opus-2020-07-26.yml

release: eng-dra/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:
   - kan
   - mal
   - tam
   - tel
use-target-labels:
   - ">>kan<<"
   - ">>mal<<"
   - ">>tam<<"
   - ">>tel<<"
test-data:
   Tatoeba-test.eng-kan: 167/783
   Tatoeba-test.eng-mal: 802/3459
   Tatoeba-test.eng-multi: 1541/6691
   Tatoeba-test.eng-tam: 311/1364
   Tatoeba-test.eng-tel: 261/1085
BLEU-scores:
   Tatoeba-test.eng-kan: 4.7
   Tatoeba-test.eng-mal: 13.1
   Tatoeba-test.eng-multi: 10.7
   Tatoeba-test.eng-tam: 9.0
   Tatoeba-test.eng-tel: 7.1
chr-F-scores:
   Tatoeba-test.eng-kan: 0.348
   Tatoeba-test.eng-mal: 0.515
   Tatoeba-test.eng-multi: 0.463
   Tatoeba-test.eng-tam: 0.444
   Tatoeba-test.eng-tel: 0.363