opus-2020-06-28.yml

release: zle-eng/opus-2020-06-28.zip
release-date: 2020-06-28
dataset-name: opus
modeltype: transformer
pre-processing: normalization + SentencePiece (spm32k,spm32k)
subwords:
   - source: spm32k
   - target: spm32k
subword-models:
   - source: source.spm
   - target: target.spm
source-languages:
   - bel
   - orv
   - rue
   - rus
   - ukr
target-languages:
   - eng
test-data:
   - Tatoeba-test.bel-eng: 2500/15535
   - Tatoeba-test.multi-eng: 0/0
   - Tatoeba-test.orv-eng: 322/1713
   - Tatoeba-test.rue-eng: 120/566
   - Tatoeba-test.rus-eng: 10000/61769
   - Tatoeba-test.ukr-eng: 10000/55352
   - newstest2012.rus-eng: 3003/63778
   - newstest2013.rus-eng: 3000/56089
   - newstest2014-ruen.rus-eng: 3003/61219
   - newstest2015-enru.rus-eng: 2818/56934
   - newstest2016-enru.rus-eng: 2998/61076
   - newstest2017-enru.rus-eng: 3001/60714
   - newstest2018-enru.rus-eng: 3000/62925
   - newstest2019-ruen.rus-eng: 2000/37348
BLEU-scores:
   - Tatoeba-test.bel-eng: 35.6
   - Tatoeba-test.multi-eng: 47.3
   - Tatoeba-test.orv-eng: 4.9
   - Tatoeba-test.rue-eng: 12.5
   - Tatoeba-test.rus-eng: 49.9
   - Tatoeba-test.ukr-eng: 49.5
   - newstest2012.rus-eng: 29.5
   - newstest2013.rus-eng: 23.8
   - newstest2014-ruen.rus-eng: 26.1
   - newstest2015-enru.rus-eng: 25.0
   - newstest2016-enru.rus-eng: 24.5
   - newstest2017-enru.rus-eng: 27.1
   - newstest2018-enru.rus-eng: 23.8
   - newstest2019-ruen.rus-eng: 25.5
chr-F-scores:
   - Tatoeba-test.bel-eng: 0.541
   - Tatoeba-test.multi-eng: 0.633
   - Tatoeba-test.orv-eng: 0.213
   - Tatoeba-test.rue-eng: 0.307
   - Tatoeba-test.rus-eng: 0.654
   - Tatoeba-test.ukr-eng: 0.654
   - newstest2012.rus-eng: 0.565
   - newstest2013.rus-eng: 0.509
   - newstest2014-ruen.rus-eng: 0.547
   - newstest2015-enru.rus-eng: 0.523
   - newstest2016-enru.rus-eng: 0.521
   - newstest2017-enru.rus-eng: 0.540
   - newstest2018-enru.rus-eng: 0.520
   - newstest2019-ruen.rus-eng: 0.528