Skip to content

Latest commit

 

History

History

eng-zho

README.md

opus-2020-07-14.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): cjy_Hans cjy_Hant cmn cmn_Hans cmn_Hant gan lzh lzh_Hans nan wuu yue yue_Hans yue_Hant
  • 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-2020-07-14.zip
  • test set translations: opus-2020-07-14.test.txt
  • test set scores: opus-2020-07-14.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng.zho 31.1 0.265

opus-2020-07-17.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): cjy_Hans cjy_Hant cmn cmn_Hans cmn_Hant gan lzh lzh_Hans nan wuu yue yue_Hans yue_Hant
  • 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-2020-07-17.zip
  • test set translations: opus-2020-07-17.test.txt
  • test set scores: opus-2020-07-17.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng.zho 31.4 0.268

opus-2021-02-23.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): cjy cmn gan hak hsn lzh nan wuu yue
  • 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: >>cmn_Hans<< >>cmn_Hant<< >>cmn<< >>yue_Hant<< >>yue_Hans<< >>nan<< >>wuu<<
  • download: opus-2021-02-23.zip
  • test set translations: opus-2021-02-23.test.txt
  • test set scores: opus-2021-02-23.eval.txt

Benchmarks

testset BLEU chr-F #sent #words BP
Tatoeba-test.eng-zho 31.6 0.267 9999 110463 0.911

opus+bt-2021-04-19.zip

  • dataset: opus+bt
  • model: transformer-align
  • source language(s): eng
  • target language(s): cjy cmn gan hak hsn lzh nan wuu yue
  • model: transformer-align
  • 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: >>cmn_Hans<< >>cmn_Hant<< >>cmn<< >>yue_Hant<< >>yue_Hans<< >>nan<< >>wuu<<
  • download: opus+bt-2021-04-19.zip
  • test set translations: opus+bt-2021-04-19.test.txt
  • test set scores: opus+bt-2021-04-19.eval.txt

Benchmarks

testset BLEU chr-F #sent #words BP
Tatoeba-test.eng-cjy_Hans 27.8 0.262 2 21 0.900
Tatoeba-test.eng-cjy_Hant 1.1 0.000 1 4 1.000
Tatoeba-test.eng-cmn 10.0 0.300 15 87 1.000
Tatoeba-test.eng-cmn_Hans 34.7 0.293 4195 48021 0.971
Tatoeba-test.eng-cmn_Hant 34.4 0.294 4418 46882 0.871
Tatoeba-test.eng-gan 18.0 0.197 1 14 1.000
Tatoeba-test.eng-lzh 0.3 0.024 98 702 1.000
Tatoeba-test.eng-lzh_Hans 0.8 0.026 3 38 1.000
Tatoeba-test.eng-nan 0.8 0.007 2 10 1.000
Tatoeba-test.eng-wuu 5.1 0.061 203 2335 1.000
Tatoeba-test.eng-yue_Hans 5.6 0.077 630 8155 0.854
Tatoeba-test.eng-yue_Hant 9.0 0.096 431 4183 0.921
Tatoeba-test.eng-zho 30.9 0.259 9999 110463 0.953

opusTCv20210807+bt_transformer-big_2022-05-14.zip

Benchmarks

testset BLEU chr-F #sent #words BP
Tatoeba-test-v2021-08-07.eng-cjy_Hans 12.9 0.13957 2 21 0.951
Tatoeba-test-v2021-08-07.eng-cjy_Hant 16.0 6.250 1 4 1.000
Tatoeba-test-v2021-08-07.eng-cmn 15.5 0.35677 15 87 1.000
Tatoeba-test-v2021-08-07.eng-cmn_Hans 34.3 0.29125 4450 50615 0.960
Tatoeba-test-v2021-08-07.eng-cmn_Hant 35.3 0.30103 4476 47363 0.897
Tatoeba-test-v2021-08-07.eng-gan 11.8 0.16078 1 14 0.926
Tatoeba-test-v2021-08-07.eng-lzh 1.3 4.033 100 776 1.000
Tatoeba-test-v2021-08-07.eng-lzh_Hans 0.8 1.515 3 38 1.000
Tatoeba-test-v2021-08-07.eng-multi 31.1 0.26573 10000 110339 0.938
Tatoeba-test-v2021-08-07.eng-nan 5.3 3.125 2 10 1.000
Tatoeba-test-v2021-08-07.eng-wuu 8.0 9.213 208 2375 1.000
Tatoeba-test-v2021-08-07.eng-yue_Hans 5.4 7.265 677 8754 0.897
Tatoeba-test-v2021-08-07.eng-yue_Hant 8.5 9.720 454 4437 0.985