eng-deu

opus-2021-02-17.zip

• dataset: opus
• model: transformer-align
• source language(s): eng
• target language(s): deu
• model: transformer-align
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opus-2021-02-17.zip
• test set translations: opus-2021-02-17.test.txt
• test set scores: opus-2021-02-17.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test.eng-deu	45.4	0.651	10000	83347	0.992

opus-2021-02-19.zip

• dataset: opus
• model: transformer-align
• source language(s): eng
• target language(s): deu
• model: transformer-align
• 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
newssyscomb2009.eng-deu	23.3	0.538	502	11271	0.992
news-test2008.eng-deu	23.4	0.530	2051	47427	1.000
newstest2009.eng-deu	22.6	0.532	2525	62816	1.000
newstest2010.eng-deu	25.5	0.548	2489	61511	0.967
newstest2011.eng-deu	22.8	0.526	3003	72981	0.993
newstest2012.eng-deu	23.4	0.527	3003	72886	0.970
newstest2013.eng-deu	27.3	0.555	3000	63737	0.984
newstest2015-ende.eng-deu	31.4	0.595	2169	44260	1.000
newstest2016-ende.eng-deu	37.5	0.640	2999	62670	0.994
newstest2017-ende.eng-deu	30.1	0.589	3004	61291	1.000
newstest2018-ende.eng-deu	45.4	0.690	2998	64276	1.000
newstest2019-ende.eng-deu	41.7	0.659	1997	48969	0.986
Tatoeba-test.eng-deu	45.3	0.651	10000	83347	0.991

opus-2021-02-22.zip

• dataset: opus
• model: transformer
• source language(s): eng
• target language(s): deu
• model: transformer
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opus-2021-02-22.zip
• test set translations: opus-2021-02-22.test.txt
• test set scores: opus-2021-02-22.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
newssyscomb2009.eng-deu	23.3	0.539	502	11271	0.990
news-test2008.eng-deu	23.9	0.533	2051	47427	1.000
newstest2009.eng-deu	22.7	0.533	2525	62816	0.999
newstest2010.eng-deu	25.9	0.550	2489	61511	0.966
newstest2011.eng-deu	22.9	0.528	3003	72981	0.993
newstest2012.eng-deu	23.8	0.530	3003	72886	0.972
newstest2013.eng-deu	27.6	0.558	3000	63737	0.983
newstest2014-deen.eng-deu	29.6	0.595	3003	62964	1.000
newstest2015-ende.eng-deu	32.0	0.601	2169	44260	1.000
newstest2016-ende.eng-deu	37.9	0.644	2999	62670	0.992
newstest2017-ende.eng-deu	30.6	0.593	3004	61291	1.000
newstest2018-ende.eng-deu	46.4	0.697	2998	64276	0.999
newstest2019-ende.eng-deu	42.4	0.664	1997	48969	0.990
Tatoeba-test.eng-deu	45.8	0.655	10000	83347	0.995

opus+bt-2021-04-13.zip

• dataset: opus+bt
• model: transformer-align
• source language(s): eng
• target language(s): deu
• model: transformer-align
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opus+bt-2021-04-13.zip
• test set translations: opus+bt-2021-04-13.test.txt
• test set scores: opus+bt-2021-04-13.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
newssyscomb2009.eng-deu	22.8	0.538	502	11271	0.991
news-test2008.eng-deu	23.7	0.533	2051	47427	1.000
newstest2009.eng-deu	22.6	0.532	2525	62816	0.999
newstest2010.eng-deu	25.5	0.552	2489	61511	0.962
newstest2011.eng-deu	22.6	0.527	3003	72981	0.988
newstest2012.eng-deu	23.4	0.530	3003	72886	0.970
newstest2013.eng-deu	27.1	0.556	3000	63737	0.979
newstest2014-deen.eng-deu	29.6	0.599	3003	62964	1.000
newstest2015-ende.eng-deu	31.6	0.600	2169	44260	1.000
newstest2016-ende.eng-deu	37.2	0.644	2999	62670	0.988
newstest2017-ende.eng-deu	30.6	0.595	3004	61291	1.000
newstest2018-ende.eng-deu	45.6	0.696	2998	64276	0.996
newstest2019-ende.eng-deu	41.3	0.659	1997	48969	0.991
Tatoeba-test.eng-deu	45.7	0.654	10000	83347	0.995

opusTCv20210807+bt-2021-12-08.zip

• dataset: opusTCv20210807+bt
• model: transformer-big
• source language(s): eng
• target language(s): deu
• raw source language(s): eng
• raw target language(s): deu
• model: transformer-big
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opusTCv20210807+bt-2021-12-08.zip
• test set translations: opusTCv20210807+bt-2021-12-08.test.txt
• test set scores: opusTCv20210807+bt-2021-12-08.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
newssyscomb2009.eng-deu	24.3	0.5462	502	11271	0.993
news-test2008.eng-deu	24.7	0.5412	2051	47427	1.000
newstest2009.eng-deu	23.6	0.5385	2525	62816	0.999
newstest2010.eng-deu	26.9	0.5589	2489	61511	0.966
newstest2011.eng-deu	24.1	0.5364	3003	72981	0.990
newstest2012.eng-deu	24.6	0.5375	3003	72886	0.972
newstest2013.eng-deu	28.3	0.5636	3000	63737	0.988
newstest2014-deen.eng-deu	30.9	0.6084	3003	62964	1.000
newstest2015-ende.eng-deu	33.2	0.6106	2169	44260	1.000
newstest2016-ende.eng-deu	39.8	0.6595	2999	62670	0.993
newstest2017-ende.eng-deu	32.0	0.6047	3004	61291	1.000
newstest2018-ende.eng-deu	48.8	0.7146	2998	64276	1.000
newstest2019-ende.eng-deu	45.0	0.6821	1997	48969	0.995
Tatoeba-test-v2021-08-07.eng-deu	43.7	0.6442	10000	85728	1.000