lr = 5e-05per_device_batch_size = 16gradient_accumulation_steps = 1weight_decay = 1e-09seed = 42numerical stability denominator constant = 0.01lambda = 0.02alpha = 2.0beta = 1.0| eval_loss | eval_accuracy | epoch |
|---|---|---|
| 8.986 | 0.938 | 1.0 |
| 9.101 | 0.932 | 2.0 |
| 8.912 | 0.941 | 3.0 |
| 8.916 | 0.941 | 4.0 |
| 8.902 | 0.942 | 5.0 |
| 8.814 | 0.946 | 6.0 |
| 8.843 | 0.945 | 7.0 |
| 8.837 | 0.945 | 8.0 |
| 8.830 | 0.945 | 9.0 |
| 8.847 | 0.944 | 10.0 |
| 8.822 | 0.945 | 11.0 |
| 8.808 | 0.946 | 12.0 |
| 8.806 | 0.946 | 13.0 |
| 8.814 | 0.946 | 14.0 |
| 8.807 | 0.946 | 15.0 |
| 8.933 | 0.941 | 16.0 |
| 8.824 | 0.945 | 17.0 |
| 8.840 | 0.945 | 18.0 |
| 8.854 | 0.944 | 19.0 |
| 8.850 | 0.944 | 20.0 |
| 8.901 | 0.942 | 21.0 |
| 8.862 | 0.944 | 22.0 |
| 8.856 | 0.944 | 23.0 |
| 8.870 | 0.943 | 24.0 |
| 8.800 | 0.946 | 25.0 |
| 8.800 | 0.947 | 26.0 |
| 8.825 | 0.945 | 27.0 |
| 8.782 | 0.947 | 28.0 |
| 8.772 | 0.948 | 29.0 |
| 8.767 | 0.948 | 30.0 |
| 8.756 | 0.948 | 31.0 |
| 8.793 | 0.947 | 32.0 |
| 8.738 | 0.949 | 33.0 |
| 8.728 | 0.949 | 34.0 |
| 8.709 | 0.950 | 35.0 |
| 8.694 | 0.951 | 36.0 |
| 8.740 | 0.949 | 37.0 |
| 8.723 | 0.950 | 38.0 |
| 8.710 | 0.950 | 39.0 |
| 8.735 | 0.949 | 40.0 |
| 8.705 | 0.950 | 41.0 |
| 8.699 | 0.951 | 42.0 |
| 8.714 | 0.950 | 43.0 |
| 8.706 | 0.950 | 44.0 |
| 8.699 | 0.951 | 45.0 |
| 8.714 | 0.950 | 46.0 |
| 8.706 | 0.950 | 47.0 |
| 8.701 | 0.951 | 48.0 |
| 8.711 | 0.950 | 49.0 |