trainer_state.json
| 1 | { |
| 2 | "best_metric": null, |
| 3 | "best_model_checkpoint": null, |
| 4 | "epoch": 3.0, |
| 5 | "global_step": 693, |
| 6 | "is_hyper_param_search": false, |
| 7 | "is_local_process_zero": true, |
| 8 | "is_world_process_zero": true, |
| 9 | "log_history": [ |
| 10 | { |
| 11 | "epoch": 2.16, |
| 12 | "learning_rate": 1.3924963924963927e-05, |
| 13 | "loss": 1.2342, |
| 14 | "step": 500 |
| 15 | }, |
| 16 | { |
| 17 | "epoch": 3.0, |
| 18 | "step": 693, |
| 19 | "total_flos": 3.5424926764630016e+16, |
| 20 | "train_runtime": 363.7124, |
| 21 | "train_samples_per_second": 1.905 |
| 22 | } |
| 23 | ], |
| 24 | "max_steps": 693, |
| 25 | "num_train_epochs": 3, |
| 26 | "total_flos": 3.5424926764630016e+16, |
| 27 | "trial_name": null, |
| 28 | "trial_params": null |
| 29 | } |
| 30 | |