| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "global_step": 2060, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 4.85, | |
| "learning_rate": 1.5145631067961166e-05, | |
| "loss": 0.4008, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 9.71, | |
| "learning_rate": 1.029126213592233e-05, | |
| "loss": 0.1154, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 14.56, | |
| "learning_rate": 5.436893203883496e-06, | |
| "loss": 0.0677, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 19.42, | |
| "learning_rate": 5.825242718446603e-07, | |
| "loss": 0.0481, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 2060, | |
| "total_flos": 2.284156053808128e+16, | |
| "train_loss": 0.15463279879208908, | |
| "train_runtime": 2030.3525, | |
| "train_samples_per_second": 64.383, | |
| "train_steps_per_second": 1.015 | |
| } | |
| ], | |
| "max_steps": 2060, | |
| "num_train_epochs": 20, | |
| "total_flos": 2.284156053808128e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |