- dataset: opus
- model: transformer-align
- source language(s): fas pes
- target language(s): fra
- model: transformer-align
- pre-processing: normalization + SentencePiece (spm32k,spm32k)
- download: opus-2021-05-15.zip
- test set translations: opus-2021-05-15.test.txt
- test set scores: opus-2021-05-15.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| Tatoeba-test.fas-fra | 32.2 | 0.537 | 750 | 6742 | 1.000 |