| { | |
| "best_metric": 28.417188512200386, | |
| "best_model_checkpoint": "./v3-turbo-low-lora-8805-qkvo/v3-turbo-low-lora-8805-qkvo/checkpoint-3500", | |
| "epoch": 10.482180293501049, | |
| "eval_steps": 500, | |
| "global_step": 5000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0482180293501049, | |
| "grad_norm": 1.8798627853393555, | |
| "learning_rate": 9.760113134728383e-06, | |
| "loss": 1.1764, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0482180293501049, | |
| "eval_loss": 0.42681750655174255, | |
| "eval_runtime": 331.0832, | |
| "eval_samples_per_second": 2.471, | |
| "eval_steps_per_second": 0.311, | |
| "eval_wer": 32.19606996329087, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0964360587002098, | |
| "grad_norm": 1.6974595785140991, | |
| "learning_rate": 9.054297904161868e-06, | |
| "loss": 0.3681, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0964360587002098, | |
| "eval_loss": 0.39048299193382263, | |
| "eval_runtime": 331.8097, | |
| "eval_samples_per_second": 2.465, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 31.48348088965666, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.1446540880503147, | |
| "grad_norm": 2.770571231842041, | |
| "learning_rate": 7.951619746781474e-06, | |
| "loss": 0.3222, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.1446540880503147, | |
| "eval_loss": 0.3774045407772064, | |
| "eval_runtime": 332.0608, | |
| "eval_samples_per_second": 2.463, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 30.598142949686892, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.1928721174004195, | |
| "grad_norm": 2.8018639087677, | |
| "learning_rate": 6.5600164834420754e-06, | |
| "loss": 0.3158, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.1928721174004195, | |
| "eval_loss": 0.36975201964378357, | |
| "eval_runtime": 330.6434, | |
| "eval_samples_per_second": 2.474, | |
| "eval_steps_per_second": 0.312, | |
| "eval_wer": 29.23774562729432, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.241090146750524, | |
| "grad_norm": 4.796682357788086, | |
| "learning_rate": 5.015707937429398e-06, | |
| "loss": 0.2983, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.241090146750524, | |
| "eval_loss": 0.36699235439300537, | |
| "eval_runtime": 332.0572, | |
| "eval_samples_per_second": 2.463, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 28.978622327790976, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 6.289308176100629, | |
| "grad_norm": 2.243896007537842, | |
| "learning_rate": 3.4698617890574972e-06, | |
| "loss": 0.283, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 6.289308176100629, | |
| "eval_loss": 0.3649148643016815, | |
| "eval_runtime": 332.7836, | |
| "eval_samples_per_second": 2.458, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 28.805873461455413, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 7.337526205450734, | |
| "grad_norm": 2.2685556411743164, | |
| "learning_rate": 2.0737962298724513e-06, | |
| "loss": 0.2714, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 7.337526205450734, | |
| "eval_loss": 0.36280229687690735, | |
| "eval_runtime": 331.6361, | |
| "eval_samples_per_second": 2.467, | |
| "eval_steps_per_second": 0.311, | |
| "eval_wer": 28.417188512200386, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 8.385744234800839, | |
| "grad_norm": 1.9245986938476562, | |
| "learning_rate": 9.641678837679985e-07, | |
| "loss": 0.263, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 8.385744234800839, | |
| "eval_loss": 0.36285853385925293, | |
| "eval_runtime": 332.0272, | |
| "eval_samples_per_second": 2.464, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 28.546750161952062, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 9.433962264150944, | |
| "grad_norm": 3.1871135234832764, | |
| "learning_rate": 2.49594904496141e-07, | |
| "loss": 0.2656, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 9.433962264150944, | |
| "eval_loss": 0.36288538575172424, | |
| "eval_runtime": 332.2149, | |
| "eval_samples_per_second": 2.462, | |
| "eval_steps_per_second": 0.31, | |
| "eval_wer": 28.762686244871517, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 10.482180293501049, | |
| "grad_norm": 1.539926528930664, | |
| "learning_rate": 3.55305337634837e-11, | |
| "loss": 0.2607, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 10.482180293501049, | |
| "eval_loss": 0.36288636922836304, | |
| "eval_runtime": 331.4383, | |
| "eval_samples_per_second": 2.468, | |
| "eval_steps_per_second": 0.311, | |
| "eval_wer": 28.762686244871517, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 10.482180293501049, | |
| "step": 5000, | |
| "total_flos": 6.84657859756032e+19, | |
| "train_loss": 0.38244903564453125, | |
| "train_runtime": 13834.795, | |
| "train_samples_per_second": 2.891, | |
| "train_steps_per_second": 0.361 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 11, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.84657859756032e+19, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |