<!-- 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-en-de-finetuned-en-to-de
This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-de on the wmt16 dataset. It achieves the following results on the evaluation set:
- Loss: 1.2799
- Bleu1: 0.5227
- Bleu2: 0.3993
- Rougelsum: 0.5577
- Gen Len: 27.2379
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: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu1 | Bleu2 | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|
| 1.5412 | 1.0 | 15625 | 1.2853 | 0.5232 | 0.4001 | 0.5595 | 27.1014 |
| 1.5375 | 2.0 | 31250 | 1.2816 | 0.5229 | 0.3996 | 0.5582 | 27.0881 |
| 1.5452 | 3.0 | 46875 | 1.2804 | 0.5227 | 0.3995 | 0.5577 | 27.2328 |
| 1.5405 | 4.0 | 62500 | 1.2800 | 0.5225 | 0.3993 | 0.5577 | 27.2365 |
| 1.5373 | 5.0 | 78125 | 1.2799 | 0.5227 | 0.3993 | 0.5577 | 27.2379 |
Framework versions
- Transformers 4.27.4
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
