| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0, |
| "eval_steps": 500, |
| "global_step": 0, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0, |
| "eval_loss": 1.851607322692871, |
| "eval_runtime": 0.987, |
| "eval_samples_per_second": 133.739, |
| "eval_steps_per_second": 5.066, |
| "step": 0 |
| }, |
| { |
| "epoch": 0, |
| "eval_loss": 1.8677929639816284, |
| "eval_runtime": 0.901, |
| "eval_samples_per_second": 146.508, |
| "eval_steps_per_second": 5.55, |
| "step": 0 |
| }, |
| { |
| "epoch": 0, |
| "eval_loss": 1.941433072090149, |
| "eval_runtime": 0.954, |
| "eval_samples_per_second": 138.368, |
| "eval_steps_per_second": 5.241, |
| "step": 0 |
| }, |
| { |
| "epoch": 0, |
| "step": 0, |
| "total_flos": 128755402014720.0, |
| "train_loss": 0.22736623883247375, |
| "train_runtime": 62.0917, |
| "train_samples_per_second": 57.012, |
| "train_steps_per_second": 0.048 |
| } |
| ], |
| "logging_steps": 5, |
| "max_steps": 3, |
| "num_train_epochs": 3, |
| "save_steps": 500, |
| "total_flos": 128755402014720.0, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
