- dataset: opus
- model: transformer-align
- source language(s): cmn nan yue
- target language(s): fin
- 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.fin-zho | 37.6 | 0.597 | 380 | 1847 | 0.942 |