| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 20.408163265306122, |
| "global_step": 1000, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 10.2, |
| "learning_rate": 2.777777777777778e-05, |
| "loss": 0.336, |
| "step": 500 |
| }, |
| { |
| "epoch": 20.41, |
| "learning_rate": 0.0, |
| "loss": 0.0116, |
| "step": 1000 |
| }, |
| { |
| "epoch": 20.41, |
| "step": 1000, |
| "total_flos": 4484754685255680.0, |
| "train_runtime": 3032.248, |
| "train_samples_per_second": 0.33 |
| } |
| ], |
| "max_steps": 1000, |
| "num_train_epochs": 21, |
| "total_flos": 4484754685255680.0, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|