| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 13.019166666666667, | |
| "global_step": 400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 4.9999999999999996e-06, | |
| "loss": 10.8955, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 0.00025, | |
| "loss": 8.1689, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "learning_rate": 0.0005, | |
| "loss": 6.7154, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 0.0005833333333333333, | |
| "loss": 6.5645, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "learning_rate": 0.0005555555555555556, | |
| "loss": 6.3217, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 8.02, | |
| "learning_rate": 0.0005277777777777777, | |
| "loss": 6.2449, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 10.01, | |
| "learning_rate": 0.0005, | |
| "loss": 6.006, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 0.00047222222222222224, | |
| "loss": 5.7678, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 13.02, | |
| "learning_rate": 0.00044444444444444436, | |
| "loss": 5.2976, | |
| "step": 400 | |
| } | |
| ], | |
| "max_steps": 1200, | |
| "num_train_epochs": 9223372036854775807, | |
| "total_flos": 1.1011630104576e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |