opusTCv20210807+bt-2021-11-08.yml

release: eng-fin/opusTCv20210807+bt-2021-11-08.zip
release-date: 2021-11-08
dataset-name: opusTCv20210807+bt
modeltype: transformer-big-align
vocabulary:
   source: opusTCv20210807+bt.spm32k-spm32k.vocab.yml
   target: opusTCv20210807+bt.spm32k-spm32k.vocab.yml
pre-processing: normalization + SentencePiece (spm32k,spm32k)
subwords:
   source: spm32k
   target: spm32k
subword-models:
   source: source.spm
   target: target.spm
source-languages:
   - eng
target-languages:
   - fin
raw-source-languages:
   - eng
raw-target-languages:
   - fin
training-data:
   eng-fin: Tatoeba-train-v2021-08-07 (103531791) wiki.aa.fin-eng (989793) wiki.ab.fin-eng (989863) wiki.ac.fin-eng (989739) wiki.ad.fin-eng (989710) wiki.ae.fin-eng (989743) wiki.af.fin-eng (989850) wiki.ag.fin-eng (989780) wiki.ah.fin-eng (479612) wikibooks.aa.fin-eng (65843) wikinews.aa.fin-eng (21283) wikiquote.aa.fin-eng (32622) wikisource.aa.fin-eng (273356) 
validation-data:
   eng-fin: Tatoeba-dev-v2021-08-07, 69896
   total-size-shuffled: 69896
   devset-selected: top 5000  lines of Tatoeba-dev-v2021-08-07.src.shuffled
test-data:
   newsdev2015-enfi.eng-fin: 1500/23375
   newstest2015-enfi.eng-fin: 1370/19968
   newstest2016-enfi.eng-fin: 3000/48116
   newstest2017-enfi.eng-fin: 3002/45718
   newstest2018-enfi.eng-fin: 3000/45475
   newstest2019-enfi.eng-fin: 1997/38369
   newstestB2016-enfi.eng-fin: 3000/45766
   Tatoeba-test-v2021-08-07.eng-fin: 10000/60797
BLEU-scores:
   newsdev2015-enfi.eng-fin: 24.6
   newstest2015-enfi.eng-fin: 26.5
   newstest2016-enfi.eng-fin: 28.2
   newstest2017-enfi.eng-fin: 31.4
   newstest2018-enfi.eng-fin: 20.7
   newstest2019-enfi.eng-fin: 26.5
   newstestB2016-enfi.eng-fin: 22.3
   Tatoeba-test-v2021-08-07.eng-fin: 39.9
chr-F-scores:
   newsdev2015-enfi.eng-fin: 0.580
   newstest2015-enfi.eng-fin: 0.595
   newstest2016-enfi.eng-fin: 0.607
   newstest2017-enfi.eng-fin: 0.636
   newstest2018-enfi.eng-fin: 0.553
   newstest2019-enfi.eng-fin: 0.581
   newstestB2016-enfi.eng-fin: 0.568
   Tatoeba-test-v2021-08-07.eng-fin: 0.643