<!-- 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. -->
es_fi_all_copy_quy
This model is a fine-tuned version of Helsinki-NLP/opus-mt-es-fi on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4826
- Bleu: 1.0774
- Chrf: 27.8374
- Gen Len: 49.2425
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: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Chrf | Gen Len |
|---|---|---|---|---|---|---|
| 0.5479 | 0.03 | 1000 | 0.6737 | 0.3055 | 13.5765 | 61.5895 |
| 0.4856 | 0.07 | 2000 | 0.6330 | 0.1975 | 16.9698 | 73.9869 |
| 0.4541 | 0.1 | 3000 | 0.6120 | 0.1727 | 17.4235 | 77.335 |
| 0.4381 | 0.13 | 4000 | 0.5962 | 0.3781 | 18.0872 | 48.1066 |
| 0.4236 | 0.17 | 5000 | 0.5880 | 0.3741 | 18.808 | 60.0463 |
| 0.4165 | 0.2 | 6000 | 0.5721 | 0.4366 | 19.5753 | 54.2746 |
| 0.3978 | 0.23 | 7000 | 0.5669 | 0.3667 | 20.4865 | 57.9366 |
| 0.3916 | 0.27 | 8000 | 0.5536 | 0.8014 | 21.485 | 47.503 |
| 0.3747 | 0.3 | 9000 | 0.5501 | 0.5344 | 21.0513 | 60.3924 |
| 0.3825 | 0.33 | 10000 | 0.5375 | 0.4356 | 22.2798 | 56.7726 |
| 0.3716 | 0.36 | 11000 | 0.5390 | 0.5945 | 23.2025 | 53.9366 |
| 0.3518 | 0.4 | 12000 | 0.5312 | 0.7343 | 23.1143 | 52.326 |
| 0.3495 | 0.43 | 13000 | 0.5217 | 1.0112 | 24.6907 | 49.3903 |
| 0.3508 | 0.46 | 14000 | 0.5222 | 0.5915 | 23.556 | 57.2656 |
| 0.3464 | 0.5 | 15000 | 0.5165 | 0.7044 | 24.2944 | 59.159 |
| 0.3476 | 0.53 | 16000 | 0.5142 | 1.1 | 25.0408 | 52.1127 |
| 0.3444 | 0.56 | 17000 | 0.5108 | 0.9361 | 24.8808 | 48.9366 |
| 0.3404 | 0.6 | 18000 | 0.5056 | 0.9007 | 25.3524 | 51.6962 |
| 0.3286 | 0.63 | 19000 | 0.5029 | 0.9896 | 24.8249 | 52.7233 |
| 0.3362 | 0.66 | 20000 | 0.5012 | 0.8774 | 26.6471 | 47.3732 |
| 0.3386 | 0.7 | 21000 | 0.5028 | 0.8963 | 25.7596 | 54.1861 |
| 0.3371 | 0.73 | 22000 | 0.5002 | 0.8477 | 24.9775 | 54.2958 |
| 0.3284 | 0.76 | 23000 | 0.4955 | 0.9197 | 25.5964 | 55.2606 |
| 0.3239 | 0.8 | 24000 | 0.4943 | 1.1353 | 27.5912 | 48.5714 |
| 0.3214 | 0.83 | 25000 | 0.4889 | 0.9846 | 26.098 | 51.0221 |
| 0.3183 | 0.86 | 26000 | 0.4889 | 1.1816 | 28.4223 | 44.1288 |
| 0.3164 | 0.89 | 27000 | 0.4874 | 1.028 | 27.4202 | 48.2616 |
| 0.3162 | 0.93 | 28000 | 0.4839 | 1.0008 | 26.1891 | 54.4095 |
| 0.3167 | 0.96 | 29000 | 0.4837 | 1.1952 | 27.7299 | 47.3159 |
| 0.3141 | 0.99 | 30000 | 0.4809 | 0.7488 | 27.6076 | 48.9638 |
| 0.3043 | 1.03 | 31000 | 0.4826 | 1.0774 | 27.8374 | 49.2425 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu117
- Datasets 2.11.0
- Tokenizers 0.13.3
