<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
opus-mt-id-en-jakarta
This model was trained from scratch on the inglish dataset. It achieves the following results on the evaluation set:
- Loss: 0.5167
- Bleu: 67.0647
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 4000
- num_epochs: 25
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu |
|---|---|---|---|---|
| 1.2178 | 1.0 | 272 | 1.0343 | 47.6405 |
| 1.1206 | 2.0 | 544 | 0.9537 | 49.2033 |
| 1.038 | 3.0 | 816 | 0.8950 | 50.6658 |
| 0.9686 | 4.0 | 1088 | 0.8473 | 51.8963 |
| 0.9085 | 5.0 | 1360 | 0.8089 | 52.9515 |
| 0.854 | 6.0 | 1632 | 0.7728 | 53.9652 |
| 0.8002 | 7.0 | 1904 | 0.7423 | 54.9825 |
| 0.7486 | 8.0 | 2176 | 0.7127 | 55.8795 |
| 0.7006 | 9.0 | 2448 | 0.6837 | 56.9391 |
| 0.6514 | 10.0 | 2720 | 0.6618 | 57.8949 |
| 0.6059 | 11.0 | 2992 | 0.6367 | 59.0581 |
| 0.5618 | 12.0 | 3264 | 0.6180 | 59.7973 |
| 0.5186 | 13.0 | 3536 | 0.5972 | 60.9435 |
| 0.4793 | 14.0 | 3808 | 0.5788 | 61.8618 |
| 0.4386 | 15.0 | 4080 | 0.5642 | 62.9536 |
| 0.4028 | 16.0 | 4352 | 0.5519 | 63.7941 |
| 0.371 | 17.0 | 4624 | 0.5410 | 64.6409 |
| 0.3455 | 18.0 | 4896 | 0.5349 | 65.1385 |
| 0.3239 | 19.0 | 5168 | 0.5291 | 65.6674 |
| 0.3067 | 20.0 | 5440 | 0.5254 | 66.0443 |
| 0.292 | 21.0 | 5712 | 0.5220 | 66.4475 |
| 0.2808 | 22.0 | 5984 | 0.5190 | 66.5645 |
| 0.2712 | 23.0 | 6256 | 0.5179 | 66.927 |
| 0.2652 | 24.0 | 6528 | 0.5167 | 66.9501 |
| 0.2603 | 25.0 | 6800 | 0.5167 | 67.0647 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.0
- Datasets 2.10.1
- Tokenizers 0.11.0
