opus-mt-en-bkm
This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-ro on the arrow dataset. It achieves the following results on the evaluation set:
- Loss: 1.1790
- Bleu: 17.7574
- Gen Len: 58.4209
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 2.1758 | 1.0 | 1113 | 1.8681 | 4.1739 | 58.6351 |
| 1.8143 | 2.0 | 2226 | 1.6288 | 6.2869 | 62.8396 |
| 1.635 | 3.0 | 3339 | 1.4789 | 7.8756 | 58.5721 |
| 1.4988 | 4.0 | 4452 | 1.3930 | 9.2821 | 59.5793 |
| 1.3753 | 5.0 | 5565 | 1.3288 | 10.4942 | 58.924 |
| 1.3015 | 6.0 | 6678 | 1.2773 | 11.3724 | 60.0849 |
| 1.2424 | 7.0 | 7791 | 1.2419 | 12.1525 | 60.724 |
| 1.1758 | 8.0 | 8904 | 1.2131 | 12.5595 | 58.5216 |
| 1.1263 | 9.0 | 10017 | 1.1882 | 13.4807 | 58.1827 |
| 1.0781 | 10.0 | 11130 | 1.1720 | 13.6583 | 56.953 |
| 1.0377 | 11.0 | 12243 | 1.1571 | 14.2744 | 58.1146 |
| 1.0014 | 12.0 | 13356 | 1.1437 | 14.5804 | 57.9928 |
| 0.9737 | 13.0 | 14469 | 1.1326 | 14.9612 | 57.4652 |
| 0.9384 | 14.0 | 15582 | 1.1263 | 15.1647 | 58.4813 |
| 0.9061 | 15.0 | 16695 | 1.1262 | 15.3948 | 57.8562 |
| 0.8854 | 16.0 | 17808 | 1.1164 | 15.7348 | 57.8652 |
| 0.8657 | 17.0 | 18921 | 1.1179 | 15.9306 | 57.5578 |
| 0.837 | 18.0 | 20034 | 1.1140 | 16.0704 | 58.2836 |
| 0.8208 | 19.0 | 21147 | 1.1135 | 16.1836 | 57.6796 |
| 0.7919 | 20.0 | 22260 | 1.1117 | 16.4418 | 57.7658 |
| 0.7645 | 21.0 | 23373 | 1.1134 | 16.3838 | 58.2189 |
| 0.7519 | 22.0 | 24486 | 1.1157 | 16.4369 | 57.7701 |
| 0.7375 | 23.0 | 25599 | 1.1178 | 16.4328 | 57.5811 |
| 0.7221 | 24.0 | 26712 | 1.1186 | 16.8289 | 57.3139 |
| 0.7009 | 25.0 | 27825 | 1.1190 | 16.9092 | 57.9038 |
| 0.6882 | 26.0 | 28938 | 1.1254 | 17.0946 | 58.229 |
| 0.6778 | 27.0 | 30051 | 1.1246 | 17.1689 | 58.5953 |
| 0.6668 | 28.0 | 31164 | 1.1281 | 17.1734 | 58.1258 |
| 0.6589 | 29.0 | 32277 | 1.1322 | 16.9988 | 58.0218 |
| 0.639 | 30.0 | 33390 | 1.1297 | 17.2725 | 58.3717 |
| 0.6318 | 31.0 | 34503 | 1.1392 | 17.3926 | 57.9088 |
| 0.6174 | 32.0 | 35616 | 1.1429 | 17.385 | 58.6474 |
| 0.6105 | 33.0 | 36729 | 1.1443 | 17.4034 | 58.7521 |
| 0.5953 | 34.0 | 37842 | 1.1485 | 17.4571 | 58.4733 |
| 0.5897 | 35.0 | 38955 | 1.1491 | 17.4854 | 58.9544 |
| 0.5807 | 36.0 | 40068 | 1.1572 | 17.544 | 58.1013 |
| 0.5774 | 37.0 | 41181 | 1.1588 | 17.5858 | 58.4694 |
| 0.5633 | 38.0 | 42294 | 1.1588 | 17.604 | 58.2328 |
| 0.5565 | 39.0 | 43407 | 1.1640 | 17.7342 | 58.3148 |
| 0.5556 | 40.0 | 44520 | 1.1642 | 17.6596 | 58.6809 |
| 0.5469 | 41.0 | 45633 | 1.1671 | 17.5064 | 58.1013 |
| 0.5428 | 42.0 | 46746 | 1.1686 | 17.7473 | 58.5171 |
| 0.5342 | 43.0 | 47859 | 1.1719 | 17.749 | 58.8335 |
| 0.5292 | 44.0 | 48972 | 1.1730 | 17.6552 | 58.4492 |
| 0.5314 | 45.0 | 50085 | 1.1728 | 17.7932 | 58.6007 |
| 0.5283 | 46.0 | 51198 | 1.1770 | 17.7351 | 58.4564 |
| 0.5252 | 47.0 | 52311 | 1.1778 | 17.803 | 58.5793 |
| 0.5227 | 48.0 | 53424 | 1.1782 | 17.7729 | 58.3533 |
| 0.5206 | 49.0 | 54537 | 1.1788 | 17.7547 | 58.5108 |
| 0.5186 | 50.0 | 55650 | 1.1790 | 17.7574 | 58.4209 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 10