opus1m-2021-05-16.zip dataset: opus1m model: transformer-align source language(s): est fin hun izh mhr target language(s): rus model: transformer-align pre-processing: normalization + SentencePiece (spm32k,spm32k) download: opus1m-2021-05-16.zip test set translations: opus1m-2021-05-16.test.txt test set scores: opus1m-2021-05-16.eval.txt Benchmarks testset BLEU chr-F #sent #words BP Tatoeba-test.chm-rus 1.4 0.150 2750 20215 1.000 Tatoeba-test.est-rus 48.5 0.673 598 3563 0.991 Tatoeba-test.fin-rus 39.9 0.613 3643 21434 1.000 Tatoeba-test.hun-rus 37.2 0.574 2500 16105 0.983 Tatoeba-test.izh-rus 34.7 0.655 8 32 1.000 Tatoeba-test.multi-rus 26.9 0.440 9499 61541 1.000 Tatoeba-test.udm-rus 0.4 0.101 54 317 1.000