t5-small-finetuned-de-to-en
This model is a fine-tuned version of t5-small on the wmt16 dataset. It achieves the following results on the evaluation set:
- Loss: 1.9417
- Bleu: 9.2166
- Gen Len: 17.3404
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| No log | 1.0 | 272 | 2.1660 | 3.8515 | 17.6289 |
| 2.6678 | 2.0 | 544 | 2.0656 | 6.4422 | 17.4842 |
| 2.6678 | 3.0 | 816 | 2.0203 | 7.4348 | 17.3741 |
| 2.4316 | 4.0 | 1088 | 1.9926 | 8.0914 | 17.3658 |
| 2.4316 | 5.0 | 1360 | 1.9739 | 8.6535 | 17.3461 |
| 2.3307 | 6.0 | 1632 | 1.9603 | 8.8757 | 17.3768 |
| 2.3307 | 7.0 | 1904 | 1.9509 | 9.0744 | 17.3511 |
| 2.2945 | 8.0 | 2176 | 1.9466 | 9.1111 | 17.3418 |
| 2.2945 | 9.0 | 2448 | 1.9427 | 9.1969 | 17.3351 |
| 2.2666 | 10.0 | 2720 | 1.9417 | 9.2166 | 17.3404 |
Framework versions
- Transformers 4.12.5
- Pytorch 1.10.0+cu111
- Datasets 1.16.1
- Tokenizers 0.10.3