opus-2020-06-28.yml

release: cel-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:
   - 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: 0/0
BLEU-scores:
   - Tatoeba-test.bre-eng: 15.2
   - Tatoeba-test.cor-eng: 3.1
   - Tatoeba-test.cym-eng: 38.3
   - Tatoeba-test.gla-eng: 14.7
   - Tatoeba-test.gle-eng: 50.0
   - Tatoeba-test.glv-eng: 30.0
   - Tatoeba-test.multi-eng: 22.1
chr-F-scores:
   - Tatoeba-test.bre-eng: 0.362
   - Tatoeba-test.cor-eng: 0.175
   - Tatoeba-test.cym-eng: 0.564
   - Tatoeba-test.gla-eng: 0.318
   - Tatoeba-test.gle-eng: 0.660
   - Tatoeba-test.glv-eng: 0.466
   - Tatoeba-test.multi-eng: 0.391