zle-ita

opusTCv20210807_transformer-big_2022-03-19.zip

• dataset: opusTCv20210807
• model: transformer-big
• source language(s): bel bel_Latn orv_Cyrl rus ukr
• target language(s): ita
• raw source language(s): bel orv rus ukr
• raw target language(s): ita
• model: transformer-big
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opusTCv20210807_transformer-big_2022-03-19.zip
• test set translations: opusTCv20210807_transformer-big_2022-03-19.test.txt
• test set scores: opusTCv20210807_transformer-big_2022-03-19.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test-v2021-08-07.bel-ita	48.5	0.65265	264	1681	0.991
Tatoeba-test-v2021-08-07.bel_Latn-ita	3.4	0.13902	1	8	1.000
Tatoeba-test-v2021-08-07.multi-ita	45.2	0.65367	10000	66592	0.958
Tatoeba-test-v2021-08-07.orv-ita	24.7	0.33756	8	53	0.942
Tatoeba-test-v2021-08-07.rus-ita	43.3	0.63958	10045	71507	0.956
Tatoeba-test-v2021-08-07.ukr-ita	49.7	0.69453	5000	27846	0.957

opusTCv20210807_transformer-big_2022-03-23.zip

• dataset: opusTCv20210807
• model: transformer-big
• source language(s): bel bel_Latn orv_Cyrl rus ukr
• target language(s): ita
• raw source language(s): bel orv rus ukr
• raw target language(s): ita
• model: transformer-big
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opusTCv20210807_transformer-big_2022-03-23.zip
• test set translations: opusTCv20210807_transformer-big_2022-03-23.test.txt
• test set scores: opusTCv20210807_transformer-big_2022-03-23.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test-v2021-08-07.bel-ita	47.6	0.64276	264	1681	0.996
Tatoeba-test-v2021-08-07.bel_Latn-ita	3.4	0.13902	1	8	1.000
Tatoeba-test-v2021-08-07.multi-ita	45.6	0.65617	10000	66592	0.957
Tatoeba-test-v2021-08-07.orv-ita	8.6	0.25517	8	53	0.901
Tatoeba-test-v2021-08-07.rus-ita	43.6	0.64139	10045	71507	0.956
Tatoeba-test-v2021-08-07.ukr-ita	50.2	0.69761	5000	27846	0.956