opus-2021-02-19.zip dataset: opus model: transformer source language(s): bos cnr hbs hrv srp target language(s): eng model: transformer pre-processing: normalization + SentencePiece (spm32k,spm32k) download: opus-2021-02-19.zip test set translations: opus-2021-02-19.test.txt test set scores: opus-2021-02-19.eval.txt Benchmarks testset BLEU chr-F #sent #words BP Tatoeba-test.bos_Latn-eng 62.0 0.778 300 1818 0.961 Tatoeba-test.hbs-eng 57.0 0.713 10000 69070 0.971 Tatoeba-test.hrv-eng 57.4 0.727 1468 10556 0.959 Tatoeba-test.srp_Cyrl-eng 53.9 0.684 1577 10162 0.955 Tatoeba-test.srp_Latn-eng 57.4 0.710 6655 46297 0.974 opusTCv20210807+bt_transformer-big_2022-02-25.zip dataset: opusTCv20210807+bt model: transformer-big source language(s): bos_Cyrl bos_Latn cnr cnr_Latn hbs hbs_Cyrl hrv srp_Cyrl srp_Latn target language(s): eng raw source language(s): bos cnr hbs hrv srp raw target language(s): eng model: transformer-big pre-processing: normalization + SentencePiece (spm32k,spm32k) download: opusTCv20210807+bt_transformer-big_2022-02-25.zip test set translations: opusTCv20210807+bt_transformer-big_2022-02-25.test.txt test set scores: opusTCv20210807+bt_transformer-big_2022-02-25.eval.txt Benchmarks testset BLEU chr-F #sent #words BP Tatoeba-test-v2021-08-07.bos_Latn-eng 66.4 0.80006 301 1824 0.973 Tatoeba-test-v2021-08-07.hrv-eng 58.7 0.73512 1480 10620 0.994 Tatoeba-test-v2021-08-07.multi-eng 56.3 0.71724 10000 68833 1.000 Tatoeba-test-v2021-08-07.srp_Cyrl-eng 44.5 0.68137 1580 10180 1.000 Tatoeba-test-v2021-08-07.srp_Latn-eng 58.3 0.71699 6656 46303 1.000