| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 24.0, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 4.9500000000000004e-05, | |
| "loss": 3.554, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.8209362030029297, | |
| "eval_runtime": 59.6048, | |
| "eval_samples_per_second": 28.186, | |
| "eval_steps_per_second": 3.523, | |
| "eval_wer": 1.027634208531459, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 9.95e-05, | |
| "loss": 0.6528, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.45033037662506104, | |
| "eval_runtime": 59.343, | |
| "eval_samples_per_second": 28.31, | |
| "eval_steps_per_second": 3.539, | |
| "eval_wer": 0.438632761353456, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 8.2e-05, | |
| "loss": 0.2126, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.4427516460418701, | |
| "eval_runtime": 58.1542, | |
| "eval_samples_per_second": 28.889, | |
| "eval_steps_per_second": 3.611, | |
| "eval_wer": 0.3861208738198608, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "learning_rate": 6.381818181818183e-05, | |
| "loss": 0.122, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 0.443359375, | |
| "eval_runtime": 60.2568, | |
| "eval_samples_per_second": 27.881, | |
| "eval_steps_per_second": 3.485, | |
| "eval_wer": 0.37357866446144306, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 4.567272727272727e-05, | |
| "loss": 0.083, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_loss": 0.46422243118286133, | |
| "eval_runtime": 59.9224, | |
| "eval_samples_per_second": 28.036, | |
| "eval_steps_per_second": 3.505, | |
| "eval_wer": 0.35331817242092206, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "learning_rate": 2.7490909090909094e-05, | |
| "loss": 0.0614, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_loss": 0.4522028863430023, | |
| "eval_runtime": 60.7885, | |
| "eval_samples_per_second": 27.637, | |
| "eval_steps_per_second": 3.455, | |
| "eval_wer": 0.34635793535938253, | |
| "step": 3000 | |
| } | |
| ], | |
| "max_steps": 3750, | |
| "num_train_epochs": 30, | |
| "total_flos": 2.4789473171977364e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |