opus-2020-07-26.yml

release: eng-gmq/opus-2020-07-26.zip
release-date: 2020-07-26
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:
   - eng
target-languages:
   - dan
   - fao
   - isl
   - nno
   - nob
   - non
   - swe
use-target-labels:
   - ">>dan<<"
   - ">>fao<<"
   - ">>isl<<"
   - ">>nno<<"
   - ">>nob<<"
   - ">>nob_Hebr<<"
   - ">>non_Latn<<"
   - ">>swe<<"
test-data:
   Tatoeba-test.eng-dan: 10000/61469
   Tatoeba-test.eng-fao: 294/1564
   Tatoeba-test.eng-isl: 2500/16288
   Tatoeba-test.eng-multi: 10000/60550
   Tatoeba-test.eng-non: 15/109
   Tatoeba-test.eng-nor: 5000/33820
   Tatoeba-test.eng-swe: 10000/54832
BLEU-scores:
   Tatoeba-test.eng-dan: 57.0
   Tatoeba-test.eng-fao: 7.0
   Tatoeba-test.eng-isl: 23.3
   Tatoeba-test.eng-multi: 52.3
   Tatoeba-test.eng-non: 0.7
   Tatoeba-test.eng-nor: 49.6
   Tatoeba-test.eng-swe: 56.9
chr-F-scores:
   Tatoeba-test.eng-dan: 0.719
   Tatoeba-test.eng-fao: 0.311
   Tatoeba-test.eng-isl: 0.500
   Tatoeba-test.eng-multi: 0.683
   Tatoeba-test.eng-non: 0.196
   Tatoeba-test.eng-nor: 0.671
   Tatoeba-test.eng-swe: 0.711