-
Notifications
You must be signed in to change notification settings - Fork 92
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
44 changed files
with
1,165 additions
and
3,165 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+news+bt_transformer-big_2023-04-13.yml |
58 changes: 58 additions & 0 deletions
58
models/eng-fin/opusTCv20210807+news+bt_transformer-big_2023-04-13.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,58 @@ | ||
| release: eng-fin/opusTCv20210807+news+bt_transformer-big_2023-04-13.zip | ||
| release-date: 2023-04-13 | ||
| dataset-name: opusTCv20210807+news+bt | ||
| modeltype: transformer-big | ||
| vocabulary: | ||
| source: opusTCv20210807+news+bt.spm32k-spm32k.vocab.yml | ||
| target: opusTCv20210807+news+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.eng-fin.strict (94159007) news.2014.fi.shuffled.deduped.aa.fin-eng (999031) news.2014.fi.shuffled.deduped.ab.fin-eng (626224) news.2015.fi.shuffled.deduped.aa.fin-eng (997577) news.2015.fi.shuffled.deduped.ab.fin-eng (997625) news.2015.fi.shuffled.deduped.ac.fin-eng (997660) news.2015.fi.shuffled.deduped.ad.fin-eng (997617) news.2015.fi.shuffled.deduped.ae.fin-eng (731994) news.2016.fi.shuffled.deduped.aa.fin-eng (996440) news.2016.fi.shuffled.deduped.ab.fin-eng (996450) news.2016.fi.shuffled.deduped.ac.fin-eng (996387) news.2016.fi.shuffled.deduped.ad.fin-eng (868304) news.2017.fi.shuffled.deduped.aa.fin-eng (996108) news.2017.fi.shuffled.deduped.ab.fin-eng (995969) news.2017.fi.shuffled.deduped.ac.fin-eng (995930) news.2017.fi.shuffled.deduped.ad.fin-eng (634818) news.2018.fi.shuffled.deduped.aa.fin-eng (997253) news.2018.fi.shuffled.deduped.ab.fin-eng (997260) news.2018.fi.shuffled.deduped.ac.fin-eng (962216) news.2019.fi.shuffled.deduped.aa.fin-eng (995768) news.2019.fi.shuffled.deduped.ab.fin-eng (995822) news.2019.fi.shuffled.deduped.ac.fin-eng (995858) news.2019.fi.shuffled.deduped.ad.fin-eng (995653) news.2019.fi.shuffled.deduped.ae.fin-eng (683615) news.2020.fi.shuffled.deduped.aa.fin-eng (998418) news.2020.fi.shuffled.deduped.ab.fin-eng (998439) news.2020.fi.shuffled.deduped.ac.fin-eng (998422) news.2020.fi.shuffled.deduped.ad.fin-eng (998462) news.2020.fi.shuffled.deduped.ae.fin-eng (51508) news.2021.fi.shuffled.deduped.aa.fin-eng (999151) news.2021.fi.shuffled.deduped.ab.fin-eng (999125) news.2021.fi.shuffled.deduped.ac.fin-eng (999103) news.2021.fi.shuffled.deduped.ad.fin-eng (202014) 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, 67377 | ||
| total-size-shuffled: 67377 | ||
| 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 | ||
| newstestB2017-enfi.eng-fin: 3002/45506 | ||
| Tatoeba-test-v2021-08-07.eng-fin: 10000/60832 | ||
| BLEU-scores: | ||
| newsdev2015-enfi.eng-fin: 26.5 | ||
| newstest2015-enfi.eng-fin: 28.9 | ||
| newstest2016-enfi.eng-fin: 30.6 | ||
| newstest2017-enfi.eng-fin: 33.1 | ||
| newstest2018-enfi.eng-fin: 22.1 | ||
| newstest2019-enfi.eng-fin: 26.0 | ||
| newstestB2016-enfi.eng-fin: 24.8 | ||
| newstestB2017-enfi.eng-fin: 28.5 | ||
| Tatoeba-test-v2021-08-07.eng-fin: 40.1 | ||
| chr-F-scores: | ||
| newsdev2015-enfi.eng-fin: 0.59380 | ||
| newstest2015-enfi.eng-fin: 0.61202 | ||
| newstest2016-enfi.eng-fin: 0.62472 | ||
| newstest2017-enfi.eng-fin: 0.65437 | ||
| newstest2018-enfi.eng-fin: 0.56551 | ||
| newstest2019-enfi.eng-fin: 0.58341 | ||
| newstestB2016-enfi.eng-fin: 0.58546 | ||
| newstestB2017-enfi.eng-fin: 0.61445 | ||
| Tatoeba-test-v2021-08-07.eng-fin: 0.64873 |
1 change: 1 addition & 0 deletions
1
models/eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-03-13.yml |
40 changes: 40 additions & 0 deletions
40
models/eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-03-13.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,40 @@ | ||
| release: eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-03-13.zip | ||
| release-date: 2023-03-13 | ||
| dataset-name: opusTCv20210807+nopar+ft95 | ||
| modeltype: transformer-small-align | ||
| vocabulary: | ||
| source: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.src.vocab | ||
| target: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.trg.vocab | ||
| pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| subwords: | ||
| source: spm32k | ||
| target: spm32k | ||
| subword-models: | ||
| source: source.spm | ||
| target: target.spm | ||
| source-languages: | ||
| - eus | ||
| target-languages: | ||
| - eng | ||
| raw-source-languages: | ||
| - eus | ||
| raw-target-languages: | ||
| - eng | ||
| training-data: | ||
| eus-eng: Tatoeba-train.eus.best95.gz (5907766) | ||
| validation-data: | ||
| eng-eus: Tatoeba-dev-v2021-08-07, 990 | ||
| total-size-shuffled: 990 | ||
| devset-selected: top 990 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 1060/8157 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 1060/8157 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 1060/8157 | ||
| BLEU-scores: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 51.6 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 51.0 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 51.0 | ||
| chr-F-scores: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 0.67608 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 0.67334 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 0.67323 |
1 change: 1 addition & 0 deletions
1
models/eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-03-13.yml |
40 changes: 40 additions & 0 deletions
40
models/eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-03-13.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,40 @@ | ||
| release: eus-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-03-13.zip | ||
| release-date: 2023-03-13 | ||
| dataset-name: opusTCv20210807+nopar+ft95 | ||
| modeltype: transformer-tiny11-align | ||
| vocabulary: | ||
| source: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.src.vocab | ||
| target: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.trg.vocab | ||
| pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| subwords: | ||
| source: spm32k | ||
| target: spm32k | ||
| subword-models: | ||
| source: source.spm | ||
| target: target.spm | ||
| source-languages: | ||
| - eus | ||
| target-languages: | ||
| - eng | ||
| raw-source-languages: | ||
| - eus | ||
| raw-target-languages: | ||
| - eng | ||
| training-data: | ||
| eus-eng: Tatoeba-train.eus.best95.gz (5907766) | ||
| validation-data: | ||
| eng-eus: Tatoeba-dev-v2021-08-07, 990 | ||
| total-size-shuffled: 990 | ||
| devset-selected: top 990 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 1060/8157 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 1060/8157 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 1060/8157 | ||
| BLEU-scores: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 51.4 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 50.5 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 50.5 | ||
| chr-F-scores: | ||
| Tatoeba-test-v2021-08-07.eus-eng: 0.67570 | ||
| Tatoeba-test-v2021-08-07.intgemm8.eus-eng: 0.66861 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.eus-eng: 0.66828 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+news+bt_transformer-big_2023-04-13.yml |
58 changes: 58 additions & 0 deletions
58
models/fin-eng/opusTCv20210807+news+bt_transformer-big_2023-04-13.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,58 @@ | ||
| release: fin-eng/opusTCv20210807+news+bt_transformer-big_2023-04-13.zip | ||
| release-date: 2023-04-13 | ||
| dataset-name: opusTCv20210807+news+bt | ||
| modeltype: transformer-big | ||
| vocabulary: | ||
| source: opusTCv20210807+news+bt.spm32k-spm32k.vocab.yml | ||
| target: opusTCv20210807+news+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: | ||
| - fin | ||
| target-languages: | ||
| - eng | ||
| raw-source-languages: | ||
| - fin | ||
| raw-target-languages: | ||
| - eng | ||
| training-data: | ||
| fin-eng: Tatoeba-train-v2021-08-07.eng-fin.strict (94159007) news.2007.en.shuffled.deduped.aa (995827) news.2007.en.shuffled.deduped.ab (995744) news.2007.en.shuffled.deduped.ac (995730) news.2007.en.shuffled.deduped.ad (386386) news.2008.en.shuffled.deduped.aa (996856) news.2008.en.shuffled.deduped.ab (996897) news.2008.en.shuffled.deduped.ac (996857) news.2008.en.shuffled.deduped.ad (996952) news.2008.en.shuffled.deduped.ae (996898) news.2008.en.shuffled.deduped.af (996867) news.2008.en.shuffled.deduped.ag (996886) news.2008.en.shuffled.deduped.ah (996948) news.2008.en.shuffled.deduped.ai (996900) news.2008.en.shuffled.deduped.aj (996894) news.2008.en.shuffled.deduped.ak (996887) news.2008.en.shuffled.deduped.al (745791) news.2009.en.shuffled.deduped.aa (997324) news.2009.en.shuffled.deduped.ab (997307) news.2009.en.shuffled.deduped.ac (997323) news.2009.en.shuffled.deduped.ad (997343) news.2009.en.shuffled.deduped.ae (997294) news.2009.en.shuffled.deduped.af (997357) news.2009.en.shuffled.deduped.ag (997255) news.2009.en.shuffled.deduped.ah (997300) news.2009.en.shuffled.deduped.ai (997332) news.2009.en.shuffled.deduped.aj (997345) news.2009.en.shuffled.deduped.ak (997374) news.2009.en.shuffled.deduped.al (997346) news.2009.en.shuffled.deduped.am (997261) news.2009.en.shuffled.deduped.an (997283) news.2009.en.shuffled.deduped.ao (401476) news.2010.en.shuffled.deduped.aa (998438) news.2010.en.shuffled.deduped.ab (998404) news.2010.en.shuffled.deduped.ac (998443) news.2010.en.shuffled.deduped.ad (998426) news.2010.en.shuffled.deduped.ae (998435) news.2010.en.shuffled.deduped.af (998463) news.2010.en.shuffled.deduped.ag (998361) news.2010.en.shuffled.deduped.ah (271688) news.2011.en.shuffled.deduped.aa (998625) news.2011.en.shuffled.deduped.ab (998559) news.2011.en.shuffled.deduped.ac (998533) news.2011.en.shuffled.deduped.ad (998596) news.2011.en.shuffled.deduped.ae (998549) news.2011.en.shuffled.deduped.af (998598) news.2011.en.shuffled.deduped.ag (998674) news.2011.en.shuffled.deduped.ah (998580) news.2011.en.shuffled.deduped.ai (998556) news.2011.en.shuffled.deduped.aj (998554) news.2011.en.shuffled.deduped.ak (998534) news.2011.en.shuffled.deduped.al (998550) news.2011.en.shuffled.deduped.am (998530) news.2011.en.shuffled.deduped.an (998627) news.2011.en.shuffled.deduped.ao (106910) news.2013.en.shuffled.deduped.aa (996357) news.2013.en.shuffled.deduped.ab (996368) news.2013.en.shuffled.deduped.ac (996368) news.2013.en.shuffled.deduped.ad (996394) news.2013.en.shuffled.deduped.ae (996367) news.2013.en.shuffled.deduped.af (996372) news.2013.en.shuffled.deduped.ag (996448) news.2013.en.shuffled.deduped.ah (996554) news.2013.en.shuffled.deduped.ai (996289) news.2013.en.shuffled.deduped.aj (996254) news.2013.en.shuffled.deduped.ak (996415) news.2013.en.shuffled.deduped.al (996342) news.2013.en.shuffled.deduped.am (996425) news.2013.en.shuffled.deduped.an (996502) news.2013.en.shuffled.deduped.ao (996299) news.2013.en.shuffled.deduped.ap (996317) news.2013.en.shuffled.deduped.aq (996376) news.2013.en.shuffled.deduped.ar (996332) news.2013.en.shuffled.deduped.as (996482) news.2013.en.shuffled.deduped.at (996391) news.2013.en.shuffled.deduped.au (996432) news.2013.en.shuffled.deduped.av (996332) news.2013.en.shuffled.deduped.aw (996313) news.2013.en.shuffled.deduped.ax (359771) news.2021.en.shuffled.deduped.aa (976234) news.2021.en.shuffled.deduped.ab (976038) news.2021.en.shuffled.deduped.ac (976001) news.2021.en.shuffled.deduped.ad (976334) news.2021.en.shuffled.deduped.ae (976322) news.2021.en.shuffled.deduped.af (976467) news.2021.en.shuffled.deduped.ag (976342) news.2021.en.shuffled.deduped.ah (976613) news.2021.en.shuffled.deduped.ai (976392) news.2021.en.shuffled.deduped.aj (976450) news.2021.en.shuffled.deduped.ak (976262) news.2021.en.shuffled.deduped.al (976579) news.2021.en.shuffled.deduped.am (976079) news.2021.en.shuffled.deduped.an (976532) news.2021.en.shuffled.deduped.ao (975872) news.2021.en.shuffled.deduped.ap (976024) news.2021.en.shuffled.deduped.aq (976142) news.2021.en.shuffled.deduped.ar (976316) news.2021.en.shuffled.deduped.as (976558) news.2021.en.shuffled.deduped.at (976068) news.2021.en.shuffled.deduped.au (976350) news.2021.en.shuffled.deduped.av (976205) news.2021.en.shuffled.deduped.aw (976298) news.2021.en.shuffled.deduped.ax (976405) news.2021.en.shuffled.deduped.ay (976341) news.2021.en.shuffled.deduped.az (976104) news.2021.en.shuffled.deduped.ba (976291) news.2021.en.shuffled.deduped.bb (976046) news.2021.en.shuffled.deduped.bc (975935) news.2021.en.shuffled.deduped.bd (976260) news.2021.en.shuffled.deduped.be (976147) news.2021.en.shuffled.deduped.bf (976419) news.2021.en.shuffled.deduped.bg (976416) news.2021.en.shuffled.deduped.bh (976403) news.2021.en.shuffled.deduped.bi (976169) news.2021.en.shuffled.deduped.bj (976598) news.2021.en.shuffled.deduped.bk (976656) news.2021.en.shuffled.deduped.bl (976495) news.2021.en.shuffled.deduped.bm (976319) news.2021.en.shuffled.deduped.bn (352694) wikibooks.aa (992334) wikinews.aa (457233) wikipedia.aa (983013) wikipedia.ab (983150) wikipedia.ac (983127) wikipedia.ad (982918) wikiquote.aa (997039) | ||
| validation-data: | ||
| eng-fin: Tatoeba-dev-v2021-08-07, 67377 | ||
| total-size-shuffled: 67377 | ||
| devset-selected: top 5000 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| newsdev2015-enfi.fin-eng: 1500/32104 | ||
| newstest2015-enfi.fin-eng: 1370/27356 | ||
| newstest2016-enfi.fin-eng: 3000/63043 | ||
| newstest2017-enfi.fin-eng: 3002/61936 | ||
| newstest2018-enfi.fin-eng: 3000/62325 | ||
| newstest2019-fien.fin-eng: 1996/36227 | ||
| newstestB2016-enfi.fin-eng: 3000/63043 | ||
| newstestB2017-enfi.fin-eng: 3002/61936 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 10000/75524 | ||
| BLEU-scores: | ||
| newsdev2015-enfi.fin-eng: 29.1 | ||
| newstest2015-enfi.fin-eng: 30.3 | ||
| newstest2016-enfi.fin-eng: 36.3 | ||
| newstest2017-enfi.fin-eng: 38.4 | ||
| newstest2018-enfi.fin-eng: 26.9 | ||
| newstest2019-fien.fin-eng: 30.9 | ||
| newstestB2016-enfi.fin-eng: 28.7 | ||
| newstestB2017-enfi.fin-eng: 32.3 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 56.4 | ||
| chr-F-scores: | ||
| newsdev2015-enfi.fin-eng: 0.56442 | ||
| newstest2015-enfi.fin-eng: 0.57413 | ||
| newstest2016-enfi.fin-eng: 0.61780 | ||
| newstest2017-enfi.fin-eng: 0.63566 | ||
| newstest2018-enfi.fin-eng: 0.54142 | ||
| newstest2019-fien.fin-eng: 0.57880 | ||
| newstestB2016-enfi.fin-eng: 0.55994 | ||
| newstestB2017-enfi.fin-eng: 0.59170 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 0.71708 |
1 change: 1 addition & 0 deletions
1
models/fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-07-03.yml |
67 changes: 67 additions & 0 deletions
67
models/fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-07-03.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,67 @@ | ||
| release: fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-07-03.zip | ||
| release-date: 2023-07-03 | ||
| dataset-name: opusTCv20210807+nopar+ft95 | ||
| modeltype: transformer-small-align | ||
| vocabulary: | ||
| source: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.src.vocab | ||
| target: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.trg.vocab | ||
| pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| subwords: | ||
| source: spm32k | ||
| target: spm32k | ||
| subword-models: | ||
| source: source.spm | ||
| target: target.spm | ||
| source-languages: | ||
| - fin | ||
| target-languages: | ||
| - eng | ||
| raw-source-languages: | ||
| - fin | ||
| raw-target-languages: | ||
| - eng | ||
| training-data: | ||
| fin-eng: Tatoeba-train.fin.best95.gz (68560504) | ||
| validation-data: | ||
| eng-fin: Tatoeba-dev-v2021-08-07, 67377 | ||
| total-size-shuffled: 67377 | ||
| devset-selected: top 5000 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| newsdev2015-enfi.fin-eng: 1500/32104 | ||
| newstest2015-enfi.fin-eng: 1370/27356 | ||
| newstest2016-enfi.fin-eng: 3000/63043 | ||
| newstest2017-enfi.fin-eng: 3002/61936 | ||
| newstest2018-enfi.fin-eng: 3000/62325 | ||
| newstest2019-fien.fin-eng: 1996/36227 | ||
| newstestB2016-enfi.fin-eng: 3000/63043 | ||
| newstestB2017-enfi.fin-eng: 3002/61936 | ||
| newstestB2017-fien.fin-eng: 3002/61936 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 10000/75524 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 10000/75524 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 10000/75524 | ||
| BLEU-scores: | ||
| newsdev2015-enfi.fin-eng: 27.7 | ||
| newstest2015-enfi.fin-eng: 29.1 | ||
| newstest2016-enfi.fin-eng: 32.6 | ||
| newstest2017-enfi.fin-eng: 35.0 | ||
| newstest2018-enfi.fin-eng: 25.5 | ||
| newstest2019-fien.fin-eng: 31.1 | ||
| newstestB2016-enfi.fin-eng: 26.7 | ||
| newstestB2017-enfi.fin-eng: 29.8 | ||
| newstestB2017-fien.fin-eng: 29.8 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 54.8 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 54.1 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 54.1 | ||
| chr-F-scores: | ||
| newsdev2015-enfi.fin-eng: 0.55510 | ||
| newstest2015-enfi.fin-eng: 0.56598 | ||
| newstest2016-enfi.fin-eng: 0.59247 | ||
| newstest2017-enfi.fin-eng: 0.61491 | ||
| newstest2018-enfi.fin-eng: 0.53101 | ||
| newstest2019-fien.fin-eng: 0.58418 | ||
| newstestB2016-enfi.fin-eng: 0.54433 | ||
| newstestB2017-enfi.fin-eng: 0.57576 | ||
| newstestB2017-fien.fin-eng: 0.57576 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 0.70645 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 0.70255 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 0.70198 |
1 change: 1 addition & 0 deletions
1
models/fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-07-03.yml |
67 changes: 67 additions & 0 deletions
67
models/fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-07-03.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,67 @@ | ||
| release: fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-tiny11-align_2023-07-03.zip | ||
| release-date: 2023-07-03 | ||
| dataset-name: opusTCv20210807+nopar+ft95 | ||
| modeltype: transformer-tiny11-align | ||
| vocabulary: | ||
| source: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.src.vocab | ||
| target: opusTCv20210807+nopar+ft95-sepvoc.spm32k-spm32k.trg.vocab | ||
| pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| subwords: | ||
| source: spm32k | ||
| target: spm32k | ||
| subword-models: | ||
| source: source.spm | ||
| target: target.spm | ||
| source-languages: | ||
| - fin | ||
| target-languages: | ||
| - eng | ||
| raw-source-languages: | ||
| - fin | ||
| raw-target-languages: | ||
| - eng | ||
| training-data: | ||
| fin-eng: Tatoeba-train.fin.best95.gz (68560504) | ||
| validation-data: | ||
| eng-fin: Tatoeba-dev-v2021-08-07, 67377 | ||
| total-size-shuffled: 67377 | ||
| devset-selected: top 5000 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| newsdev2015-enfi.fin-eng: 1500/32104 | ||
| newstest2015-enfi.fin-eng: 1370/27356 | ||
| newstest2016-enfi.fin-eng: 3000/63043 | ||
| newstest2017-enfi.fin-eng: 3002/61936 | ||
| newstest2018-enfi.fin-eng: 3000/62325 | ||
| newstest2019-fien.fin-eng: 1996/36227 | ||
| newstestB2016-enfi.fin-eng: 3000/63043 | ||
| newstestB2017-enfi.fin-eng: 3002/61936 | ||
| newstestB2017-fien.fin-eng: 3002/61936 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 10000/75524 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 10000/75524 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 10000/75524 | ||
| BLEU-scores: | ||
| newsdev2015-enfi.fin-eng: 26.4 | ||
| newstest2015-enfi.fin-eng: 27.2 | ||
| newstest2016-enfi.fin-eng: 30.5 | ||
| newstest2017-enfi.fin-eng: 33.3 | ||
| newstest2018-enfi.fin-eng: 24.1 | ||
| newstest2019-fien.fin-eng: 30.2 | ||
| newstestB2016-enfi.fin-eng: 25.0 | ||
| newstestB2017-enfi.fin-eng: 28.2 | ||
| newstestB2017-fien.fin-eng: 28.2 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 52.5 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 51.8 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 51.7 | ||
| chr-F-scores: | ||
| newsdev2015-enfi.fin-eng: 0.54342 | ||
| newstest2015-enfi.fin-eng: 0.55250 | ||
| newstest2016-enfi.fin-eng: 0.57846 | ||
| newstest2017-enfi.fin-eng: 0.60025 | ||
| newstest2018-enfi.fin-eng: 0.52001 | ||
| newstest2019-fien.fin-eng: 0.57468 | ||
| newstestB2016-enfi.fin-eng: 0.53223 | ||
| newstestB2017-enfi.fin-eng: 0.56231 | ||
| newstestB2017-fien.fin-eng: 0.56231 | ||
| Tatoeba-test-v2021-08-07.fin-eng: 0.69073 | ||
| Tatoeba-test-v2021-08-07.intgemm8.fin-eng: 0.68494 | ||
| Tatoeba-test-v2021-08-07.intgemm8.shortlist.fin-eng: 0.68448 |
1 change: 1 addition & 0 deletions
1
models/glg-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align.yml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1 @@ | ||
| opusTCv20210807+nopar+ft95-sepvoc_transformer-small-align_2023-03-13.yml |
Oops, something went wrong.