opus2m-2020-08-01.yml

release: ine-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:
   - afr
   - aln
   - ang
   - arg
   - asm
   - ast
   - awa
   - bel
   - ben
   - bho
   - bos
   - bre
   - bul
   - cat
   - ces
   - cor
   - cos
   - csb
   - cym
   - dan
   - deu
   - dsb
   - egl
   - ell
   - enm
   - ext
   - fao
   - fra
   - frm
   - frr
   - fry
   - gcf
   - gla
   - gle
   - glg
   - glv
   - gom
   - gos
   - got
   - grc
   - gsw
   - guj
   - hat
   - hif
   - hin
   - hrv
   - hsb
   - hye
   - ind
   - isl
   - ita
   - jdt
   - ksh
   - kur
   - lad
   - lat
   - lav
   - lij
   - lit
   - lld
   - lmo
   - ltg
   - ltz
   - mai
   - mar
   - max
   - mfe
   - min
   - mkd
   - mwl
   - nds
   - nld
   - nno
   - nob
   - non
   - npi
   - oci
   - ori
   - orv
   - oss
   - pan
   - pap
   - pdc
   - pes
   - pms
   - pnb
   - pol
   - por
   - prg
   - pus
   - roh
   - rom
   - ron
   - rue
   - rus
   - san
   - scn
   - sco
   - sgs
   - sin
   - slv
   - snd
   - spa
   - sqi
   - srp
   - stq
   - swe
   - swg
   - tgk
   - tly
   - tmw
   - ukr
   - urd
   - vec
   - wln
   - yid
   - zlm
   - zsm
   - zza
target-languages:
   - eng
test-data:
   Tatoeba-test.afr-eng: 1374/8101
   Tatoeba-test.ang-eng: 189/1693
   Tatoeba-test.arg-eng: 105/343
   Tatoeba-test.asm-eng: 117/573
   Tatoeba-test.ast-eng: 99/682
   Tatoeba-test.awa-eng: 279/1042
   Tatoeba-test.bel-eng: 2500/15535
   Tatoeba-test.ben-eng: 2500/11321
   Tatoeba-test.bho-eng: 42/229
   Tatoeba-test.bre-eng: 383/1644
   Tatoeba-test.bul-eng: 10000/60705
   Tatoeba-test.cat-eng: 1631/10625
   Tatoeba-test.ces-eng: 10000/64235
   Tatoeba-test.cor-eng: 3198/13500
   Tatoeba-test.cos-eng: 5/28
   Tatoeba-test.csb-eng: 27/197
   Tatoeba-test.cym-eng: 818/4618
   Tatoeba-test.dan-eng: 10000/62015
   Tatoeba-test.deu-eng: 10000/69484
   Tatoeba-test.dsb-eng: 34/168
   Tatoeba-test.egl-eng: 84/348
   Tatoeba-test.ell-eng: 10000/50149
   Tatoeba-test.enm-eng: 49/238
   Tatoeba-test.ext-eng: 69/279
   Tatoeba-test.fao-eng: 294/1631
   Tatoeba-test.fas-eng: 3768/26936
   Tatoeba-test.fra-eng: 10000/65688
   Tatoeba-test.frm-eng: 18/183
   Tatoeba-test.frr-eng: 2/7
   Tatoeba-test.fry-eng: 205/1234
   Tatoeba-test.gcf-eng: 99/455
   Tatoeba-test.gla-eng: 917/5130
   Tatoeba-test.gle-eng: 1924/9187
   Tatoeba-test.glg-eng: 1008/7171
   Tatoeba-test.glv-eng: 5/22
   Tatoeba-test.gos-eng: 1152/4319
   Tatoeba-test.got-eng: 207/865
   Tatoeba-test.grc-eng: 592/4157
   Tatoeba-test.gsw-eng: 205/770
   Tatoeba-test.guj-eng: 154/767
   Tatoeba-test.hat-eng: 64/281
   Tatoeba-test.hbs-eng: 10000/57608
   Tatoeba-test.hif-eng: 36/187
   Tatoeba-test.hin-eng: 5000/28315
   Tatoeba-test.hsb-eng: 40/194
   Tatoeba-test.hye-eng: 1121/3888
   Tatoeba-test.isl-eng: 2500/16860
   Tatoeba-test.ita-eng: 10000/56443
   Tatoeba-test.jdt-eng: 2/6
   Tatoeba-test.kok-eng: 1/6
   Tatoeba-test.ksh-eng: 26/175
   Tatoeba-test.kur-eng: 296/1404
   Tatoeba-test.lad-eng: 629/2774
   Tatoeba-test.lah-eng: 32/162
   Tatoeba-test.lat-eng: 10000/81934
   Tatoeba-test.lav-eng: 1631/9273
   Tatoeba-test.lij-eng: 94/586
   Tatoeba-test.lit-eng: 2500/14897
   Tatoeba-test.lld-eng: 21/189
   Tatoeba-test.lmo-eng: 17/104
   Tatoeba-test.ltz-eng: 283/1378
   Tatoeba-test.mai-eng: 8/17
   Tatoeba-test.mar-eng: 10000/53713
   Tatoeba-test.mfe-eng: 7/28
   Tatoeba-test.mkd-eng: 10000/54846
   Tatoeba-test.msa-eng: 5000/31183
   Tatoeba-test.multi-eng: 10000/60375
   Tatoeba-test.mwl-eng: 4/19
   Tatoeba-test.nds-eng: 2500/14803
   Tatoeba-test.nep-eng: 115/389
   Tatoeba-test.nld-eng: 10000/58489
   Tatoeba-test.non-eng: 15/112
   Tatoeba-test.nor-eng: 5000/34479
   Tatoeba-test.oci-eng: 841/4376
   Tatoeba-test.ori-eng: 33/198
   Tatoeba-test.orv-eng: 322/1713
   Tatoeba-test.oss-eng: 57/312
   Tatoeba-test.pan-eng: 87/514
   Tatoeba-test.pap-eng: 70/275
   Tatoeba-test.pdc-eng: 53/295
   Tatoeba-test.pms-eng: 268/1708
   Tatoeba-test.pol-eng: 10000/63754
   Tatoeba-test.por-eng: 10000/63761
   Tatoeba-test.prg-eng: 213/1410
   Tatoeba-test.pus-eng: 31/239
   Tatoeba-test.roh-eng: 16/186
   Tatoeba-test.rom-eng: 671/3765
   Tatoeba-test.ron-eng: 5000/31418
   Tatoeba-test.rue-eng: 120/566
   Tatoeba-test.rus-eng: 10000/61769
   Tatoeba-test.san-eng: 144/504
   Tatoeba-test.scn-eng: 4/31
   Tatoeba-test.sco-eng: 27/183
   Tatoeba-test.sgs-eng: 52/128
   Tatoeba-test.sin-eng: 45/210
   Tatoeba-test.slv-eng: 2007/11389
   Tatoeba-test.snd-eng: 4/15
   Tatoeba-test.spa-eng: 10000/67744
   Tatoeba-test.sqi-eng: 1065/6477
   Tatoeba-test.stq-eng: 5/28
   Tatoeba-test.swe-eng: 10000/54997
   Tatoeba-test.swg-eng: 33/205
   Tatoeba-test.tgk-eng: 29/176
   Tatoeba-test.tly-eng: 46/167
   Tatoeba-test.ukr-eng: 10000/55352
   Tatoeba-test.urd-eng: 1663/10180
   Tatoeba-test.vec-eng: 19/90
   Tatoeba-test.wln-eng: 89/363
   Tatoeba-test.yid-eng: 1168/6380
   Tatoeba-test.zza-eng: 523/2522
   news-test2008.deu-eng: 2051/43455
   news-test2008.fra-eng: 2051/43455
   news-test2008.spa-eng: 2051/43455
   newsdev2014.hin-eng: 520/9210
   newsdev2016-enro.ron-eng: 1999/43535
   newsdev2017-enlv.lav-eng: 2003/42560
   newsdev2019-engu.guj-eng: 1998/36882
   newsdev2019-enlt.lit-eng: 2000/43193
   newsdiscussdev2015-enfr.fra-eng: 1500/24201
   newsdiscusstest2015-enfr.fra-eng: 1500/23668
   newssyscomb2009.ces-eng: 502/10424
   newssyscomb2009.deu-eng: 502/10424
   newssyscomb2009.fra-eng: 502/10424
   newssyscomb2009.ita-eng: 502/10424
   newssyscomb2009.spa-eng: 502/10424
   newstest2009.ces-eng: 2525/57625
   newstest2009.deu-eng: 2525/57625
   newstest2009.fra-eng: 2525/57625
   newstest2009.ita-eng: 2525/57625
   newstest2009.spa-eng: 2525/57625
   newstest2010.ces-eng: 2489/54021
   newstest2010.deu-eng: 2489/54021
   newstest2010.fra-eng: 2489/54021
   newstest2010.spa-eng: 2489/54021
   newstest2011.ces-eng: 3003/65829
   newstest2011.deu-eng: 3003/65829
   newstest2011.fra-eng: 3003/65829
   newstest2011.spa-eng: 3003/65829
   newstest2012.ces-eng: 3003/63778
   newstest2012.deu-eng: 3003/63778
   newstest2012.fra-eng: 3003/63778
   newstest2012.rus-eng: 3003/63778
   newstest2012.spa-eng: 3003/63778
   newstest2013.ces-eng: 3000/56089
   newstest2013.deu-eng: 3000/56089
   newstest2013.fra-eng: 3000/56089
   newstest2013.rus-eng: 3000/56089
   newstest2013.spa-eng: 3000/56089
   newstest2014-csen.ces-eng: 3003/60441
   newstest2014-deen.deu-eng: 3003/59325
   newstest2014-fren.fra-eng: 3003/62338
   newstest2014-hien.hin-eng: 2507/49394
   newstest2014-ruen.rus-eng: 3003/61219
   newstest2015-encs.ces-eng: 2656/47160
   newstest2015-ende.deu-eng: 2169/40771
   newstest2015-enru.rus-eng: 2818/56934
   newstest2016-encs.ces-eng: 2999/57418
   newstest2016-ende.deu-eng: 2999/56789
   newstest2016-enro.ron-eng: 1999/42055
   newstest2016-enru.rus-eng: 2998/61076
   newstest2017-encs.ces-eng: 3005/54762
   newstest2017-ende.deu-eng: 3004/56435
   newstest2017-enlv.lav-eng: 2001/42079
   newstest2017-enru.rus-eng: 3001/60714
   newstest2018-encs.ces-eng: 2983/55920
   newstest2018-ende.deu-eng: 2998/58628
   newstest2018-enru.rus-eng: 3000/62925
   newstest2019-deen.deu-eng: 2000/34386
   newstest2019-guen.guj-eng: 1016/15734
   newstest2019-lten.lit-eng: 1000/22953
   newstest2019-ruen.rus-eng: 2000/37348
BLEU-scores:
   Tatoeba-test.afr-eng: 59.0
   Tatoeba-test.ang-eng: 9.9
   Tatoeba-test.arg-eng: 41.6
   Tatoeba-test.asm-eng: 22.8
   Tatoeba-test.ast-eng: 36.1
   Tatoeba-test.awa-eng: 11.6
   Tatoeba-test.bel-eng: 42.2
   Tatoeba-test.ben-eng: 45.8
   Tatoeba-test.bho-eng: 34.4
   Tatoeba-test.bre-eng: 24.4
   Tatoeba-test.bul-eng: 50.8
   Tatoeba-test.cat-eng: 51.2
   Tatoeba-test.ces-eng: 47.6
   Tatoeba-test.cor-eng: 5.4
   Tatoeba-test.cos-eng: 61.0
   Tatoeba-test.csb-eng: 22.5
   Tatoeba-test.cym-eng: 34.7
   Tatoeba-test.dan-eng: 56.2
   Tatoeba-test.deu-eng: 44.9
   Tatoeba-test.dsb-eng: 21.0
   Tatoeba-test.egl-eng: 6.9
   Tatoeba-test.ell-eng: 62.1
   Tatoeba-test.enm-eng: 22.6
   Tatoeba-test.ext-eng: 33.2
   Tatoeba-test.fao-eng: 28.1
   Tatoeba-test.fas-eng: 9.6
   Tatoeba-test.fra-eng: 50.3
   Tatoeba-test.frm-eng: 30.0
   Tatoeba-test.frr-eng: 15.2
   Tatoeba-test.fry-eng: 34.4
   Tatoeba-test.gcf-eng: 18.4
   Tatoeba-test.gla-eng: 24.1
   Tatoeba-test.gle-eng: 52.2
   Tatoeba-test.glg-eng: 50.5
   Tatoeba-test.glv-eng: 5.7
   Tatoeba-test.gos-eng: 19.2
   Tatoeba-test.got-eng: 0.1
   Tatoeba-test.grc-eng: 0.9
   Tatoeba-test.gsw-eng: 23.9
   Tatoeba-test.guj-eng: 28.0
   Tatoeba-test.hat-eng: 44.2
   Tatoeba-test.hbs-eng: 51.6
   Tatoeba-test.hif-eng: 22.3
   Tatoeba-test.hin-eng: 41.7
   Tatoeba-test.hsb-eng: 46.4
   Tatoeba-test.hye-eng: 40.4
   Tatoeba-test.isl-eng: 43.8
   Tatoeba-test.ita-eng: 60.7
   Tatoeba-test.jdt-eng: 5.5
   Tatoeba-test.kok-eng: 7.8
   Tatoeba-test.ksh-eng: 15.8
   Tatoeba-test.kur-eng: 11.6
   Tatoeba-test.lad-eng: 30.7
   Tatoeba-test.lah-eng: 11.0
   Tatoeba-test.lat-eng: 24.4
   Tatoeba-test.lav-eng: 47.2
   Tatoeba-test.lij-eng: 9.0
   Tatoeba-test.lit-eng: 51.7
   Tatoeba-test.lld-eng: 22.4
   Tatoeba-test.lmo-eng: 26.1
   Tatoeba-test.ltz-eng: 39.8
   Tatoeba-test.mai-eng: 72.3
   Tatoeba-test.mar-eng: 32.0
   Tatoeba-test.mfe-eng: 63.1
   Tatoeba-test.mkd-eng: 49.5
   Tatoeba-test.msa-eng: 38.6
   Tatoeba-test.multi-eng: 45.6
   Tatoeba-test.mwl-eng: 40.4
   Tatoeba-test.nds-eng: 35.5
   Tatoeba-test.nep-eng: 4.9
   Tatoeba-test.nld-eng: 54.2
   Tatoeba-test.non-eng: 39.3
   Tatoeba-test.nor-eng: 50.9
   Tatoeba-test.oci-eng: 19.6
   Tatoeba-test.ori-eng: 16.2
   Tatoeba-test.orv-eng: 13.6
   Tatoeba-test.oss-eng: 9.4
   Tatoeba-test.pan-eng: 17.1
   Tatoeba-test.pap-eng: 57.0
   Tatoeba-test.pdc-eng: 41.6
   Tatoeba-test.pms-eng: 13.7
   Tatoeba-test.pol-eng: 46.5
   Tatoeba-test.por-eng: 56.4
   Tatoeba-test.prg-eng: 2.3
   Tatoeba-test.pus-eng: 3.2
   Tatoeba-test.roh-eng: 17.5
   Tatoeba-test.rom-eng: 5.0
   Tatoeba-test.ron-eng: 51.4
   Tatoeba-test.rue-eng: 26.0
   Tatoeba-test.rus-eng: 47.8
   Tatoeba-test.san-eng: 4.0
   Tatoeba-test.scn-eng: 45.1
   Tatoeba-test.sco-eng: 41.9
   Tatoeba-test.sgs-eng: 38.7
   Tatoeba-test.sin-eng: 29.7
   Tatoeba-test.slv-eng: 38.2
   Tatoeba-test.snd-eng: 12.7
   Tatoeba-test.spa-eng: 53.2
   Tatoeba-test.sqi-eng: 51.9
   Tatoeba-test.stq-eng: 9.0
   Tatoeba-test.swe-eng: 57.4
   Tatoeba-test.swg-eng: 18.0
   Tatoeba-test.tgk-eng: 24.3
   Tatoeba-test.tly-eng: 1.1
   Tatoeba-test.ukr-eng: 48.0
   Tatoeba-test.urd-eng: 27.2
   Tatoeba-test.vec-eng: 28.0
   Tatoeba-test.wln-eng: 17.5
   Tatoeba-test.yid-eng: 26.9
   Tatoeba-test.zza-eng: 1.7
   news-test2008.deu-eng: 24.0
   news-test2008.fra-eng: 24.5
   news-test2008.spa-eng: 25.5
   newsdev2014.hin-eng: 11.2
   newsdev2016-enro.ron-eng: 35.5
   newsdev2017-enlv.lav-eng: 25.1
   newsdev2019-engu.guj-eng: 16.0
   newsdev2019-enlt.lit-eng: 24.0
   newsdiscussdev2015-enfr.fra-eng: 30.1
   newsdiscusstest2015-enfr.fra-eng: 33.4
   newssyscomb2009.ces-eng: 24.0
   newssyscomb2009.deu-eng: 25.7
   newssyscomb2009.fra-eng: 27.9
   newssyscomb2009.ita-eng: 31.4
   newssyscomb2009.spa-eng: 28.3
   newstest2009.ces-eng: 23.3
   newstest2009.deu-eng: 23.2
   newstest2009.fra-eng: 27.3
   newstest2009.ita-eng: 30.3
   newstest2009.spa-eng: 27.9
   newstest2010.ces-eng: 23.8
   newstest2010.deu-eng: 26.2
   newstest2010.fra-eng: 28.6
   newstest2010.spa-eng: 31.4
   newstest2011.ces-eng: 24.2
   newstest2011.deu-eng: 23.9
   newstest2011.fra-eng: 29.5
   newstest2011.spa-eng: 30.3
   newstest2012.ces-eng: 23.5
   newstest2012.deu-eng: 24.9
   newstest2012.fra-eng: 30.0
   newstest2012.rus-eng: 29.9
   newstest2012.spa-eng: 33.3
   newstest2013.ces-eng: 25.6
   newstest2013.deu-eng: 27.7
   newstest2013.fra-eng: 30.0
   newstest2013.rus-eng: 24.4
   newstest2013.spa-eng: 30.8
   newstest2014-csen.ces-eng: 27.7
   newstest2014-deen.deu-eng: 27.7
   newstest2014-fren.fra-eng: 32.2
   newstest2014-hien.hin-eng: 16.7
   newstest2014-ruen.rus-eng: 27.2
   newstest2015-encs.ces-eng: 25.4
   newstest2015-ende.deu-eng: 28.8
   newstest2015-enru.rus-eng: 25.6
   newstest2016-encs.ces-eng: 27.0
   newstest2016-ende.deu-eng: 33.5
   newstest2016-enro.ron-eng: 32.8
   newstest2016-enru.rus-eng: 24.8
   newstest2017-encs.ces-eng: 23.7
   newstest2017-ende.deu-eng: 29.3
   newstest2017-enlv.lav-eng: 18.9
   newstest2017-enru.rus-eng: 28.0
   newstest2018-encs.ces-eng: 24.9
   newstest2018-ende.deu-eng: 36.0
   newstest2018-enru.rus-eng: 23.8
   newstest2019-deen.deu-eng: 31.5
   newstest2019-guen.guj-eng: 12.1
   newstest2019-lten.lit-eng: 26.6
   newstest2019-ruen.rus-eng: 27.5
chr-F-scores:
   Tatoeba-test.afr-eng: 0.724
   Tatoeba-test.ang-eng: 0.254
   Tatoeba-test.arg-eng: 0.487
   Tatoeba-test.asm-eng: 0.392
   Tatoeba-test.ast-eng: 0.521
   Tatoeba-test.awa-eng: 0.280
   Tatoeba-test.bel-eng: 0.597
   Tatoeba-test.ben-eng: 0.598
   Tatoeba-test.bho-eng: 0.518
   Tatoeba-test.bre-eng: 0.405
   Tatoeba-test.bul-eng: 0.660
   Tatoeba-test.cat-eng: 0.677
   Tatoeba-test.ces-eng: 0.641
   Tatoeba-test.cor-eng: 0.214
   Tatoeba-test.cos-eng: 0.675
   Tatoeba-test.csb-eng: 0.394
   Tatoeba-test.cym-eng: 0.522
   Tatoeba-test.dan-eng: 0.708
   Tatoeba-test.deu-eng: 0.625
   Tatoeba-test.dsb-eng: 0.383
   Tatoeba-test.egl-eng: 0.221
   Tatoeba-test.ell-eng: 0.741
   Tatoeba-test.enm-eng: 0.466
   Tatoeba-test.ext-eng: 0.496
   Tatoeba-test.fao-eng: 0.460
   Tatoeba-test.fas-eng: 0.306
   Tatoeba-test.fra-eng: 0.661
   Tatoeba-test.frm-eng: 0.457
   Tatoeba-test.frr-eng: 0.301
   Tatoeba-test.fry-eng: 0.525
   Tatoeba-test.gcf-eng: 0.317
   Tatoeba-test.gla-eng: 0.400
   Tatoeba-test.gle-eng: 0.671
   Tatoeba-test.glg-eng: 0.669
   Tatoeba-test.glv-eng: 0.189
   Tatoeba-test.gos-eng: 0.378
   Tatoeba-test.got-eng: 0.022
   Tatoeba-test.grc-eng: 0.095
   Tatoeba-test.gsw-eng: 0.390
   Tatoeba-test.guj-eng: 0.428
   Tatoeba-test.hat-eng: 0.567
   Tatoeba-test.hbs-eng: 0.666
   Tatoeba-test.hif-eng: 0.451
   Tatoeba-test.hin-eng: 0.585
   Tatoeba-test.hsb-eng: 0.590
   Tatoeba-test.hye-eng: 0.564
   Tatoeba-test.isl-eng: 0.605
   Tatoeba-test.ita-eng: 0.735
   Tatoeba-test.jdt-eng: 0.091
   Tatoeba-test.kok-eng: 0.205
   Tatoeba-test.ksh-eng: 0.284
   Tatoeba-test.kur-eng: 0.232
   Tatoeba-test.lad-eng: 0.484
   Tatoeba-test.lah-eng: 0.286
   Tatoeba-test.lat-eng: 0.432
   Tatoeba-test.lav-eng: 0.646
   Tatoeba-test.lij-eng: 0.287
   Tatoeba-test.lit-eng: 0.670
   Tatoeba-test.lld-eng: 0.369
   Tatoeba-test.lmo-eng: 0.381
   Tatoeba-test.ltz-eng: 0.536
   Tatoeba-test.mai-eng: 0.758
   Tatoeba-test.mar-eng: 0.554
   Tatoeba-test.mfe-eng: 0.822
   Tatoeba-test.mkd-eng: 0.638
   Tatoeba-test.msa-eng: 0.566
   Tatoeba-test.multi-eng: 0.615
   Tatoeba-test.mwl-eng: 0.767
   Tatoeba-test.nds-eng: 0.538
   Tatoeba-test.nep-eng: 0.209
   Tatoeba-test.nld-eng: 0.694
   Tatoeba-test.non-eng: 0.573
   Tatoeba-test.nor-eng: 0.663
   Tatoeba-test.oci-eng: 0.386
   Tatoeba-test.ori-eng: 0.364
   Tatoeba-test.orv-eng: 0.288
   Tatoeba-test.oss-eng: 0.301
   Tatoeba-test.pan-eng: 0.389
   Tatoeba-test.pap-eng: 0.680
   Tatoeba-test.pdc-eng: 0.526
   Tatoeba-test.pms-eng: 0.333
   Tatoeba-test.pol-eng: 0.632
   Tatoeba-test.por-eng: 0.710
   Tatoeba-test.prg-eng: 0.193
   Tatoeba-test.pus-eng: 0.194
   Tatoeba-test.roh-eng: 0.420
   Tatoeba-test.rom-eng: 0.237
   Tatoeba-test.ron-eng: 0.670
   Tatoeba-test.rue-eng: 0.447
   Tatoeba-test.rus-eng: 0.634
   Tatoeba-test.san-eng: 0.195
   Tatoeba-test.scn-eng: 0.440
   Tatoeba-test.sco-eng: 0.582
   Tatoeba-test.sgs-eng: 0.498
   Tatoeba-test.sin-eng: 0.499
   Tatoeba-test.slv-eng: 0.564
   Tatoeba-test.snd-eng: 0.342
   Tatoeba-test.spa-eng: 0.687
   Tatoeba-test.sqi-eng: 0.679
   Tatoeba-test.stq-eng: 0.391
   Tatoeba-test.swe-eng: 0.705
   Tatoeba-test.swg-eng: 0.338
   Tatoeba-test.tgk-eng: 0.413
   Tatoeba-test.tly-eng: 0.094
   Tatoeba-test.ukr-eng: 0.639
   Tatoeba-test.urd-eng: 0.471
   Tatoeba-test.vec-eng: 0.398
   Tatoeba-test.wln-eng: 0.320
   Tatoeba-test.yid-eng: 0.457
   Tatoeba-test.zza-eng: 0.131
   news-test2008.deu-eng: 0.515
   news-test2008.fra-eng: 0.524
   news-test2008.spa-eng: 0.533
   newsdev2014.hin-eng: 0.375
   newsdev2016-enro.ron-eng: 0.614
   newsdev2017-enlv.lav-eng: 0.542
   newsdev2019-engu.guj-eng: 0.420
   newsdev2019-enlt.lit-eng: 0.522
   newsdiscussdev2015-enfr.fra-eng: 0.550
   newsdiscusstest2015-enfr.fra-eng: 0.572
   newssyscomb2009.ces-eng: 0.520
   newssyscomb2009.deu-eng: 0.526
   newssyscomb2009.fra-eng: 0.550
   newssyscomb2009.ita-eng: 0.574
   newssyscomb2009.spa-eng: 0.555
   newstest2009.ces-eng: 0.516
   newstest2009.deu-eng: 0.512
   newstest2009.fra-eng: 0.545
   newstest2009.ita-eng: 0.567
   newstest2009.spa-eng: 0.549
   newstest2010.ces-eng: 0.523
   newstest2010.deu-eng: 0.545
   newstest2010.fra-eng: 0.562
   newstest2010.spa-eng: 0.581
   newstest2011.ces-eng: 0.521
   newstest2011.deu-eng: 0.522
   newstest2011.fra-eng: 0.570
   newstest2011.spa-eng: 0.570
   newstest2012.ces-eng: 0.516
   newstest2012.deu-eng: 0.529
   newstest2012.fra-eng: 0.568
   newstest2012.rus-eng: 0.565
   newstest2012.spa-eng: 0.593
   newstest2013.ces-eng: 0.531
   newstest2013.deu-eng: 0.545
   newstest2013.fra-eng: 0.561
   newstest2013.rus-eng: 0.514
   newstest2013.spa-eng: 0.577
   newstest2014-csen.ces-eng: 0.558
   newstest2014-deen.deu-eng: 0.545
   newstest2014-fren.fra-eng: 0.592
   newstest2014-hien.hin-eng: 0.450
   newstest2014-ruen.rus-eng: 0.552
   newstest2015-encs.ces-eng: 0.518
   newstest2015-ende.deu-eng: 0.552
   newstest2015-enru.rus-eng: 0.527
   newstest2016-encs.ces-eng: 0.540
   newstest2016-ende.deu-eng: 0.592
   newstest2016-enro.ron-eng: 0.591
   newstest2016-enru.rus-eng: 0.523
   newstest2017-encs.ces-eng: 0.510
   newstest2017-ende.deu-eng: 0.556
   newstest2017-enlv.lav-eng: 0.486
   newstest2017-enru.rus-eng: 0.546
   newstest2018-encs.ces-eng: 0.521
   newstest2018-ende.deu-eng: 0.604
   newstest2018-enru.rus-eng: 0.517
   newstest2019-deen.deu-eng: 0.570
   newstest2019-guen.guj-eng: 0.377
   newstest2019-lten.lit-eng: 0.555
   newstest2019-ruen.rus-eng: 0.541