t5-small-finetuned-en-to-pt
This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3295
- Bleu: 5.6807
- Gen Len: 18.6772
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: 0.005
- 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
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 0.5787 | 1.0 | 6250 | 0.4928 | 4.1007 | 18.638 |
| 0.5089 | 2.0 | 12500 | 0.4463 | 4.3492 | 18.663 |
| 0.4652 | 3.0 | 18750 | 0.4215 | 4.68 | 18.6652 |
| 0.4353 | 4.0 | 25000 | 0.3980 | 4.8172 | 18.6708 |
| 0.4042 | 5.0 | 31250 | 0.3799 | 4.9719 | 18.6514 |
| 0.3734 | 6.0 | 37500 | 0.3676 | 5.2226 | 18.6572 |
| 0.3396 | 7.0 | 43750 | 0.3513 | 5.2693 | 18.6596 |
| 0.308 | 8.0 | 50000 | 0.3400 | 5.4546 | 18.676 |
| 0.2767 | 9.0 | 56250 | 0.3331 | 5.5649 | 18.6708 |
| 0.2424 | 10.0 | 62500 | 0.3295 | 5.6807 | 18.6772 |
Framework versions
- Transformers 4.15.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.0
- Tokenizers 0.10.3