| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.239283429302623, | |
| "global_step": 3500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 1.872040946896993e-05, | |
| "loss": 0.3165, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.744081893793986e-05, | |
| "loss": 0.2508, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.616122840690979e-05, | |
| "loss": 0.2268, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 1.488163787587972e-05, | |
| "loss": 0.1718, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 1.3602047344849649e-05, | |
| "loss": 0.1513, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 1.2322456813819578e-05, | |
| "loss": 0.156, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 1.1042866282789508e-05, | |
| "loss": 0.0976, | |
| "step": 3500 | |
| } | |
| ], | |
| "max_steps": 7815, | |
| "num_train_epochs": 5, | |
| "total_flos": 7342723792185888.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |