eng-gla

opus-2021-02-11.zip

• dataset: opus
• model: transformer-align
• source language(s): eng
• target language(s): gla
• model: transformer-align
• pre-processing: normalization + SentencePiece (spm1k,spm1k)
• download: opus-2021-02-11.zip
• test set translations: opus-2021-02-11.test.txt
• test set scores: opus-2021-02-11.eval.txt

Benchmarks

testset	BLEU	chr-F
Tatoeba-test.eng.gla	0.4	0.159
Tatoeba-test.eng-gla.eng.gla	0.4	0.159

opus+bt-2021-03-09.zip

• dataset: opus+bt
• model: transformer-align
• source language(s): eng
• target language(s): gla
• model: transformer-align
• pre-processing: normalization + SentencePiece (spm1k,spm1k)
• download: opus+bt-2021-03-09.zip
• test set translations: opus+bt-2021-03-09.test.txt
• test set scores: opus+bt-2021-03-09.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test.eng-gla	0.5	0.162	917	7072	1.000