opus2m-2020-08-01.yml

release: sem-eng/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:
   - acm
   - afb
   - amh
   - apc
   - ara
   - arq
   - ary
   - arz
   - heb
   - mlt
   - tir
target-languages:
   - eng
test-data:
   Tatoeba-test.amh-eng: 190/802
   Tatoeba-test.ara-eng: 10000/62449
   Tatoeba-test.heb-eng: 10000/62332
   Tatoeba-test.mlt-eng: 203/935
   Tatoeba-test.multi-eng: 10000/61612
   Tatoeba-test.tir-eng: 69/426
BLEU-scores:
   Tatoeba-test.amh-eng: 37.5
   Tatoeba-test.ara-eng: 38.9
   Tatoeba-test.heb-eng: 44.6
   Tatoeba-test.mlt-eng: 53.7
   Tatoeba-test.multi-eng: 41.7
   Tatoeba-test.tir-eng: 18.3
chr-F-scores:
   Tatoeba-test.amh-eng: 0.565
   Tatoeba-test.ara-eng: 0.566
   Tatoeba-test.heb-eng: 0.610
   Tatoeba-test.mlt-eng: 0.688
   Tatoeba-test.multi-eng: 0.588
   Tatoeba-test.tir-eng: 0.370