mirror of
https://github.com/Helsinki-NLP/OPUS-MT-train.git
synced 2024-10-26 21:19:02 +03:00
.. | ||
README.md |
opus-2019-12-04.zip
- dataset: opus
- model: transformer-align
- pre-processing: normalization + tokenization + BPE
- download: opus-2019-12-04.zip
- test set translations: opus-2019-12-04.test.txt
- test set scores: opus-2019-12-04.eval.txt
Benchmarks
testset | BLEU | chr-F |
---|---|---|
newsdev2015-enfi.fi.en | 22.5 | 0.513 |
newstest2015-enfi.fi.en | 24.0 | 0.524 |
newstest2016-enfi.fi.en | 26.5 | 0.548 |
newstest2017-enfi.fi.en | 29.0 | 0.569 |
newstest2018-enfi.fi.en | 21.4 | 0.493 |
newstest2019-fien.fi.en | 25.8 | 0.547 |
newstestB2016-enfi.fi.en | 21.9 | 0.507 |
newstestB2017-enfi.fi.en | 24.9 | 0.536 |
newstestB2017-fien.fi.en | 24.9 | 0.536 |
Tatoeba.fi.en | 56.8 | 0.704 |
opus-2019-12-18.zip
- dataset: opus
- model: transformer-align
- pre-processing: normalization + SentencePiece
- download: opus-2019-12-18.zip
- test set translations: opus-2019-12-18.test.txt
- test set scores: opus-2019-12-18.eval.txt
Benchmarks
testset | BLEU | chr-F |
---|---|---|
newsdev2015-enfi.fi.en | 25.3 | 0.537 |
newstest2015-enfi.fi.en | 26.9 | 0.548 |
newstest2016-enfi.fi.en | 29.4 | 0.570 |
newstest2017-enfi.fi.en | 32.2 | 0.592 |
newstest2018-enfi.fi.en | 24.3 | 0.519 |
newstest2019-fien.fi.en | 28.6 | 0.562 |
newstestB2016-enfi.fi.en | 24.5 | 0.526 |
newstestB2017-enfi.fi.en | 27.3 | 0.556 |
newstestB2017-fien.fi.en | 27.3 | 0.556 |
Tatoeba.fi.en | 55.3 | 0.705 |
opus-2020-02-11.zip
- dataset: opus
- model: transformer-align
- pre-processing: normalization + SentencePiece
- download: opus-2020-02-11.zip
- test set translations: opus-2020-02-11.test.txt
- test set scores: opus-2020-02-11.eval.txt
Benchmarks
testset | BLEU | chr-F |
---|---|---|
newsdev2015-enfi.fi.en | 25.1 | 0.535 |
newstest2015-enfi.fi.en | 26.8 | 0.548 |
newstest2016-enfi.fi.en | 29.1 | 0.569 |
newstest2017-enfi.fi.en | 32.7 | 0.596 |
newstest2018-enfi.fi.en | 23.9 | 0.518 |
newstest2019-fien.fi.en | 28.7 | 0.564 |
newstestB2016-enfi.fi.en | 24.2 | 0.525 |
newstestB2017-enfi.fi.en | 27.7 | 0.559 |
newstestB2017-fien.fi.en | 27.7 | 0.559 |
Tatoeba.fi.en | 57.2 | 0.717 |
opus-2020-02-13.zip
- dataset: opus
- model: transformer-align
- pre-processing: normalization + SentencePiece
- download: opus-2020-02-13.zip
- test set translations: opus-2020-02-13.test.txt
- test set scores: opus-2020-02-13.eval.txt
Benchmarks
testset | BLEU | chr-F |
---|---|---|
newsdev2015-enfi.fi.en | 25.4 | 0.538 |
newstest2015-enfi.fi.en | 27.1 | 0.549 |
newstest2016-enfi.fi.en | 29.5 | 0.572 |
newstest2017-enfi.fi.en | 33.1 | 0.598 |
newstest2018-enfi.fi.en | 24.0 | 0.519 |
newstest2019-fien.fi.en | 28.9 | 0.566 |
newstestB2016-enfi.fi.en | 24.5 | 0.527 |
newstestB2017-enfi.fi.en | 27.9 | 0.560 |
newstestB2017-fien.fi.en | 27.9 | 0.560 |
Tatoeba.fi.en | 57.4 | 0.718 |