| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.8095238095238093, | |
| "eval_steps": 500, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.38095238095238093, | |
| "grad_norm": 128.80056652631617, | |
| "learning_rate": 0.0001, | |
| "loss": 1.939, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.7619047619047619, | |
| "grad_norm": 128.356467747052, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 1.9333, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 1.1428571428571428, | |
| "grad_norm": 360.8624759690317, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 9.7012, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 1.5238095238095237, | |
| "grad_norm": 123.43648733761829, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 3.0347, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 1.9047619047619047, | |
| "grad_norm": 269.2858901335159, | |
| "learning_rate": 5.868240888334653e-05, | |
| "loss": 5.0547, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 2.2857142857142856, | |
| "grad_norm": 26.38091309419736, | |
| "learning_rate": 4.131759111665349e-05, | |
| "loss": 0.7275, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 2.6666666666666665, | |
| "grad_norm": 25.39288571666346, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 0.9666, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 3.0476190476190474, | |
| "grad_norm": 10.098295434315187, | |
| "learning_rate": 1.1697777844051105e-05, | |
| "loss": 0.3615, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 3.4285714285714284, | |
| "grad_norm": 11.223311238174132, | |
| "learning_rate": 3.0153689607045845e-06, | |
| "loss": 0.2268, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 3.8095238095238093, | |
| "grad_norm": 4.0409752690257745, | |
| "learning_rate": 0.0, | |
| "loss": 0.0909, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 3.8095238095238093, | |
| "step": 10, | |
| "total_flos": 5.956250982416384e+16, | |
| "train_loss": 2.4036094665527346, | |
| "train_runtime": 166.6494, | |
| "train_samples_per_second": 4.981, | |
| "train_steps_per_second": 0.06 | |
| } | |
| ], | |
| "logging_steps": 1.0, | |
| "max_steps": 10, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 5.956250982416384e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |