| { | |
| "best_metric": 0.4810967445373535, | |
| "best_model_checkpoint": "whisper-m-wo/checkpoint-7000", | |
| "epoch": 3.912363067292645, | |
| "eval_steps": 1000, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.19561815336463223, | |
| "grad_norm": 14.263213157653809, | |
| "learning_rate": 9.714285714285715e-06, | |
| "loss": 1.2317, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.39123630672926446, | |
| "grad_norm": 11.0503568649292, | |
| "learning_rate": 9.401002506265666e-06, | |
| "loss": 0.7816, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.39123630672926446, | |
| "eval_loss": 0.7274152040481567, | |
| "eval_runtime": 1154.7126, | |
| "eval_samples_per_second": 0.875, | |
| "eval_steps_per_second": 0.11, | |
| "eval_wer": 0.6368868366459212, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.5868544600938967, | |
| "grad_norm": 9.666540145874023, | |
| "learning_rate": 9.087719298245615e-06, | |
| "loss": 0.7047, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.7824726134585289, | |
| "grad_norm": 9.399227142333984, | |
| "learning_rate": 8.774436090225564e-06, | |
| "loss": 0.6368, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.7824726134585289, | |
| "eval_loss": 0.6092658042907715, | |
| "eval_runtime": 1128.5812, | |
| "eval_samples_per_second": 0.895, | |
| "eval_steps_per_second": 0.113, | |
| "eval_wer": 0.5042494813625109, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.9780907668231612, | |
| "grad_norm": 9.21610164642334, | |
| "learning_rate": 8.461152882205514e-06, | |
| "loss": 0.5824, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.1737089201877935, | |
| "grad_norm": 7.924953460693359, | |
| "learning_rate": 8.147869674185465e-06, | |
| "loss": 0.3921, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.1737089201877935, | |
| "eval_loss": 0.5506494641304016, | |
| "eval_runtime": 1090.6405, | |
| "eval_samples_per_second": 0.926, | |
| "eval_steps_per_second": 0.116, | |
| "eval_wer": 0.4280265007026701, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.3693270735524257, | |
| "grad_norm": 5.910419464111328, | |
| "learning_rate": 7.834586466165414e-06, | |
| "loss": 0.3708, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.5649452269170578, | |
| "grad_norm": 7.41471529006958, | |
| "learning_rate": 7.521303258145364e-06, | |
| "loss": 0.3494, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.5649452269170578, | |
| "eval_loss": 0.5246758460998535, | |
| "eval_runtime": 999.9485, | |
| "eval_samples_per_second": 1.01, | |
| "eval_steps_per_second": 0.127, | |
| "eval_wer": 0.3114501773405608, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.76056338028169, | |
| "grad_norm": 4.512611389160156, | |
| "learning_rate": 7.208646616541354e-06, | |
| "loss": 0.3506, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.9561815336463224, | |
| "grad_norm": 7.675483226776123, | |
| "learning_rate": 6.895363408521304e-06, | |
| "loss": 0.3264, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.9561815336463224, | |
| "eval_loss": 0.4906909167766571, | |
| "eval_runtime": 1031.2068, | |
| "eval_samples_per_second": 0.979, | |
| "eval_steps_per_second": 0.123, | |
| "eval_wer": 0.32925115438666935, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.151799687010955, | |
| "grad_norm": 7.257579326629639, | |
| "learning_rate": 6.582080200501254e-06, | |
| "loss": 0.204, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.347417840375587, | |
| "grad_norm": 6.650781154632568, | |
| "learning_rate": 6.269423558897244e-06, | |
| "loss": 0.1734, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.347417840375587, | |
| "eval_loss": 0.49680861830711365, | |
| "eval_runtime": 997.6386, | |
| "eval_samples_per_second": 1.012, | |
| "eval_steps_per_second": 0.127, | |
| "eval_wer": 0.2973298534430837, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.543035993740219, | |
| "grad_norm": 8.005096435546875, | |
| "learning_rate": 5.956140350877193e-06, | |
| "loss": 0.1718, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.7386541471048513, | |
| "grad_norm": 6.306284427642822, | |
| "learning_rate": 5.6428571428571435e-06, | |
| "loss": 0.1808, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.7386541471048513, | |
| "eval_loss": 0.4810967445373535, | |
| "eval_runtime": 972.537, | |
| "eval_samples_per_second": 1.039, | |
| "eval_steps_per_second": 0.131, | |
| "eval_wer": 0.2595195074616877, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.9342723004694835, | |
| "grad_norm": 10.663891792297363, | |
| "learning_rate": 5.329573934837093e-06, | |
| "loss": 0.1747, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.1298904538341157, | |
| "grad_norm": 5.245405673980713, | |
| "learning_rate": 5.016290726817044e-06, | |
| "loss": 0.1064, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.1298904538341157, | |
| "eval_loss": 0.4989220201969147, | |
| "eval_runtime": 978.6774, | |
| "eval_samples_per_second": 1.032, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.2490129157465034, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.325508607198748, | |
| "grad_norm": 3.174140691757202, | |
| "learning_rate": 4.703007518796993e-06, | |
| "loss": 0.0804, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.52112676056338, | |
| "grad_norm": 0.9650879502296448, | |
| "learning_rate": 4.389724310776943e-06, | |
| "loss": 0.0802, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.52112676056338, | |
| "eval_loss": 0.4974755346775055, | |
| "eval_runtime": 945.4502, | |
| "eval_samples_per_second": 1.068, | |
| "eval_steps_per_second": 0.134, | |
| "eval_wer": 0.22753128555176336, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.7167449139280127, | |
| "grad_norm": 6.479980945587158, | |
| "learning_rate": 4.076441102756892e-06, | |
| "loss": 0.0817, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.912363067292645, | |
| "grad_norm": 4.768033981323242, | |
| "learning_rate": 3.7631578947368426e-06, | |
| "loss": 0.0745, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.912363067292645, | |
| "eval_loss": 0.48827001452445984, | |
| "eval_runtime": 978.3234, | |
| "eval_samples_per_second": 1.032, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.24292310780967677, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.912363067292645, | |
| "step": 10000, | |
| "total_flos": 8.16483926016e+19, | |
| "train_loss": 0.35271793479919433, | |
| "train_runtime": 106272.5194, | |
| "train_samples_per_second": 1.204, | |
| "train_steps_per_second": 0.151 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 16000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 7, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.16483926016e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |