Skip to content

Latest commit

 

History

History

itc-gmw

README.md

opusTCv20210807_transformer-big_2022-08-23.zip

  • dataset: opusTCv20210807
  • model: transformer-big
  • source language(s): arg ast cat cbk_Latn cos crs egl ext fra frm_Latn frp frp_Arab frp_Beng frp_Cyrl frp_Deva frp_Grek frp_Gujr frp_Guru frp_Hang frp_Hani frp_Mlym frp_Orya frp_Taml fur fur_Latn gcf_Latn glg hat ita kea lad lad_Latn lat_Latn lij lld_Latn lmo mfe mol mwl nap oci osp_Latn pap pms pob por roh ron rup rup_Cyrl scn spa srd srd_Grek vec wln
  • target language(s): afr afr_Arab ang_Latn bar deu drt_Latn eng enm_Latn frr fry gos gsw hrx_Latn jam ksh lim ltz nds nld sco stq swg tpi yid zea
  • raw source language(s): arg ast cat cbk cos crs egl ext fra frm frp fur gcf glg hat ita kea lad lat lij lld lmo mfe mol mwl nap oci osp pap pms pob por roh ron rup scn spa srd vec wln
  • raw target language(s): afr ang bar deu drt eng enm frr fry gos gsw hrx jam ksh lim ltz nds nld sco stq swg tpi yid zea
  • model: transformer-big
  • pre-processing: normalization + SentencePiece (spm32k,spm32k)
  • a sentence initial language token is required in the form of >>id<< (id = valid target language ID)
  • valid language labels:
  • download: opusTCv20210807_transformer-big_2022-08-23.zip
  • test set translations: opusTCv20210807_transformer-big_2022-08-23.test.txt
  • test set scores: opusTCv20210807_transformer-big_2022-08-23.eval.txt

Benchmarks

testset BLEU chr-F #sent #words BP
euelections_dev2019.fr-de.fra-deu 28.4 0.58062 1512 33478 1.000
newsdev2016-enro.ron-eng 40.7 0.65164 1999 49526 0.986
newsdiscussdev2015-enfr.fra-eng 33.0 0.57963 1500 27759 0.942
newsdiscusstest2015-enfr.fra-eng 38.5 0.61383 1500 26995 0.973
newssyscomb2009.fra-deu 24.1 0.53603 502 11271 0.979
newssyscomb2009.fra-eng 30.4 0.56884 502 11821 0.981
newssyscomb2009.ita-deu 23.3 0.54029 502 11271 0.967
newssyscomb2009.ita-eng 34.4 0.59782 502 11821 0.967
newssyscomb2009.spa-deu 21.9 0.52997 502 11271 0.988
newssyscomb2009.spa-eng 30.3 0.56883 502 11821 0.977
news-test2008.fra-deu 23.8 0.53383 2051 47427 0.995
news-test2008.fra-eng 26.3 0.54179 2051 49380 0.985
news-test2008.spa-deu 22.0 0.52028 2051 47427 0.994
news-test2008.spa-eng 27.0 0.54795 2051 49380 0.982
newstest2009.fra-deu 23.0 0.52672 2525 62816 0.990
newstest2009.fra-eng 29.6 0.56280 2525 65402 0.978
newstest2009.ita-deu 22.9 0.52847 2525 62816 0.976
newstest2009.ita-eng 32.7 0.58787 2525 65402 0.955
newstest2009.spa-deu 22.3 0.52592 2525 62816 0.990
newstest2009.spa-eng 29.4 0.56508 2525 65402 0.972
newstest2010.fra-deu 24.0 0.53776 2489 61511 0.971
newstest2010.fra-eng 32.7 0.58998 2489 61724 0.993
newstest2010.spa-deu 26.0 0.55131 2489 61511 0.959
newstest2010.spa-eng 35.6 0.61101 2489 61724 0.980
newstest2011.fra-deu 23.3 0.52811 3003 72981 1.000
newstest2011.fra-eng 32.6 0.59388 3003 74681 0.991
newstest2011.spa-deu 23.9 0.52982 3003 72981 0.993
newstest2011.spa-eng 33.1 0.59179 3003 74681 0.985
newstest2012.fra-deu 23.9 0.52781 3003 72886 0.973
newstest2012.fra-eng 32.4 0.58894 3003 72812 0.987
newstest2012.spa-deu 24.5 0.53307 3003 72886 0.983
newstest2012.spa-eng 37.2 0.61832 3003 72812 1.000
newstest2013.fra-deu 26.0 0.54215 3000 63737 1.000
newstest2013.fra-eng 34.2 0.58974 3000 64505 1.000
newstest2013.spa-deu 26.6 0.54907 3000 63737 1.000
newstest2013.spa-eng 34.6 0.60398 3000 64505 1.000
newstest2014-fren.fra-eng 37.7 0.63405 3003 70708 0.985
newstest2016-enro.ron-eng 38.9 0.63291 1999 47563 0.989
newstest2019-frde.fra-deu 30.2 0.60761 1701 36571 1.000
Tatoeba-test-v2021-08-07.arg-eng 39.4 0.50054 105 451 0.975
Tatoeba-test-v2021-08-07.ast-deu 35.4 0.56781 35 260 1.000
Tatoeba-test-v2021-08-07.ast-eng 33.2 0.49395 107 860 1.000
Tatoeba-test-v2021-08-07.ast-gos 6.6 0.10499 1 6 1.000
Tatoeba-test-v2021-08-07.ast-nds 10.4 0.23383 1 6 0.819
Tatoeba-test-v2021-08-07.ast-nld 48.9 0.82914 1 6 1.000
Tatoeba-test-v2021-08-07.cat-deu 46.7 0.65951 723 5673 0.994
Tatoeba-test-v2021-08-07.cat-eng 56.2 0.71161 1631 12625 0.979
Tatoeba-test-v2021-08-07.cat-enm 20.1 0.44289 2 8 1.000
Tatoeba-test-v2021-08-07.cat-nld 50.3 0.67573 578 4184 0.980
Tatoeba-test-v2021-08-07.cat-yid 5.3 0.20211 6 36 0.717
Tatoeba-test-v2021-08-07.cbk-eng 24.0 0.44205 1498 10024 1.000
Tatoeba-test-v2021-08-07.cos-deu 0.0 0.87724 1 2 1.000
Tatoeba-test-v2021-08-07.cos-eng 70.8 0.78597 5 42 1.000
Tatoeba-test-v2021-08-07.egl-deu 4.7 5.631 3 15 1.000
Tatoeba-test-v2021-08-07.egl-eng 2.8 0.17754 84 444 1.000
Tatoeba-test-v2021-08-07.ext-eng 41.8 0.56307 69 396 0.932
Tatoeba-test-v2021-08-07.fra-afr 62.7 0.74968 195 1273 1.000
Tatoeba-test-v2021-08-07.fra-afr_Arab 3.3 0.689 2 14 1.000
Tatoeba-test-v2021-08-07.fra-ang 1.1 0.15885 40 358 1.000
Tatoeba-test-v2021-08-07.fra-bar 9.7 0.11311 1 5 1.000
Tatoeba-test-v2021-08-07.fra-deu 48.9 0.67582 12418 100525 0.993
Tatoeba-test-v2021-08-07.fra-drt 6.5 0.26848 3 20 0.949
Tatoeba-test-v2021-08-07.fra-eng 57.0 0.71751 12681 101729 0.983
Tatoeba-test-v2021-08-07.fra-enm 7.9 0.30365 9 62 0.984
Tatoeba-test-v2021-08-07.fra-frr 3.5 0.22769 4 18 1.000
Tatoeba-test-v2021-08-07.fra-fry 16.7 0.37674 50 306 0.977
Tatoeba-test-v2021-08-07.fra-gos 1.8 0.16470 31 149 1.000
Tatoeba-test-v2021-08-07.fra-jam 16.0 0.22980 1 3 1.000
Tatoeba-test-v2021-08-07.fra-ksh 4.8 0.14421 1 9 1.000
Tatoeba-test-v2021-08-07.fra-lim 8.1 0.26206 3 20 0.895
Tatoeba-test-v2021-08-07.fra-ltz 16.8 0.30920 33 171 0.964
Tatoeba-test-v2021-08-07.fra-nds 12.1 0.34245 857 5760 0.975
Tatoeba-test-v2021-08-07.fra-nld 47.6 0.66054 11548 82129 0.981
Tatoeba-test-v2021-08-07.fra-sco 5.6 0.28568 7 53 1.000
Tatoeba-test-v2021-08-07.fra-stq 5.8 0.26814 1 9 0.882
Tatoeba-test-v2021-08-07.fra-swg 2.3 0.12423 4 25 1.000
Tatoeba-test-v2021-08-07.fra-tpi 25.3 0.63430 6 29 1.000
Tatoeba-test-v2021-08-07.fra-yid 1.4 0.12947 384 2379 0.886
Tatoeba-test-v2021-08-07.fra-zea 7.7 0.21533 2 15 0.931
Tatoeba-test-v2021-08-07.frm-eng 21.3 0.44084 18 231 0.928
Tatoeba-test-v2021-08-07.fur-eng 34.3 0.54198 10 47 0.934
Tatoeba-test-v2021-08-07.gcf-eng 23.7 0.35334 101 580 0.991
Tatoeba-test-v2021-08-07.glg-deu 48.4 0.67340 103 688 1.000
Tatoeba-test-v2021-08-07.glg-eng 54.9 0.69963 1015 8420 0.972
Tatoeba-test-v2021-08-07.glg-nld 53.1 0.70452 118 788 0.972
Tatoeba-test-v2021-08-07.glg-yid 3.6 0.14346 4 21 1.000
Tatoeba-test-v2021-08-07.hat-eng 52.1 0.69063 64 384 1.000
Tatoeba-test-v2021-08-07.hat-nld 52.9 0.62525 60 345 1.000
Tatoeba-test-v2021-08-07.hat-yid 16.0 0.13783 1 4 1.000
Tatoeba-test-v2021-08-07.ita-afr 68.2 0.80876 124 773 0.997
Tatoeba-test-v2021-08-07.ita-ang 4.5 0.18347 10 73 1.000
Tatoeba-test-v2021-08-07.ita-deu 49.2 0.68139 10094 79754 0.983
Tatoeba-test-v2021-08-07.ita-eng 69.7 0.80575 17320 119198 0.992
Tatoeba-test-v2021-08-07.ita-enm 6.9 0.23141 3 13 0.920
Tatoeba-test-v2021-08-07.ita-fry 7.6 0.30084 37 226 0.931
Tatoeba-test-v2021-08-07.ita-gos 11.8 0.28957 5 27 0.923
Tatoeba-test-v2021-08-07.ita-ltz 4.5 0.26261 50 226 1.000
Tatoeba-test-v2021-08-07.ita-nds 15.3 0.36738 313 2152 0.993
Tatoeba-test-v2021-08-07.ita-nld 56.8 0.72513 2578 16680 0.987
Tatoeba-test-v2021-08-07.ita-yid 2.1 0.12967 206 1198 0.865
Tatoeba-test-v2021-08-07.lad-ang 2.8 9.414 15 71 1.000
Tatoeba-test-v2021-08-07.lad-ang_Latn 2.4 3.341 7 31 1.000
Tatoeba-test-v2021-08-07.lad-deu 25.7 0.44959 220 1175 1.000
Tatoeba-test-v2021-08-07.lad-drt 3.6 0.13857 4 28 1.000
Tatoeba-test-v2021-08-07.lad-drt_Latn 2.7 0.10976 2 14 1.000
Tatoeba-test-v2021-08-07.lad-eng 33.0 0.47694 768 4184 1.000
Tatoeba-test-v2021-08-07.lad-enm 10.0 0.28880 51 266 1.000
Tatoeba-test-v2021-08-07.lad-enm_Latn 0.9 6.521 16 86 1.000
Tatoeba-test-v2021-08-07.lad-fry 11.4 0.28211 6 38 0.918
Tatoeba-test-v2021-08-07.lad-gos 1.9 0.12450 11 67 1.000
Tatoeba-test-v2021-08-07.lad_Latn-ang_Latn 4.3 0.14253 8 40 1.000
Tatoeba-test-v2021-08-07.lad_Latn-deu 38.4 0.55250 173 925 0.979
Tatoeba-test-v2021-08-07.lad_Latn-drt_Latn 8.3 0.16907 2 14 1.000
Tatoeba-test-v2021-08-07.lad_Latn-eng 37.5 0.53583 672 3665 0.992
Tatoeba-test-v2021-08-07.lad_Latn-enm_Latn 15.4 0.38814 35 180 0.972
Tatoeba-test-v2021-08-07.lad_Latn-fry 19.3 0.46995 3 19 1.000
Tatoeba-test-v2021-08-07.lad_Latn-gos 4.7 0.16696 5 31 0.967
Tatoeba-test-v2021-08-07.lad_Latn-ltz 4.7 0.19216 4 29 0.770
Tatoeba-test-v2021-08-07.lad_Latn-nds 12.4 0.52708 3 17 1.000
Tatoeba-test-v2021-08-07.lad_Latn-nld 48.7 0.60881 21 126 0.992
Tatoeba-test-v2021-08-07.lad_Latn-stq 7.7 0.13687 2 15 0.931
Tatoeba-test-v2021-08-07.lad_Latn-swg 16.0 6.720 1 4 1.000
Tatoeba-test-v2021-08-07.lad_Latn-yid 2.6 0.13542 438 2499 0.853
Tatoeba-test-v2021-08-07.lad-ltz 2.6 0.15314 7 51 1.000
Tatoeba-test-v2021-08-07.lad-nds 10.5 0.43169 4 22 0.905
Tatoeba-test-v2021-08-07.lad-nld 30.8 0.41743 33 199 1.000
Tatoeba-test-v2021-08-07.lad-sco 10.7 5.712 1 5 1.000
Tatoeba-test-v2021-08-07.lad-stq 3.6 0.11271 4 30 1.000
Tatoeba-test-v2021-08-07.lad-swg 7.6 5.249 2 8 1.000
Tatoeba-test-v2021-08-07.lad-yid 2.1 0.12171 604 3439 0.778
Tatoeba-test-v2021-08-07.lat-afr 6.2 0.10653 2 11 0.565
Tatoeba-test-v2021-08-07.lat-ang 2.3 0.13974 23 116 0.842
Tatoeba-test-v2021-08-07.lat-bar 16.0 0.10447 1 4 1.000
Tatoeba-test-v2021-08-07.lat-deu 4.8 0.19938 2016 13323 0.808
Tatoeba-test-v2021-08-07.lat-drt 19.0 0.12967 1 4 1.000
Tatoeba-test-v2021-08-07.lat-eng 3.2 0.19296 10298 100151 0.763
Tatoeba-test-v2021-08-07.lat-enm 3.2 0.14934 50 331 0.960
Tatoeba-test-v2021-08-07.lat-fry 6.0 0.22674 5 22 1.000
Tatoeba-test-v2021-08-07.lat-gos 3.8 0.13067 6 27 0.962
Tatoeba-test-v2021-08-07.lat-ltz 6.5 0.16498 4 20 1.000
Tatoeba-test-v2021-08-07.lat-nds 3.9 0.12609 3 20 0.838
Tatoeba-test-v2021-08-07.lat-nld 7.4 0.22532 366 2423 0.802
Tatoeba-test-v2021-08-07.lat-sco 1.7 0.14804 4 26 1.000
Tatoeba-test-v2021-08-07.lat-stq 19.0 0.13677 1 4 1.000
Tatoeba-test-v2021-08-07.lat-swg 0.0 9.225 1 4 0.717
Tatoeba-test-v2021-08-07.lat-yid 0.4 0.10022 458 2794 0.660
Tatoeba-test-v2021-08-07.lij-eng 15.1 0.34328 96 716 1.000
Tatoeba-test-v2021-08-07.lld-eng 16.1 0.35867 21 226 0.945
Tatoeba-test-v2021-08-07.lmo-eng 6.5 0.28460 17 132 0.913
Tatoeba-test-v2021-08-07.mfe-eng 47.3 0.68388 7 35 1.000
Tatoeba-test-v2021-08-07.multi-multi 53.5 0.69346 10000 76969 0.982
Tatoeba-test-v2021-08-07.mwl-eng 25.9 0.52245 4 24 0.913
Tatoeba-test-v2021-08-07.mwl-enm 20.1 0.44289 2 8 1.000
Tatoeba-test-v2021-08-07.oci-deu 16.1 0.40921 174 1319 1.000
Tatoeba-test-v2021-08-07.oci-eng 22.0 0.39540 841 5299 1.000
Tatoeba-test-v2021-08-07.oci-enm 20.1 0.44289 2 8 1.000
Tatoeba-test-v2021-08-07.oci-nld 26.2 0.45886 62 425 1.000
Tatoeba-test-v2021-08-07.oci-yid 39.8 0.50245 1 4 1.000
Tatoeba-test-v2021-08-07.osp-ang 5.5 0.13414 2 10 1.000
Tatoeba-test-v2021-08-07.osp-deu 84.9 0.93445 4 20 1.000
Tatoeba-test-v2021-08-07.osp-eng 46.9 0.65151 3 21 1.000
Tatoeba-test-v2021-08-07.osp-enm 42.7 0.60060 1 5 1.000
Tatoeba-test-v2021-08-07.osp-gos 12.7 0.19548 1 4 1.000
Tatoeba-test-v2021-08-07.osp-yid 3.9 0.18914 5 26 0.920
Tatoeba-test-v2021-08-07.pap-drt 19.0 0.17882 1 4 1.000
Tatoeba-test-v2021-08-07.pap-eng 62.0 0.70284 72 374 1.000
Tatoeba-test-v2021-08-07.pap-enm 25.0 0.55832 2 8 1.000
Tatoeba-test-v2021-08-07.pap-fry 100.0 10.00000 1 4 1.000
Tatoeba-test-v2021-08-07.pap-gos 19.0 0.20135 1 4 1.000
Tatoeba-test-v2021-08-07.pap-ltz 19.0 0.22419 1 4 1.000
Tatoeba-test-v2021-08-07.pap-nld 60.8 0.72131 54 231 1.000
Tatoeba-test-v2021-08-07.pap-stq 19.0 0.13853 1 4 1.000
Tatoeba-test-v2021-08-07.pap-yid 10.7 0.12042 1 5 1.000
Tatoeba-test-v2021-08-07.pms-deu 15.5 0.34499 12 91 1.000
Tatoeba-test-v2021-08-07.pms-eng 16.0 0.35543 269 2059 1.000
Tatoeba-test-v2021-08-07.por-afr 60.4 0.76500 94 599 0.968
Tatoeba-test-v2021-08-07.por-ang 16.8 0.41590 2 9 0.882
Tatoeba-test-v2021-08-07.por-deu 48.8 0.67795 10000 81221 0.998
Tatoeba-test-v2021-08-07.por-drt 9.0 0.16566 2 8 1.000
Tatoeba-test-v2021-08-07.por-eng 63.2 0.76344 13222 105318 0.976
Tatoeba-test-v2021-08-07.por-enm 8.6 0.43777 7 30 1.000
Tatoeba-test-v2021-08-07.por-fry 12.9 0.32196 21 128 0.944
Tatoeba-test-v2021-08-07.por-gos 4.7 0.14692 4 18 1.000
Tatoeba-test-v2021-08-07.por-hrx 1.5 0.13188 8 54 0.981
Tatoeba-test-v2021-08-07.por-ltz 16.6 0.33385 5 22 1.000
Tatoeba-test-v2021-08-07.por-nds 18.5 0.39676 207 1292 1.000
Tatoeba-test-v2021-08-07.por-nld 52.6 0.69563 2500 17810 0.978
Tatoeba-test-v2021-08-07.por-stq 9.0 0.13786 2 8 1.000
Tatoeba-test-v2021-08-07.por-swg 16.0 6.510 1 4 1.000
Tatoeba-test-v2021-08-07.por-yid 2.0 0.11450 138 837 0.880
Tatoeba-test-v2021-08-07.roh-deu 16.1 0.36863 14 141 1.000
Tatoeba-test-v2021-08-07.roh-eng 29.5 0.50132 16 214 0.902
Tatoeba-test-v2021-08-07.ron-afr 50.0 0.78358 2 9 1.000
Tatoeba-test-v2021-08-07.ron-deu 48.1 0.66211 1141 7889 0.987
Tatoeba-test-v2021-08-07.ron-eng 55.9 0.70853 5508 40715 0.974
Tatoeba-test-v2021-08-07.ron-enm 20.1 0.44289 2 8 1.000
Tatoeba-test-v2021-08-07.ron-gos 9.7 0.20366 1 6 1.000
Tatoeba-test-v2021-08-07.ron-nds 12.7 0.15651 1 5 1.000
Tatoeba-test-v2021-08-07.ron-nld 45.0 0.63599 2269 16690 0.959
Tatoeba-test-v2021-08-07.ron-yid 3.8 0.12540 16 78 0.757
Tatoeba-test-v2021-08-07.scn-deu 9.3 0.26518 5 36 1.000
Tatoeba-test-v2021-08-07.scn-eng 37.0 0.47724 9 98 0.904
Tatoeba-test-v2021-08-07.scn-gos 9.7 0.24412 1 5 1.000
Tatoeba-test-v2021-08-07.scn-nld 66.9 0.71999 1 5 1.000
Tatoeba-test-v2021-08-07.spa-afr 61.4 0.75948 448 3044 1.000
Tatoeba-test-v2021-08-07.spa-ang 1.4 0.19248 18 168 1.000
Tatoeba-test-v2021-08-07.spa-bar 4.5 0.18380 4 23 1.000
Tatoeba-test-v2021-08-07.spa-deu 48.5 0.67476 10521 86407 0.989
Tatoeba-test-v2021-08-07.spa-drt 5.5 0.24762 4 30 0.966
Tatoeba-test-v2021-08-07.spa-eng 59.8 0.74172 16583 138085 0.981
Tatoeba-test-v2021-08-07.spa-enm 8.4 0.40907 7 58 1.000
Tatoeba-test-v2021-08-07.spa-fry 13.9 0.33018 48 307 1.000
Tatoeba-test-v2021-08-07.spa-gos 1.9 0.19785 116 576 1.000
Tatoeba-test-v2021-08-07.spa-gsw 3.8 0.10455 4 15 1.000
Tatoeba-test-v2021-08-07.spa-ksh 0.8 0.12464 10 99 0.959
Tatoeba-test-v2021-08-07.spa-lim 7.2 0.27754 4 26 0.961
Tatoeba-test-v2021-08-07.spa-ltz 11.7 0.27459 16 90 0.943
Tatoeba-test-v2021-08-07.spa-nds 16.5 0.38428 923 5940 0.984
Tatoeba-test-v2021-08-07.spa-nld 50.2 0.68215 10113 79143 0.975
Tatoeba-test-v2021-08-07.spa-stq 5.8 0.28657 4 28 0.926
Tatoeba-test-v2021-08-07.spa-yid 1.6 0.11789 407 2599 0.832
Tatoeba-test-v2021-08-07.spa-zea 7.7 0.21623 2 15 0.931
Tatoeba-test-v2021-08-07.vec-eng 23.1 0.37141 19 127 1.000
Tatoeba-test-v2021-08-07.wln-eng 1.7 0.18133 89 465 1.000
Tatoeba-test-v2021-08-07.wln-nld 6.2 0.21694 87 432 1.000
tico19-test.fra-eng 39.6 0.62226 2100 56347 0.995
tico19-test.pob-eng 51.9 0.74484 2100 56339 1.000
tico19-test.por-eng 51.9 0.74484 2100 56339 1.000
tico19-test.spa-eng 50.5 0.73239 2100 56339 1.000