- dataset: opus
- model: transformer
- source language(s): ces dsb hsb pol
- target language(s): ces dsb hsb pol
- model: transformer
- pre-processing: normalization + SentencePiece (spm32k,spm32k)
- a sentence initial language token is required in the form of
>>id<<(id = valid target language ID) - download: opus-2020-07-27.zip
- test set translations: opus-2020-07-27.test.txt
- test set scores: opus-2020-07-27.eval.txt
| testset | BLEU | chr-F |
|---|---|---|
| Tatoeba-test.ces-hsb.ces.hsb | 2.6 | 0.167 |
| Tatoeba-test.ces-pol.ces.pol | 44.0 | 0.649 |
| Tatoeba-test.dsb-pol.dsb.pol | 8.5 | 0.250 |
| Tatoeba-test.hsb-ces.hsb.ces | 9.6 | 0.276 |
| Tatoeba-test.multi.multi | 38.8 | 0.580 |
| Tatoeba-test.pol-ces.pol.ces | 43.4 | 0.620 |
| Tatoeba-test.pol-dsb.pol.dsb | 2.1 | 0.159 |
- dataset: opus
- model: transformer
- source language(s): ces csb csb_Latn dsb eng hsb pol
- target language(s): ces csb csb_Latn dsb eng hsb pol
- model: transformer
- pre-processing: normalization + SentencePiece (spm32k,spm32k)
- a sentence initial language token is required in the form of
>>id<<(id = valid target language ID) - download: opus-2020-10-04.zip
- test set translations: opus-2020-10-04.test.txt
- test set scores: opus-2020-10-04.eval.txt
| testset | BLEU | chr-F |
|---|---|---|
| newssyscomb2009-ceseng.ces.eng | 24.1 | 0.525 |
| newssyscomb2009-engces.eng.ces | 18.6 | 0.467 |
| news-test2008-engces.eng.ces | 16.0 | 0.444 |
| newstest2009-ceseng.ces.eng | 23.0 | 0.517 |
| newstest2009-engces.eng.ces | 17.2 | 0.461 |
| newstest2010-ceseng.ces.eng | 22.8 | 0.526 |
| newstest2010-engces.eng.ces | 17.6 | 0.464 |
| newstest2011-ceseng.ces.eng | 24.2 | 0.527 |
| newstest2011-engces.eng.ces | 18.9 | 0.469 |
| newstest2012-ceseng.ces.eng | 23.1 | 0.519 |
| newstest2012-engces.eng.ces | 16.8 | 0.447 |
| newstest2013-ceseng.ces.eng | 25.5 | 0.533 |
| newstest2013-engces.eng.ces | 19.3 | 0.472 |
| newstest2014-csen-ceseng.ces.eng | 27.2 | 0.560 |
| newstest2015-encs-ceseng.ces.eng | 25.0 | 0.523 |
| newstest2015-encs-engces.eng.ces | 19.2 | 0.476 |
| newstest2016-encs-ceseng.ces.eng | 26.4 | 0.545 |
| newstest2016-encs-engces.eng.ces | 22.0 | 0.497 |
| newstest2017-encs-ceseng.ces.eng | 23.4 | 0.517 |
| newstest2017-encs-engces.eng.ces | 17.9 | 0.457 |
| newstest2018-encs-ceseng.ces.eng | 24.3 | 0.524 |
| newstest2018-encs-engces.eng.ces | 17.7 | 0.458 |
| newstest2019-encs-engces.eng.ces | 18.6 | 0.464 |
| Tatoeba-test.ces-eng.ces.eng | 50.3 | 0.668 |
| Tatoeba-test.ces-hsb.ces.hsb | 1.6 | 0.148 |
| Tatoeba-test.ces-pol.ces.pol | 45.7 | 0.663 |
| Tatoeba-test.csb-eng.csb.eng | 12.6 | 0.319 |
| Tatoeba-test.dsb-eng.dsb.eng | 15.8 | 0.327 |
| Tatoeba-test.dsb-pol.dsb.pol | 15.6 | 0.336 |
| Tatoeba-test.eng-ces.eng.ces | 40.6 | 0.606 |
| Tatoeba-test.eng-csb.eng.csb | 1.1 | 0.188 |
| Tatoeba-test.eng-dsb.eng.dsb | 1.3 | 0.141 |
| Tatoeba-test.eng-hsb.eng.hsb | 2.9 | 0.175 |
| Tatoeba-test.eng-pol.eng.pol | 40.6 | 0.619 |
| Tatoeba-test.hsb-ces.hsb.ces | 17.2 | 0.341 |
| Tatoeba-test.hsb-eng.hsb.eng | 29.7 | 0.481 |
| Tatoeba-test.multi.multi | 45.2 | 0.633 |
| Tatoeba-test.pol-ces.pol.ces | 44.1 | 0.627 |
| Tatoeba-test.pol-dsb.pol.dsb | 1.1 | 0.095 |
| Tatoeba-test.pol-eng.pol.eng | 47.8 | 0.648 |
- dataset: opus
- model: transformer
- source language(s): ces dsb hsb pol
- target language(s): ces dsb hsb pol
- model: transformer
- 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: >>eng<< >>ces<< >>pol<< >>hsb<< >>csb_Latn<< >>csb<<
- download: opus-2021-02-23.zip
- test set translations: opus-2021-02-23.test.txt
- test set scores: opus-2021-02-23.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| newssyscomb2009.ces-eng | 24.1 | 0.525 | 502 | 11821 | 0.973 |
| newssyscomb2009.eng-ces | 18.6 | 0.467 | 502 | 10032 | 0.978 |
| news-test2008.eng-ces | 16.0 | 0.444 | 2051 | 42484 | 0.982 |
| newstest2009.ces-eng | 23.0 | 0.517 | 2525 | 65402 | 0.963 |
| newstest2009.eng-ces | 17.2 | 0.461 | 2525 | 55533 | 0.978 |
| newstest2010.ces-eng | 22.8 | 0.526 | 2489 | 61724 | 0.979 |
| newstest2010.eng-ces | 17.6 | 0.464 | 2489 | 52958 | 0.982 |
| newstest2011.ces-eng | 24.2 | 0.527 | 3003 | 74681 | 0.990 |
| newstest2011.eng-ces | 18.9 | 0.469 | 3003 | 65653 | 0.952 |
| newstest2012.ces-eng | 23.1 | 0.519 | 3003 | 72812 | 1.000 |
| newstest2012.eng-ces | 16.8 | 0.447 | 3003 | 65456 | 0.935 |
| newstest2013.ces-eng | 25.5 | 0.533 | 3000 | 64505 | 1.000 |
| newstest2013.eng-ces | 19.3 | 0.472 | 3000 | 57250 | 0.956 |
| newstest2014-csen.ces-eng | 27.2 | 0.560 | 3003 | 68065 | 1.000 |
| newstest2015-encs.ces-eng | 25.0 | 0.523 | 2656 | 53572 | 0.970 |
| newstest2015-encs.eng-ces | 19.2 | 0.476 | 2656 | 45931 | 1.000 |
| newstest2016-encs.ces-eng | 26.4 | 0.545 | 2999 | 64670 | 0.995 |
| newstest2016-encs.eng-ces | 22.0 | 0.497 | 2999 | 57013 | 0.987 |
| newstest2017-encs.ces-eng | 23.4 | 0.517 | 3005 | 61725 | 0.993 |
| newstest2017-encs.eng-ces | 17.9 | 0.457 | 3005 | 54461 | 0.971 |
| newstest2018-encs.ces-eng | 24.3 | 0.524 | 2983 | 63496 | 0.993 |
| newstest2018-encs.eng-ces | 17.7 | 0.458 | 2983 | 54772 | 0.996 |
| newstest2019-encs.eng-ces | 18.6 | 0.464 | 1997 | 43373 | 0.975 |
| Tatoeba-test.ces-hsb | 1.6 | 0.148 | 84 | 472 | 1.000 |
| Tatoeba-test.ces-pol | 45.7 | 0.664 | 561 | 3366 | 0.960 |
| Tatoeba-test.dsb-pol | 15.6 | 0.336 | 13 | 70 | 0.956 |
| Tatoeba-test.hsb-ces | 17.2 | 0.341 | 84 | 457 | 1.000 |
| Tatoeba-test.multi-multi | 40.2 | 0.592 | 1316 | 7792 | 0.985 |
| Tatoeba-test.pol-ces | 44.1 | 0.626 | 561 | 3348 | 0.987 |
| Tatoeba-test.pol-dsb | 1.1 | 0.095 | 13 | 75 | 1.000 |