-
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
31 changed files
with
625 additions
and
18 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 +1,20 @@ | ||
| # opusTCv20210807+pbt_transformer-align_2022-03-07.zip | ||
|
|
||
| * dataset: opusTCv20210807+pbt | ||
| * model: transformer-align | ||
| * source language(s): dan | ||
| * target language(s): ukr | ||
| * raw source language(s): dan | ||
| * raw target language(s): ukr | ||
| * model: transformer-align | ||
| * pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| * download: [opusTCv20210807+pbt_transformer-align_2022-03-07.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/dan-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.zip) | ||
| * test set translations: [opusTCv20210807+pbt_transformer-align_2022-03-07.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/dan-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.test.txt) | ||
| * test set scores: [opusTCv20210807+pbt_transformer-align_2022-03-07.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/dan-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.eval.txt) | ||
|
|
||
| ## Benchmarks | ||
|
|
||
| | testset | BLEU | chr-F | #sent | #words | BP | | ||
| |---------|-------|-------|-------|--------|----| | ||
| | Tatoeba-test-v2021-08-07.dan-ukr | 52.6 | 0.71901 | 10 | 47 | 1.000 | | ||
|
|
34 changes: 34 additions & 0 deletions
34
models/dan-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.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 +1,34 @@ | ||
| release: dan-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.zip | ||
| release-date: 2022-03-07 | ||
| dataset-name: opusTCv20210807+pbt | ||
| modeltype: transformer-align | ||
| vocabulary: | ||
| source: opusTCv20210807+pbt.spm32k-spm32k.vocab.yml | ||
| target: opusTCv20210807+pbt.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: | ||
| - dan | ||
| target-languages: | ||
| - ukr | ||
| raw-source-languages: | ||
| - dan | ||
| raw-target-languages: | ||
| - ukr | ||
| training-data: | ||
| dan-ukr: Tatoeba-train-v2021-08-07.dan-ukr.strict (5377143) Tatoeba-train.eng-dan-ukr.aa (1000000) Tatoeba-train.eng-dan-ukr.ab (1000000) Tatoeba-train.eng-dan-ukr.ac (1000000) Tatoeba-train.eng-dan-ukr.ad (1000000) Tatoeba-train.eng-dan-ukr.ae (1000000) Tatoeba-train.eng-dan-ukr.af (1000000) Tatoeba-train.eng-dan-ukr.ag (1000000) Tatoeba-train.eng-dan-ukr.ah (1000000) Tatoeba-train.eng-dan-ukr.ai (1000000) Tatoeba-train.eng-dan-ukr.aj (1000000) Tatoeba-train.eng-dan-ukr.ak (1000000) Tatoeba-train.eng-dan-ukr.al (1000000) Tatoeba-train.eng-dan-ukr.am (1000000) Tatoeba-train.eng-dan-ukr.an (1000000) Tatoeba-train.eng-dan-ukr.ao (1000000) Tatoeba-train.eng-dan-ukr.ap (1000000) Tatoeba-train.eng-dan-ukr.aq (1000000) Tatoeba-train.eng-dan-ukr.ar (1000000) Tatoeba-train.eng-dan-ukr.as (1000000) Tatoeba-train.eng-dan-ukr.at (1000000) Tatoeba-train.eng-dan-ukr.au (1000000) Tatoeba-train.eng-dan-ukr.av (1000000) Tatoeba-train.eng-dan-ukr.aw (1000000) Tatoeba-train.eng-dan-ukr.ax (1000000) Tatoeba-train.eng-dan-ukr.ay (1000000) Tatoeba-train.eng-dan-ukr.az (1000000) Tatoeba-train.eng-dan-ukr.ba (1000000) Tatoeba-train.eng-dan-ukr.bb (1000000) Tatoeba-train.eng-dan-ukr.bc (1000000) Tatoeba-train.eng-dan-ukr.bd (1000000) Tatoeba-train.eng-dan-ukr.be (1000000) Tatoeba-train.eng-dan-ukr.bf (1000000) Tatoeba-train.eng-dan-ukr.bg (1000000) Tatoeba-train.eng-dan-ukr.bh (220604) | ||
| validation-data: | ||
| dan-ukr: Tatoeba-dev-v2021-08-07, 1001 | ||
| total-size-shuffled: 1001 | ||
| devset-selected: top 1001 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| Tatoeba-test-v2021-08-07.dan-ukr: 10/47 | ||
| BLEU-scores: | ||
| Tatoeba-test-v2021-08-07.dan-ukr: 52.6 | ||
| chr-F-scores: | ||
| Tatoeba-test-v2021-08-07.dan-ukr: 0.71901 |
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
34 changes: 34 additions & 0 deletions
34
models/deu-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.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 +1,34 @@ | ||
| release: deu-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.zip | ||
| release-date: 2022-03-07 | ||
| dataset-name: opusTCv20210807+pbt | ||
| modeltype: transformer-align | ||
| vocabulary: | ||
| source: opusTCv20210807+pbt.spm32k-spm32k.vocab.yml | ||
| target: opusTCv20210807+pbt.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: | ||
| - deu | ||
| target-languages: | ||
| - ukr | ||
| raw-source-languages: | ||
| - deu | ||
| raw-target-languages: | ||
| - ukr | ||
| training-data: | ||
| deu-ukr: Tatoeba-train-v2021-08-07.deu-ukr.strict (11133937) Tatoeba-train.eng-deu-ukr.aa (1000000) Tatoeba-train.eng-deu-ukr.ab (1000000) Tatoeba-train.eng-deu-ukr.ac (1000000) Tatoeba-train.eng-deu-ukr.ad (1000000) Tatoeba-train.eng-deu-ukr.ae (1000000) Tatoeba-train.eng-deu-ukr.af (1000000) Tatoeba-train.eng-deu-ukr.ag (1000000) Tatoeba-train.eng-deu-ukr.ah (1000000) Tatoeba-train.eng-deu-ukr.ai (1000000) Tatoeba-train.eng-deu-ukr.aj (1000000) Tatoeba-train.eng-deu-ukr.ak (1000000) Tatoeba-train.eng-deu-ukr.al (1000000) Tatoeba-train.eng-deu-ukr.am (1000000) Tatoeba-train.eng-deu-ukr.an (1000000) Tatoeba-train.eng-deu-ukr.ao (1000000) Tatoeba-train.eng-deu-ukr.ap (1000000) Tatoeba-train.eng-deu-ukr.aq (1000000) Tatoeba-train.eng-deu-ukr.ar (1000000) Tatoeba-train.eng-deu-ukr.as (1000000) Tatoeba-train.eng-deu-ukr.at (1000000) Tatoeba-train.eng-deu-ukr.au (1000000) Tatoeba-train.eng-deu-ukr.av (1000000) Tatoeba-train.eng-deu-ukr.aw (1000000) Tatoeba-train.eng-deu-ukr.ax (1000000) Tatoeba-train.eng-deu-ukr.ay (1000000) Tatoeba-train.eng-deu-ukr.az (1000000) Tatoeba-train.eng-deu-ukr.ba (1000000) Tatoeba-train.eng-deu-ukr.bb (1000000) Tatoeba-train.eng-deu-ukr.bc (1000000) Tatoeba-train.eng-deu-ukr.bd (1000000) Tatoeba-train.eng-deu-ukr.be (1000000) Tatoeba-train.eng-deu-ukr.bf (1000000) Tatoeba-train.eng-deu-ukr.bg (1000000) Tatoeba-train.eng-deu-ukr.bh (220604) | ||
| validation-data: | ||
| deu-ukr: Tatoeba-dev-v2021-08-07, 11580 | ||
| total-size-shuffled: 11580 | ||
| devset-selected: top 5000 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| Tatoeba-test-v2021-08-07.deu-ukr: 10000/54528 | ||
| BLEU-scores: | ||
| Tatoeba-test-v2021-08-07.deu-ukr: 39.1 | ||
| chr-F-scores: | ||
| Tatoeba-test-v2021-08-07.deu-ukr: 0.61779 |
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 +1,20 @@ | ||
| # opusTCv20210807+pbt_transformer-align_2022-03-07.zip | ||
|
|
||
| * dataset: opusTCv20210807+pbt | ||
| * model: transformer-align | ||
| * source language(s): fin | ||
| * target language(s): ukr | ||
| * raw source language(s): fin | ||
| * raw target language(s): ukr | ||
| * model: transformer-align | ||
| * pre-processing: normalization + SentencePiece (spm32k,spm32k) | ||
| * download: [opusTCv20210807+pbt_transformer-align_2022-03-07.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/fin-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.zip) | ||
| * test set translations: [opusTCv20210807+pbt_transformer-align_2022-03-07.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/fin-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.test.txt) | ||
| * test set scores: [opusTCv20210807+pbt_transformer-align_2022-03-07.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/fin-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.eval.txt) | ||
|
|
||
| ## Benchmarks | ||
|
|
||
| | testset | BLEU | chr-F | #sent | #words | BP | | ||
| |---------|-------|-------|-------|--------|----| | ||
| | Tatoeba-test-v2021-08-07.fin-ukr | 40.1 | 0.58239 | 33 | 218 | 0.953 | | ||
|
|
34 changes: 34 additions & 0 deletions
34
models/fin-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.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 +1,34 @@ | ||
| release: fin-ukr/opusTCv20210807+pbt_transformer-align_2022-03-07.zip | ||
| release-date: 2022-03-07 | ||
| dataset-name: opusTCv20210807+pbt | ||
| modeltype: transformer-align | ||
| vocabulary: | ||
| source: opusTCv20210807+pbt.spm32k-spm32k.vocab.yml | ||
| target: opusTCv20210807+pbt.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: | ||
| - ukr | ||
| raw-source-languages: | ||
| - fin | ||
| raw-target-languages: | ||
| - ukr | ||
| training-data: | ||
| fin-ukr: Tatoeba-train-v2021-08-07.fin-ukr.strict (5400101) Tatoeba-train.eng-fin-ukr.aa (1000000) Tatoeba-train.eng-fin-ukr.ab (1000000) Tatoeba-train.eng-fin-ukr.ac (1000000) Tatoeba-train.eng-fin-ukr.ad (1000000) Tatoeba-train.eng-fin-ukr.ae (1000000) Tatoeba-train.eng-fin-ukr.af (1000000) Tatoeba-train.eng-fin-ukr.ag (1000000) Tatoeba-train.eng-fin-ukr.ah (1000000) Tatoeba-train.eng-fin-ukr.ai (1000000) Tatoeba-train.eng-fin-ukr.aj (1000000) Tatoeba-train.eng-fin-ukr.ak (1000000) Tatoeba-train.eng-fin-ukr.al (1000000) Tatoeba-train.eng-fin-ukr.am (1000000) Tatoeba-train.eng-fin-ukr.an (1000000) Tatoeba-train.eng-fin-ukr.ao (1000000) Tatoeba-train.eng-fin-ukr.ap (1000000) Tatoeba-train.eng-fin-ukr.aq (1000000) Tatoeba-train.eng-fin-ukr.ar (1000000) Tatoeba-train.eng-fin-ukr.as (1000000) Tatoeba-train.eng-fin-ukr.at (1000000) Tatoeba-train.eng-fin-ukr.au (1000000) Tatoeba-train.eng-fin-ukr.av (1000000) Tatoeba-train.eng-fin-ukr.aw (1000000) Tatoeba-train.eng-fin-ukr.ax (1000000) Tatoeba-train.eng-fin-ukr.ay (1000000) Tatoeba-train.eng-fin-ukr.az (1000000) Tatoeba-train.eng-fin-ukr.ba (1000000) Tatoeba-train.eng-fin-ukr.bb (1000000) Tatoeba-train.eng-fin-ukr.bc (1000000) Tatoeba-train.eng-fin-ukr.bd (1000000) Tatoeba-train.eng-fin-ukr.be (1000000) Tatoeba-train.eng-fin-ukr.bf (1000000) Tatoeba-train.eng-fin-ukr.bg (1000000) Tatoeba-train.eng-fin-ukr.bh (220604) | ||
| validation-data: | ||
| fin-ukr: Tatoeba-dev-v2021-08-07, 1000 | ||
| total-size-shuffled: 1000 | ||
| devset-selected: top 1000 lines of Tatoeba-dev-v2021-08-07.src.shuffled | ||
| test-data: | ||
| Tatoeba-test-v2021-08-07.fin-ukr: 33/218 | ||
| BLEU-scores: | ||
| Tatoeba-test-v2021-08-07.fin-ukr: 40.1 | ||
| chr-F-scores: | ||
| Tatoeba-test-v2021-08-07.fin-ukr: 0.58239 |
Large diffs are not rendered by default.
Oops, something went wrong.
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
Oops, something went wrong.