| Model | Language | Best Validation BLEU |
|---|---|---|
| RNN encoder without attention | Vi-EN | 11.55 |
| RNN encoder with attention | Vi-EN | 27.62 |
| Convolution encoder with attention | Vi-EN | 7.5 |
| Self attention encoder | Vi-EN | 16.14 |
| RNN encoder without attention | Zh-EN | 7.59 |
| RNN encoder with attention | Zh-EN | 17.19 |
| Self attention encoder | Vi-EN | 7.31 |