- dataset: opusTCv20210807+bt
- model: transformer-big
- source language(s): bel bel_Latn rus ukr
- target language(s): bos_Cyrl bos_Latn bul hbs hbs_Cyrl hrv mkd slv srp_Cyrl srp_Latn
- raw source language(s): bel rus ukr
- raw target language(s): bos bul hbs hrv mkd slv srp
- 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+bt_transformer-big_2022-03-16.zip
- test set translations: opusTCv20210807+bt_transformer-big_2022-03-16.test.txt
- test set scores: opusTCv20210807+bt_transformer-big_2022-03-16.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| Tatoeba-test-v2021-08-07.bel-bul | 19.3 | 0.44655 | 1 | 6 | 1.000 |
| Tatoeba-test-v2021-08-07.bel-hbs | 50.5 | 0.67163 | 38 | 213 | 0.981 |
| Tatoeba-test-v2021-08-07.bel-mkd | 38.0 | 0.71758 | 1 | 6 | 1.000 |
| Tatoeba-test-v2021-08-07.bel-slv | 6.3 | 0.23166 | 12 | 73 | 0.899 |
| Tatoeba-test-v2021-08-07.bel-srp_Cyrl | 50.9 | 0.64865 | 22 | 137 | 0.985 |
| Tatoeba-test-v2021-08-07.bel-srp_Latn | 49.0 | 0.71404 | 16 | 76 | 0.973 |
| Tatoeba-test-v2021-08-07.multi-multi | 46.3 | 0.63906 | 7341 | 41890 | 0.971 |
| Tatoeba-test-v2021-08-07.rus-bos_Latn | 62.1 | 0.76976 | 12 | 54 | 1.000 |
| Tatoeba-test-v2021-08-07.rus-bul | 52.9 | 0.71365 | 1247 | 8239 | 0.952 |
| Tatoeba-test-v2021-08-07.rus-hbs | 49.3 | 0.69085 | 2500 | 14723 | 0.954 |
| Tatoeba-test-v2021-08-07.rus-hrv | 49.6 | 0.68766 | 124 | 723 | 0.978 |
| Tatoeba-test-v2021-08-07.rus-mkd | 48.9 | 0.82007 | 3 | 15 | 0.857 |
| Tatoeba-test-v2021-08-07.rus-slv | 21.3 | 0.38105 | 657 | 3969 | 0.996 |
| Tatoeba-test-v2021-08-07.rus-srp_Cyrl | 46.7 | 0.66723 | 881 | 5400 | 0.944 |
| Tatoeba-test-v2021-08-07.rus-srp_Latn | 50.9 | 0.70617 | 1483 | 8546 | 0.957 |
| Tatoeba-test-v2021-08-07.ukr-bul | 60.2 | 0.76823 | 1020 | 5181 | 0.984 |
| Tatoeba-test-v2021-08-07.ukr-hbs | 51.9 | 0.69886 | 942 | 5130 | 0.974 |
| Tatoeba-test-v2021-08-07.ukr-hrv | 50.6 | 0.68256 | 389 | 2302 | 0.978 |
| Tatoeba-test-v2021-08-07.ukr-mkd | 27.5 | 0.67226 | 5 | 22 | 1.000 |
| Tatoeba-test-v2021-08-07.ukr-slv | 14.9 | 0.29052 | 915 | 4265 | 1.000 |
| Tatoeba-test-v2021-08-07.ukr-srp_Cyrl | 53.6 | 0.69849 | 205 | 1112 | 0.963 |
| Tatoeba-test-v2021-08-07.ukr-srp_Latn | 52.6 | 0.72380 | 348 | 1716 | 0.976 |
- dataset: opusTCv20210807+bt
- model: transformer-big
- source language(s): bel bel_Latn rus ukr
- target language(s): bos_Cyrl bos_Latn bul hbs hbs_Cyrl hrv mkd slv srp_Cyrl srp_Latn
- raw source language(s): bel rus ukr
- raw target language(s): bos bul hbs hrv mkd slv srp
- 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+bt_transformer-big_2022-03-23.zip
- test set translations: opusTCv20210807+bt_transformer-big_2022-03-23.test.txt
- test set scores: opusTCv20210807+bt_transformer-big_2022-03-23.eval.txt
| testset | BLEU | chr-F | #sent | #words | BP |
|---|---|---|---|---|---|
| Tatoeba-test-v2021-08-07.bel-bul | 10.4 | 0.26328 | 1 | 6 | 0.819 |
| Tatoeba-test-v2021-08-07.bel-hbs | 51.0 | 0.69287 | 38 | 213 | 0.976 |
| Tatoeba-test-v2021-08-07.bel-mkd | 38.0 | 0.71758 | 1 | 6 | 1.000 |
| Tatoeba-test-v2021-08-07.bel-slv | 3.3 | 0.20892 | 12 | 73 | 0.899 |
| Tatoeba-test-v2021-08-07.bel-srp_Cyrl | 50.1 | 0.66070 | 22 | 137 | 0.978 |
| Tatoeba-test-v2021-08-07.bel-srp_Latn | 52.3 | 0.75239 | 16 | 76 | 0.973 |
| Tatoeba-test-v2021-08-07.multi-multi | 46.2 | 0.63754 | 7341 | 41890 | 0.971 |
| Tatoeba-test-v2021-08-07.rus-bos_Latn | 57.0 | 0.76347 | 12 | 54 | 1.000 |
| Tatoeba-test-v2021-08-07.rus-bul | 52.7 | 0.71217 | 1247 | 8239 | 0.951 |
| Tatoeba-test-v2021-08-07.rus-hbs | 49.1 | 0.68937 | 2500 | 14723 | 0.954 |
| Tatoeba-test-v2021-08-07.rus-hrv | 47.9 | 0.67774 | 124 | 723 | 0.969 |
| Tatoeba-test-v2021-08-07.rus-mkd | 46.5 | 0.80914 | 3 | 15 | 0.931 |
| Tatoeba-test-v2021-08-07.rus-slv | 21.5 | 0.38037 | 657 | 3969 | 1.000 |
| Tatoeba-test-v2021-08-07.rus-srp_Cyrl | 46.0 | 0.66398 | 881 | 5400 | 0.944 |
| Tatoeba-test-v2021-08-07.rus-srp_Latn | 51.1 | 0.70663 | 1483 | 8546 | 0.958 |
| Tatoeba-test-v2021-08-07.ukr-bul | 60.4 | 0.76820 | 1020 | 5181 | 0.986 |
| Tatoeba-test-v2021-08-07.ukr-hbs | 51.9 | 0.69314 | 942 | 5130 | 0.971 |
| Tatoeba-test-v2021-08-07.ukr-hrv | 50.1 | 0.67224 | 389 | 2302 | 0.971 |
| Tatoeba-test-v2021-08-07.ukr-mkd | 24.0 | 0.65445 | 5 | 22 | 1.000 |
| Tatoeba-test-v2021-08-07.ukr-slv | 14.5 | 0.28784 | 915 | 4265 | 1.000 |
| Tatoeba-test-v2021-08-07.ukr-srp_Cyrl | 54.7 | 0.69993 | 205 | 1112 | 0.957 |
| Tatoeba-test-v2021-08-07.ukr-srp_Latn | 52.9 | 0.72138 | 348 | 1716 | 0.981 |