| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.999388960554009, | |
| "global_step": 230, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.9971014492753625e-05, | |
| "loss": 11.4287, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 0.171, | |
| "step": 230 | |
| } | |
| ], | |
| "max_steps": 690, | |
| "num_train_epochs": 3, | |
| "total_flos": 1.1494730347472486e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |