| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.090909090909092, | |
| "global_step": 3500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.3506493506493503e-05, | |
| "loss": 0.3165, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 3.701298701298702e-05, | |
| "loss": 0.178, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 3.051948051948052e-05, | |
| "loss": 0.1107, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.19, | |
| "learning_rate": 2.4025974025974027e-05, | |
| "loss": 0.058, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.49, | |
| "learning_rate": 1.7532467532467535e-05, | |
| "loss": 0.0306, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 1.103896103896104e-05, | |
| "loss": 0.0186, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "learning_rate": 4.5454545454545455e-06, | |
| "loss": 0.0117, | |
| "step": 3500 | |
| } | |
| ], | |
| "max_steps": 3850, | |
| "num_train_epochs": 10, | |
| "total_flos": 7364741550581760.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |