Skip to content

Latest commit

 

History

History
 
 

zle-fiu

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
opus-2021-02-18.yml
opus-2021-02-18.yml
 
 

README.md

opus-2021-02-10.zip

  • dataset: opus
  • model: transformer
  • source language(s): bel orv rue rus ukr
  • target language(s): chm est fin fkv hun izh kom krl liv mdf myv olo sma sme udm vep
  • 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-2021-02-10.zip
  • test set translations: opus-2021-02-10.test.txt
  • test set scores: opus-2021-02-10.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.bel-est.bel.est 3.7 0.437
Tatoeba-test.bel-hun.bel.hun 54.8 0.771
Tatoeba-test.multi.multi 26.5 0.464
Tatoeba-test.rus-chm.rus.chm 1.3 0.155
Tatoeba-test.rus-est.rus.est 54.4 0.716
Tatoeba-test.rus-fin.rus.fin 36.9 0.611
Tatoeba-test.rus-hun.rus.hun 34.9 0.582
Tatoeba-test.rus-izh.rus.izh 0.7 0.047
Tatoeba-test.rus-udm.rus.udm 2.7 0.226
Tatoeba-test.ukr-est.ukr.est 35.8 0.687
Tatoeba-test.ukr-fin.ukr.fin 42.8 0.653
Tatoeba-test.ukr-hun.ukr.hun 40.5 0.647

opus-2021-02-17.zip

  • dataset: opus
  • model: transformer
  • source language(s): bel rus ukr
  • target language(s): est fin hun izh mhr udm
  • 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: >>hun<< >>est<< >>fin<< >>udm<< >>mhr<<
  • 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.multi-multi 26.5 0.464 10000 59597 0.926

opus-2021-02-18.zip

  • dataset: opus
  • model: transformer
  • source language(s): bel rus ukr
  • target language(s): est fin hun izh mhr udm
  • 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: >>hun<< >>est<< >>fin<< >>udm<< >>mhr<<
  • download: opus-2021-02-18.zip
  • test set translations: opus-2021-02-18.test.txt
  • test set scores: opus-2021-02-18.eval.txt

Benchmarks

testset BLEU chr-F #sent #words BP
Tatoeba-test.bel-est 3.7 0.437 2 31 1.000
Tatoeba-test.bel-hun 54.8 0.771 3 16 1.000
Tatoeba-test.multi-multi 26.5 0.464 10000 59597 0.926
Tatoeba-test.rus-chm 0.9 0.154 2750 18927 0.862
Tatoeba-test.rus-est 54.0 0.714 598 3572 0.971
Tatoeba-test.rus-fin 36.8 0.610 3643 19307 0.926
Tatoeba-test.rus-hun 34.8 0.581 2500 15042 0.971
Tatoeba-test.rus-izh 0.7 0.047 8 33 1.000
Tatoeba-test.rus-udm 2.7 0.224 54 300 1.000
Tatoeba-test.ukr-est 35.8 0.687 16 85 0.988
Tatoeba-test.ukr-fin 41.8 0.640 32 186 0.984
Tatoeba-test.ukr-hun 40.5 0.647 464 2426 0.993