opus2m-2020-08-01.yml

release: eng-pqe/opus2m-2020-08-01.zip
release-date: 2020-08-01
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:
   - eng
target-languages:
   - fij
   - gil
   - haw
   - lkt
   - mah
   - mri
   - nau
   - niu
   - rap
   - smo
   - tah
   - ton
   - tvl
use-target-labels:
   - ">>fij<<"
   - ">>gil<<"
   - ">>haw<<"
   - ">>lkt<<"
   - ">>mah<<"
   - ">>mri<<"
   - ">>nau<<"
   - ">>niu<<"
   - ">>rap<<"
   - ">>smo<<"
   - ">>tah<<"
   - ">>ton<<"
   - ">>tvl<<"
test-data:
   Tatoeba-test.eng-fij: 44/122
   Tatoeba-test.eng-gil: 14/60
   Tatoeba-test.eng-haw: 92/353
   Tatoeba-test.eng-lkt: 21/69
   Tatoeba-test.eng-mah: 29/115
   Tatoeba-test.eng-mri: 363/3300
   Tatoeba-test.eng-multi: 772/4786
   Tatoeba-test.eng-nau: 20/60
   Tatoeba-test.eng-niu: 29/111
   Tatoeba-test.eng-rap: 26/85
   Tatoeba-test.eng-smo: 78/325
   Tatoeba-test.eng-tah: 21/59
   Tatoeba-test.eng-ton: 20/65
   Tatoeba-test.eng-tvl: 15/62
BLEU-scores:
   Tatoeba-test.eng-fij: 22.1
   Tatoeba-test.eng-gil: 41.9
   Tatoeba-test.eng-haw: 0.6
   Tatoeba-test.eng-lkt: 0.5
   Tatoeba-test.eng-mah: 9.7
   Tatoeba-test.eng-mri: 7.7
   Tatoeba-test.eng-multi: 11.3
   Tatoeba-test.eng-nau: 0.5
   Tatoeba-test.eng-niu: 42.5
   Tatoeba-test.eng-rap: 3.3
   Tatoeba-test.eng-smo: 27.0
   Tatoeba-test.eng-tah: 11.3
   Tatoeba-test.eng-ton: 27.0
   Tatoeba-test.eng-tvl: 29.3
chr-F-scores:
   Tatoeba-test.eng-fij: 0.396
   Tatoeba-test.eng-gil: 0.673
   Tatoeba-test.eng-haw: 0.114
   Tatoeba-test.eng-lkt: 0.075
   Tatoeba-test.eng-mah: 0.386
   Tatoeba-test.eng-mri: 0.301
   Tatoeba-test.eng-multi: 0.306
   Tatoeba-test.eng-nau: 0.071
   Tatoeba-test.eng-niu: 0.560
   Tatoeba-test.eng-rap: 0.122
   Tatoeba-test.eng-smo: 0.462
   Tatoeba-test.eng-tah: 0.307
   Tatoeba-test.eng-ton: 0.528
   Tatoeba-test.eng-tvl: 0.513