metadata
base_model: Tobius/opus-mt-en-lg-finetuned-en-to-lg
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: opus-mt-en-lg-finetuned-en-to-lg-finetuned-en-to-lm
results: []
opus-mt-en-lg-finetuned-en-to-lg-finetuned-en-to-lm
This model is a fine-tuned version of Tobius/opus-mt-en-lg-finetuned-en-to-lg on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7639
- Bleu: 30.9462
- Gen Len: 21.6305
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: 0.0001
- 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
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| No log | 1.0 | 146 | 3.6537 | 0.5489 | 30.8399 |
| No log | 2.0 | 292 | 3.0866 | 0.5298 | 22.7759 |
| No log | 3.0 | 438 | 2.7528 | 0.8364 | 26.5419 |
| 3.8771 | 4.0 | 584 | 2.4983 | 1.6885 | 25.3966 |
| 3.8771 | 5.0 | 730 | 2.2893 | 2.265 | 23.0296 |
| 3.8771 | 6.0 | 876 | 2.1164 | 3.5594 | 21.9409 |
| 2.5379 | 7.0 | 1022 | 1.9606 | 3.8191 | 22.8325 |
| 2.5379 | 8.0 | 1168 | 1.8145 | 4.7176 | 22.5025 |
| 2.5379 | 9.0 | 1314 | 1.6974 | 5.5036 | 21.7931 |
| 2.5379 | 10.0 | 1460 | 1.5895 | 7.0622 | 21.9483 |
| 2.0088 | 11.0 | 1606 | 1.4934 | 7.9064 | 22.2143 |
| 2.0088 | 12.0 | 1752 | 1.4120 | 8.7831 | 20.9754 |
| 2.0088 | 13.0 | 1898 | 1.3244 | 10.7815 | 21.899 |
| 1.6503 | 14.0 | 2044 | 1.2544 | 11.9959 | 21.835 |
| 1.6503 | 15.0 | 2190 | 1.1843 | 12.9721 | 21.766 |
| 1.6503 | 16.0 | 2336 | 1.1271 | 15.3546 | 21.5542 |
| 1.6503 | 17.0 | 2482 | 1.0670 | 15.9538 | 21.7906 |
| 1.3905 | 18.0 | 2628 | 1.0184 | 19.1229 | 21.9212 |
| 1.3905 | 19.0 | 2774 | 0.9797 | 20.3025 | 21.0148 |
| 1.3905 | 20.0 | 2920 | 0.9358 | 22.3064 | 21.6946 |
| 1.1903 | 21.0 | 3066 | 0.9012 | 24.9192 | 21.3128 |
| 1.1903 | 22.0 | 3212 | 0.8704 | 25.8138 | 21.3202 |
| 1.1903 | 23.0 | 3358 | 0.8481 | 26.8542 | 21.5591 |
| 1.0573 | 24.0 | 3504 | 0.8228 | 28.4838 | 21.3251 |
| 1.0573 | 25.0 | 3650 | 0.8069 | 29.1448 | 21.399 |
| 1.0573 | 26.0 | 3796 | 0.7892 | 29.8567 | 21.5985 |
| 1.0573 | 27.0 | 3942 | 0.7777 | 29.8862 | 21.6404 |
| 0.9649 | 28.0 | 4088 | 0.7703 | 30.9483 | 21.7315 |
| 0.9649 | 29.0 | 4234 | 0.7660 | 30.9831 | 21.6108 |
| 0.9649 | 30.0 | 4380 | 0.7639 | 30.9462 | 21.6305 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2