OPUS-MT-train/models/en-cs

opus-2019-12-04.zip

• dataset: opus
• model: transformer
• 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
newssyscomb2009.en.cs	22.6	0.499
news-test2008.en.cs	20.2	0.473
newstest2009.en.cs	21.2	0.488
newstest2010.en.cs	21.2	0.493
newstest2011.en.cs	22.5	0.491
newstest2012.en.cs	20.2	0.468
newstest2013.en.cs	24.2	0.501
newstest2015-encs.en.cs	24.7	0.516
newstest2016-encs.en.cs	26.9	0.534
newstest2017-encs.en.cs	22.3	0.493
newstest2018-encs.en.cs	22.3	0.489
newstest2019-encs.en.cs	24.3	0.499
Tatoeba.en.cs	48.2	0.658

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
newssyscomb2009.en.cs	22.8	0.507
news-test2008.en.cs	20.7	0.485
newstest2009.en.cs	21.8	0.500
newstest2010.en.cs	22.1	0.505
newstest2011.en.cs	23.2	0.507
newstest2012.en.cs	20.8	0.482
newstest2013.en.cs	24.7	0.514
newstest2015-encs.en.cs	24.9	0.527
newstest2016-encs.en.cs	26.7	0.540
newstest2017-encs.en.cs	22.7	0.503
newstest2018-encs.en.cs	22.9	0.504
newstest2019-encs.en.cs	24.9	0.518
Tatoeba.en.cs	46.1	0.647