| loss grad_norm learning_rate epoch step eval_loss eval_f1_micro eval_f1_macro eval_precision eval_recall eval_runtime eval_samples_per_second eval_steps_per_second train_runtime train_samples_per_second train_steps_per_second total_flos train_loss | |
| 0.0141 0.009887202642858028 5.1690821256038647e-05 0.4830917874396135 500 | |
| 0.0007 0.002767572645097971 3.3816425120772947e-06 0.966183574879227 1000 | |
| 1.0 1035 0.0052988119423389435 0.9985652797704447 0.9981785063752276 1.0 0.997134670487106 8.4238 107.196 6.767 | |
| 1.0 1035 95.0212 87.138 10.892 7716417760419840.0 0.00718910519555571 | |