- dataset: opus
- model: transformer-align
- source language(s): tur
- target language(s): fin
- model: transformer-align
- pre-processing: normalization + SentencePiece (spm32k,spm32k)
- download: opus-2021-03-08.zip
- test set translations: opus-2021-03-08.test.txt
- test set scores: opus-2021-03-08.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| Tatoeba-test.tur-fin | 33.2 | 0.575 | 1796 | 10461 | 0.918 |