| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "global_step": 3057, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 4.182204775924109e-05, | |
| "loss": 2.5329, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.3644095518482174e-05, | |
| "loss": 2.0633, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 2.546614327772326e-05, | |
| "loss": 1.6381, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "learning_rate": 1.7288191036964345e-05, | |
| "loss": 1.558, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 9.110238796205431e-06, | |
| "loss": 1.2778, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 9.322865554465163e-07, | |
| "loss": 1.2325, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 3057, | |
| "total_flos": 6372710852198400.0, | |
| "train_loss": 1.7085963665054715, | |
| "train_runtime": 944.735, | |
| "train_samples_per_second": 32.358, | |
| "train_steps_per_second": 3.236 | |
| } | |
| ], | |
| "max_steps": 3057, | |
| "num_train_epochs": 3, | |
| "total_flos": 6372710852198400.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |