_sec_per_gpu: 15.3K | global_batch_size: 1.02K | lm_loss: 7.23 | lr: 2.42e-05 | model_tflops_per_gpu: 139 | hardware_tflops_per_gpu: 139 | grad_norm: 3.54 [default0]:07/02/2024 18:54:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:54:27 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 18 / 20 | consumed_tokens: 75.5M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 242K | tokens_per_sec_per_gpu: 15.1K | global_batch_size: 1.02K | lm_loss: 7.16 | lr: 1.95e-05 | model_tflops_per_gpu: 137 | hardware_tflops_per_gpu: 137 | grad_norm: 3.28 [default0]:07/02/2024 18:54:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:54:44 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 19 / 20 | consumed_tokens: 79.7M | elapsed_time_per_iteration_ms: 17.3K | tokens_per_sec: 242K | tokens_per_sec_per_gpu: 15.1K | global_batch_size: 1.02K | lm_loss: 7.09 | lr: 1.47e-05 | model_tflops_per_gpu: 137 | hardware_tflops_per_gpu: 137 | grad_norm: 3.2 [default0]:07/02/2024 18:54:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-24]: Memory usage: 1572.71MiB. Peak allocated 31525.69MiB. Peak reserved: 31874.00MiB [default0]:07/02/2024 18:55:02 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-160]: iteration: 20 / 20 | consumed_tokens: 83.9M | elapsed_time_per_iteration_ms: 17.5K | tokens_per_sec: 240K | tokens_per_sec_per_gpu: 15K | global_batch_size: 1.02K | lm_loss: 7.03 | lr: 1e-05 | model_tflops_per_gpu: 136 | hardware_tflops_per_gpu: 136 | grad_norm: 3.1 Traceback (most recent call last): File "/fsx/ferdinandmom/ferdinand-hf/bench_cluster/main.py", line 4, in from bench_cluster.submit_jobs import submit_jobs, check_status ImportError: cannot import name 'check_status' from 'bench_cluster.submit_jobs' (/fsx/ferdinandmom/ferdinand-hf/bench_cluster/bench_cluster/submit_jobs.py) Traceback (most recent call last): File "/fsx/ferdinandmom/ferdinand-hf/bench_cluster/main.py", line 4, in from bench_cluster.submit_jobs import submit_jobs, check_status ImportError: cannot import name 'check_status' from 'bench_cluster.submit_jobs' (/fsx/ferdinandmom/ferdinand-hf/bench_cluster/bench_cluster/submit_jobs.py) 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. ip-26-0-165-24_655682.1719946219098357028.pt.trace.json: 0%| | 0.00/5.25G [00:00