Skip to content

Latest commit

 

History

History
124 lines (108 loc) · 5.7 KB

README.md

File metadata and controls

124 lines (108 loc) · 5.7 KB

opus-2020-06-28.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): kdx kin lin lug nya run sna swh toi_Latn tso umb xho zul
  • 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-2020-06-28.zip
  • test set translations: opus-2020-06-28.test.txt
  • test set scores: opus-2020-06-28.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng-kdx.eng.kdx 2.8 0.266
Tatoeba-test.eng-kin.eng.kin 5.1 0.522
Tatoeba-test.eng-lin.eng.lin 1.0 0.267
Tatoeba-test.eng-lug.eng.lug 15.2 0.576
Tatoeba-test.eng.multi 13.6 0.490
Tatoeba-test.eng-nya.eng.nya 16.2 0.600
Tatoeba-test.eng-run.eng.run 12.6 0.476
Tatoeba-test.eng-sna.eng.sna 24.9 0.633
Tatoeba-test.eng-swa.eng.swa 1.5 0.149
Tatoeba-test.eng-toi.eng.toi 8.3 0.210
Tatoeba-test.eng-tso.eng.tso 41.3 0.698
Tatoeba-test.eng-umb.eng.umb 4.6 0.349
Tatoeba-test.eng-xho.eng.xho 29.1 0.619
Tatoeba-test.eng-zul.eng.zul 30.4 0.749

opus-2020-07-06.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): kin lin lug nya run sna swh toi_Latn tso umb xho zul
  • 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-kin.eng.kin 8.1 0.596
Tatoeba-test.eng-lin.eng.lin 1.3 0.276
Tatoeba-test.eng-lug.eng.lug 4.8 0.503
Tatoeba-test.eng.multi 13.7 0.491
Tatoeba-test.eng-nya.eng.nya 20.1 0.623
Tatoeba-test.eng-run.eng.run 13.0 0.478
Tatoeba-test.eng-sna.eng.sna 29.6 0.618
Tatoeba-test.eng-swa.eng.swa 1.3 0.156
Tatoeba-test.eng-toi.eng.toi 14.1 0.290
Tatoeba-test.eng-tso.eng.tso 32.9 0.607
Tatoeba-test.eng-umb.eng.umb 4.0 0.330
Tatoeba-test.eng-xho.eng.xho 23.4 0.600
Tatoeba-test.eng-zul.eng.zul 33.0 0.725

opus-2020-07-14.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): kin lin lug nya run sna swh toi_Latn tso umb xho zul
  • 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-kin.eng.kin 10.2 0.540
Tatoeba-test.eng-lin.eng.lin 1.1 0.275
Tatoeba-test.eng-lug.eng.lug 5.1 0.433
Tatoeba-test.eng.multi 12.0 0.444
Tatoeba-test.eng-nya.eng.nya 25.7 0.621
Tatoeba-test.eng-run.eng.run 13.2 0.487
Tatoeba-test.eng-sna.eng.sna 32.3 0.652
Tatoeba-test.eng-toi.eng.toi 10.7 0.255
Tatoeba-test.eng-tso.eng.tso 41.3 0.698
Tatoeba-test.eng-umb.eng.umb 4.4 0.329
Tatoeba-test.eng-xho.eng.xho 24.9 0.613
Tatoeba-test.eng-zul.eng.zul 35.7 0.753

opus-2020-07-26.zip

  • dataset: opus
  • model: transformer
  • source language(s): eng
  • target language(s): kin lin lug nya run sna swh toi_Latn tso umb xho zul
  • 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-2020-07-26.zip
  • test set translations: opus-2020-07-26.test.txt
  • test set scores: opus-2020-07-26.eval.txt

Benchmarks

testset BLEU chr-F
Tatoeba-test.eng-kin.eng.kin 12.5 0.519
Tatoeba-test.eng-lin.eng.lin 1.1 0.277
Tatoeba-test.eng-lug.eng.lug 4.8 0.415
Tatoeba-test.eng.multi 12.1 0.449
Tatoeba-test.eng-nya.eng.nya 22.1 0.616
Tatoeba-test.eng-run.eng.run 13.2 0.492
Tatoeba-test.eng-sna.eng.sna 32.1 0.669
Tatoeba-test.eng-swa.eng.swa 1.7 0.180
Tatoeba-test.eng-toi.eng.toi 10.7 0.266
Tatoeba-test.eng-tso.eng.tso 26.9 0.631
Tatoeba-test.eng-umb.eng.umb 5.2 0.295
Tatoeba-test.eng-xho.eng.xho 22.6 0.615
Tatoeba-test.eng-zul.eng.zul 41.1 0.769