opusTCv20210807_transformer-align_2022-03-13.zip dataset: opusTCv20210807 model: transformer-align source language(s): lav lit ltg target language(s): rus ukr raw source language(s): lav lit ltg raw target language(s): rus ukr model: transformer-align pre-processing: normalization + SentencePiece (spm32k,spm32k) a sentence initial language token is required in the form of >>id<< (id = valid target language ID) valid language labels: download: opusTCv20210807_transformer-align_2022-03-13.zip test set translations: opusTCv20210807_transformer-align_2022-03-13.test.txt test set scores: opusTCv20210807_transformer-align_2022-03-13.eval.txt Benchmarks testset BLEU chr-F #sent #words BP Tatoeba-test-v2021-08-07.lav-rus 58.4 0.74782 274 1533 1.000 Tatoeba-test-v2021-08-07.lav-ukr 8.6 0.62064 1 7 1.000 Tatoeba-test-v2021-08-07.lit-rus 53.5 0.71945 3598 21831 1.000 Tatoeba-test-v2021-08-07.lit-ukr 24.2 0.69077 3 23 0.956 Tatoeba-test-v2021-08-07.multi-multi 55.2 0.72972 3876 23479 0.998