| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 11.003333333333334, | |
| "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": 10.1072, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 0.00025, | |
| "loss": 6.9379, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 0.0005, | |
| "loss": 4.9342, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0005833333333333333, | |
| "loss": 4.3732, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "learning_rate": 0.0005555555555555556, | |
| "loss": 4.0714, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 0.0005277777777777777, | |
| "loss": 3.9498, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 8.01, | |
| "learning_rate": 0.0005, | |
| "loss": 3.8436, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 9.02, | |
| "learning_rate": 0.00047222222222222224, | |
| "loss": 3.625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 0.00044444444444444436, | |
| "loss": 3.5012, | |
| "step": 400 | |
| } | |
| ], | |
| "max_steps": 1200, | |
| "num_train_epochs": 9223372036854775807, | |
| "total_flos": 1.08611781525504e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |