| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 4.0, |
| "eval_steps": 500, |
| "global_step": 15432, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 1.0, |
| "grad_norm": 15.83103084564209, |
| "learning_rate": 4.500777604976672e-06, |
| "loss": 0.5039, |
| "step": 3858 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.8813880586133994, |
| "eval_f1_score": 0.8024135610662164, |
| "eval_loss": 0.3608132600784302, |
| "eval_runtime": 22.2351, |
| "eval_samples_per_second": 616.907, |
| "eval_steps_per_second": 19.294, |
| "step": 3858 |
| }, |
| { |
| "epoch": 2.0, |
| "grad_norm": 5.164194583892822, |
| "learning_rate": 4.000907205806118e-06, |
| "loss": 0.3467, |
| "step": 7716 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.8847415615659401, |
| "eval_f1_score": 0.8144559299384152, |
| "eval_loss": 0.34593138098716736, |
| "eval_runtime": 22.1954, |
| "eval_samples_per_second": 618.011, |
| "eval_steps_per_second": 19.328, |
| "step": 7716 |
| }, |
| { |
| "epoch": 3.0, |
| "grad_norm": 25.590179443359375, |
| "learning_rate": 3.5010368066355626e-06, |
| "loss": 0.3017, |
| "step": 11574 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_accuracy": 0.8904279361376394, |
| "eval_f1_score": 0.8128025275921421, |
| "eval_loss": 0.3391861319541931, |
| "eval_runtime": 22.1483, |
| "eval_samples_per_second": 619.325, |
| "eval_steps_per_second": 19.369, |
| "step": 11574 |
| }, |
| { |
| "epoch": 4.0, |
| "grad_norm": 14.198585510253906, |
| "learning_rate": 3.001296008294453e-06, |
| "loss": 0.2662, |
| "step": 15432 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_accuracy": 0.8891885980899613, |
| "eval_f1_score": 0.8177542293803012, |
| "eval_loss": 0.35364070534706116, |
| "eval_runtime": 22.1641, |
| "eval_samples_per_second": 618.884, |
| "eval_steps_per_second": 19.356, |
| "step": 15432 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 38580, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 10, |
| "save_steps": 500, |
| "total_flos": 5.752183370873242e+16, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
