| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 89.28571428571429, | |
| "global_step": 2500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 17.86, | |
| "learning_rate": 4.94e-05, | |
| "loss": 19.5604, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "eval_loss": 3.8375885486602783, | |
| "eval_runtime": 2.2223, | |
| "eval_samples_per_second": 86.395, | |
| "eval_steps_per_second": 10.799, | |
| "eval_wer": 1.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "learning_rate": 9.94e-05, | |
| "loss": 2.0837, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "eval_loss": 1.0887540578842163, | |
| "eval_runtime": 2.2007, | |
| "eval_samples_per_second": 87.246, | |
| "eval_steps_per_second": 10.906, | |
| "eval_wer": 0.6354166666666666, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "learning_rate": 7.255555555555555e-05, | |
| "loss": 0.3185, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "eval_loss": 0.3722864091396332, | |
| "eval_runtime": 2.1581, | |
| "eval_samples_per_second": 88.967, | |
| "eval_steps_per_second": 11.121, | |
| "eval_wer": 0.296875, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "learning_rate": 4.477777777777778e-05, | |
| "loss": 0.1425, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "eval_loss": 0.22876793146133423, | |
| "eval_runtime": 2.2468, | |
| "eval_samples_per_second": 85.454, | |
| "eval_steps_per_second": 10.682, | |
| "eval_wer": 0.28125, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 89.29, | |
| "learning_rate": 1.7000000000000003e-05, | |
| "loss": 0.0932, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 89.29, | |
| "eval_loss": 0.16792654991149902, | |
| "eval_runtime": 2.2364, | |
| "eval_samples_per_second": 85.85, | |
| "eval_steps_per_second": 10.731, | |
| "eval_wer": 0.18229166666666666, | |
| "step": 2500 | |
| } | |
| ], | |
| "max_steps": 2800, | |
| "num_train_epochs": 100, | |
| "total_flos": 3.050974757946286e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |