| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.2718129926610492, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.269e-05, | |
| "loss": 0.7905, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 1.128e-05, | |
| "loss": 0.7419, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 9.87e-06, | |
| "loss": 0.7201, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 8.46e-06, | |
| "loss": 0.7073, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 7.05e-06, | |
| "loss": 0.6932, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 5.64e-06, | |
| "loss": 0.6966, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.23e-06, | |
| "loss": 0.6999, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 2.82e-06, | |
| "loss": 0.694, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 1.41e-06, | |
| "loss": 0.6961, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 0.0, | |
| "loss": 0.7012, | |
| "step": 10000 | |
| } | |
| ], | |
| "max_steps": 10000, | |
| "num_train_epochs": 1, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |