- dataset: opus
- model: transformer-align
- source language(s): isl
- target language(s): epo
- model: transformer-align
- pre-processing: normalization + SentencePiece (spm4k,spm4k)
- download: opus-2021-02-19.zip
- test set translations: opus-2021-02-19.test.txt
- test set scores: opus-2021-02-19.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| Tatoeba-test.isl-epo | 11.8 | 0.314 | 233 | 1528 | 1.000 |