opus2m-2020-07-31.yml

release: cel-eng/opus2m-2020-07-31.zip
release-date: 2020-07-31
dataset-name: opus2m
modeltype: transformer
pre-processing: normalization + SentencePiece (spm32k,spm32k)
subwords:
   source: spm32k
   target: spm32k
subword-models:
   source: source.spm
   target: target.spm
source-languages:
   - bre
   - cor
   - cym
   - gla
   - gle
   - glv
target-languages:
   - eng
test-data:
   Tatoeba-test.bre-eng: 383/1644
   Tatoeba-test.cor-eng: 3198/13500
   Tatoeba-test.cym-eng: 818/4618
   Tatoeba-test.gla-eng: 917/5130
   Tatoeba-test.gle-eng: 1924/9187
   Tatoeba-test.glv-eng: 5/22
   Tatoeba-test.multi-eng: 7245/34101
BLEU-scores:
   Tatoeba-test.bre-eng: 17.2
   Tatoeba-test.cor-eng: 3.0
   Tatoeba-test.cym-eng: 41.5
   Tatoeba-test.gla-eng: 15.4
   Tatoeba-test.gle-eng: 50.8
   Tatoeba-test.glv-eng: 11.0
   Tatoeba-test.multi-eng: 22.8
chr-F-scores:
   Tatoeba-test.bre-eng: 0.385
   Tatoeba-test.cor-eng: 0.172
   Tatoeba-test.cym-eng: 0.582
   Tatoeba-test.gla-eng: 0.330
   Tatoeba-test.gle-eng: 0.668
   Tatoeba-test.glv-eng: 0.297
   Tatoeba-test.multi-eng: 0.398