Skip to content

Latest commit

 

History

History
96 lines (88 loc) · 4.72 KB

README.md

File metadata and controls

96 lines (88 loc) · 4.72 KB

opus-2020-07-06.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): akl_Latn ceb cha dtp fij gil haw hil iba ilo ind jav jav_Java lkt mad mah max_Latn min mlg mri nau niu pag pau rap smo sun tah tet tmw_Latn ton tvl war zlm_Latn zsm_Latn
  • 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-2020-07-06.zip
  • test set translations: opus-2020-07-06.test.txt
  • test set scores: opus-2020-07-06.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng-akl.eng.akl 2.0 0.087
Tatoeba-test.eng-ceb.eng.ceb 11.4 0.427
Tatoeba-test.eng-cha.eng.cha 2.6 0.193
Tatoeba-test.eng-dtp.eng.dtp 0.5 0.121
Tatoeba-test.eng-fij.eng.fij 25.2 0.426
Tatoeba-test.eng-gil.eng.gil 64.2 0.790
Tatoeba-test.eng-haw.eng.haw 0.8 0.114
Tatoeba-test.eng-hil.eng.hil 18.6 0.504
Tatoeba-test.eng-iba.eng.iba 14.1 0.389
Tatoeba-test.eng-ilo.eng.ilo 33.0 0.582
Tatoeba-test.eng-jav.eng.jav 7.2 0.297
Tatoeba-test.eng-lkt.eng.lkt 0.8 0.064
Tatoeba-test.eng-mad.eng.mad 1.1 0.136
Tatoeba-test.eng-mah.eng.mah 11.8 0.366
Tatoeba-test.eng-mlg.eng.mlg 34.3 0.526
Tatoeba-test.eng-mri.eng.mri 8.1 0.311
Tatoeba-test.eng-msa.eng.msa 30.7 0.563
Tatoeba-test.eng.multi 16.8 0.409
Tatoeba-test.eng-nau.eng.nau 0.4 0.082
Tatoeba-test.eng-niu.eng.niu 42.5 0.540
Tatoeba-test.eng-pag.eng.pag 16.4 0.522
Tatoeba-test.eng-pau.eng.pau 1.7 0.144
Tatoeba-test.eng-rap.eng.rap 2.7 0.090
Tatoeba-test.eng-smo.eng.smo 25.1 0.468
Tatoeba-test.eng-sun.eng.sun 28.5 0.421
Tatoeba-test.eng-tah.eng.tah 8.2 0.262
Tatoeba-test.eng-tet.eng.tet 1.4 0.129
Tatoeba-test.eng-ton.eng.ton 26.0 0.505
Tatoeba-test.eng-tvl.eng.tvl 29.9 0.523
Tatoeba-test.eng-war.eng.war 12.9 0.435

opus-2020-07-27.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): akl_Latn ceb cha dtp fij gil haw hil iba ilo ind jav jav_Java lkt mad mah max_Latn min mlg mri nau niu pag pau rap smo sun tah tet tmw_Latn ton tvl war zlm_Latn zsm_Latn
  • 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-2020-07-27.zip
  • test set translations: opus-2020-07-27.test.txt
  • test set scores: opus-2020-07-27.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng-akl.eng.akl 2.2 0.103
Tatoeba-test.eng-ceb.eng.ceb 10.7 0.425
Tatoeba-test.eng-cha.eng.cha 3.2 0.201
Tatoeba-test.eng-dtp.eng.dtp 0.5 0.120
Tatoeba-test.eng-fij.eng.fij 26.8 0.453
Tatoeba-test.eng-gil.eng.gil 59.3 0.762
Tatoeba-test.eng-haw.eng.haw 1.0 0.116
Tatoeba-test.eng-hil.eng.hil 19.0 0.517
Tatoeba-test.eng-iba.eng.iba 15.5 0.400
Tatoeba-test.eng-ilo.eng.ilo 33.6 0.591
Tatoeba-test.eng-jav.eng.jav 7.8 0.301
Tatoeba-test.eng-lkt.eng.lkt 1.0 0.064
Tatoeba-test.eng-mad.eng.mad 1.1 0.142
Tatoeba-test.eng-mah.eng.mah 9.1 0.374
Tatoeba-test.eng-mlg.eng.mlg 35.4 0.526
Tatoeba-test.eng-mri.eng.mri 7.6 0.309
Tatoeba-test.eng-msa.eng.msa 31.1 0.565
Tatoeba-test.eng.multi 17.6 0.411
Tatoeba-test.eng-nau.eng.nau 1.4 0.098
Tatoeba-test.eng-niu.eng.niu 40.1 0.560
Tatoeba-test.eng-pag.eng.pag 16.8 0.526
Tatoeba-test.eng-pau.eng.pau 1.9 0.139
Tatoeba-test.eng-rap.eng.rap 2.7 0.090
Tatoeba-test.eng-smo.eng.smo 24.9 0.453
Tatoeba-test.eng-sun.eng.sun 33.2 0.439
Tatoeba-test.eng-tah.eng.tah 12.5 0.278
Tatoeba-test.eng-tet.eng.tet 1.6 0.140
Tatoeba-test.eng-ton.eng.ton 25.8 0.530
Tatoeba-test.eng-tvl.eng.tvl 31.1 0.523
Tatoeba-test.eng-war.eng.war 12.8 0.436