opus-2020-10-04.yml

release: gmq-gmq/opus-2020-10-04.zip
release-date: 2020-10-04
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:
   - dan
   - eng
   - fao
   - isl
   - nno
   - nob
   - non
   - swe
target-languages:
   - dan
   - eng
   - fao
   - isl
   - nno
   - nob
   - non
   - swe
use-target-labels:
   - >>dan<<
   - >>eng<<
   - >>fao<<
   - >>isl<<
   - >>nno<<
   - >>nob<<
   - >>nob_Hebr<<
   - >>non_Latn<<
   - >>swe<<
test-data:
   - Tatoeba-test.dan-eng: 10000/62015
   - Tatoeba-test.dan-fao: 1/5
   - Tatoeba-test.dan-isl: 1/8
   - Tatoeba-test.dan-nor: 1311/8216
   - Tatoeba-test.dan-swe: 1550/8385
   - Tatoeba-test.eng-dan: 10000/61469
   - Tatoeba-test.eng-fao: 294/1564
   - Tatoeba-test.eng-isl: 2500/16288
   - Tatoeba-test.eng-non: 15/109
   - Tatoeba-test.eng-nor: 5000/33820
   - Tatoeba-test.eng-swe: 10000/54832
   - Tatoeba-test.fao-dan: 1/5
   - Tatoeba-test.fao-eng: 294/1631
   - Tatoeba-test.fao-isl: 2/5
   - Tatoeba-test.fao-nor: 21/102
   - Tatoeba-test.fao-swe: 1/2
   - Tatoeba-test.isl-dan: 1/10
   - Tatoeba-test.isl-eng: 2500/16860
   - Tatoeba-test.isl-fao: 2/5
   - Tatoeba-test.isl-nor: 126/775
   - Tatoeba-test.isl-swe: 1/10
   - Tatoeba-test.multi-multi: 10000/60854
   - Tatoeba-test.non-eng: 15/112
   - Tatoeba-test.nor-dan: 1311/8343
   - Tatoeba-test.nor-eng: 5000/34479
   - Tatoeba-test.nor-fao: 21/101
   - Tatoeba-test.nor-isl: 126/739
   - Tatoeba-test.nor-nor: 948/5317
   - Tatoeba-test.nor-swe: 562/3035
   - Tatoeba-test.swe-dan: 1550/8503
   - Tatoeba-test.swe-eng: 10000/54997
   - Tatoeba-test.swe-fao: 1/2
   - Tatoeba-test.swe-isl: 1/8
   - Tatoeba-test.swe-nor: 562/3039
BLEU-scores:
   - Tatoeba-test.dan-eng: 59.2
   - Tatoeba-test.dan-fao: 6.6
   - Tatoeba-test.dan-isl: 46.7
   - Tatoeba-test.dan-nor: 74.1
   - Tatoeba-test.dan-swe: 69.3
   - Tatoeba-test.eng-dan: 57.1
   - Tatoeba-test.eng-fao: 8.9
   - Tatoeba-test.eng-isl: 24.3
   - Tatoeba-test.eng-non: 0.4
   - Tatoeba-test.eng-nor: 51.9
   - Tatoeba-test.eng-swe: 57.3
   - Tatoeba-test.fao-dan: 23.0
   - Tatoeba-test.fao-eng: 25.4
   - Tatoeba-test.fao-isl: 15.2
   - Tatoeba-test.fao-nor: 43.9
   - Tatoeba-test.fao-swe: 0.0
   - Tatoeba-test.isl-dan: 77.7
   - Tatoeba-test.isl-eng: 49.7
   - Tatoeba-test.isl-fao: 14.5
   - Tatoeba-test.isl-nor: 61.9
   - Tatoeba-test.isl-swe: 60.0
   - Tatoeba-test.multi-multi: 57.1
   - Tatoeba-test.non-eng: 30.0
   - Tatoeba-test.nor-dan: 71.1
   - Tatoeba-test.nor-eng: 54.2
   - Tatoeba-test.nor-fao: 9.9
   - Tatoeba-test.nor-isl: 29.3
   - Tatoeba-test.nor-nor: 61.0
   - Tatoeba-test.nor-swe: 72.5
   - Tatoeba-test.swe-dan: 68.7
   - Tatoeba-test.swe-eng: 61.1
   - Tatoeba-test.swe-fao: 19.0
   - Tatoeba-test.swe-isl: 46.7
   - Tatoeba-test.swe-nor: 72.9
chr-F-scores:
   - Tatoeba-test.dan-eng: 0.738
   - Tatoeba-test.dan-fao: 0.171
   - Tatoeba-test.dan-isl: 0.758
   - Tatoeba-test.dan-nor: 0.849
   - Tatoeba-test.dan-swe: 0.818
   - Tatoeba-test.eng-dan: 0.719
   - Tatoeba-test.eng-fao: 0.315
   - Tatoeba-test.eng-isl: 0.506
   - Tatoeba-test.eng-non: 0.182
   - Tatoeba-test.eng-nor: 0.682
   - Tatoeba-test.eng-swe: 0.712
   - Tatoeba-test.fao-dan: 0.536
   - Tatoeba-test.fao-eng: 0.453
   - Tatoeba-test.fao-isl: 0.304
   - Tatoeba-test.fao-nor: 0.602
   - Tatoeba-test.fao-swe: 1.000
   - Tatoeba-test.isl-dan: 0.901
   - Tatoeba-test.isl-eng: 0.660
   - Tatoeba-test.isl-fao: 0.207
   - Tatoeba-test.isl-nor: 0.772
   - Tatoeba-test.isl-swe: 0.800
   - Tatoeba-test.multi-multi: 0.717
   - Tatoeba-test.non-eng: 0.523
   - Tatoeba-test.nor-dan: 0.831
   - Tatoeba-test.nor-eng: 0.695
   - Tatoeba-test.nor-fao: 0.343
   - Tatoeba-test.nor-isl: 0.543
   - Tatoeba-test.nor-nor: 0.780
   - Tatoeba-test.nor-swe: 0.839
   - Tatoeba-test.swe-dan: 0.808
   - Tatoeba-test.swe-eng: 0.738
   - Tatoeba-test.swe-fao: 0.164
   - Tatoeba-test.swe-isl: 0.784
   - Tatoeba-test.swe-nor: 0.835