Neurora/opus-tatoeba-zho-eng

Opus Tatoeba | Chinese -> English

• dataset: opus
• model: transformer
• source language(s): cjy cmn gan hak hsn lzh nan wuu yue
• target language(s): eng
• model: transformer
• pre-processing: normalization + SentencePiece (spm32k,spm32k)
• download: opus-2021-02-18.zip
• test set translations: opus-2021-02-18.test.txt
• test set scores: opus-2021-02-18.eval.txt

Benchmarks

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test.cjy_Hans-eng	20.9	0.394	2	18	0.751
Tatoeba-test.cjy_Hant-eng	16.0	0.173	1	4	1.000
Tatoeba-test.cmn-eng	17.0	0.480	15	82	1.000
Tatoeba-test.cmn_Hans-eng	36.7	0.558	4195	36106	0.935
Tatoeba-test.cmn_Hant-eng	40.4	0.585	4418	34780	0.955
Tatoeba-test.gan-eng	43.2	0.565	1	9	1.000
Tatoeba-test.lzh-eng	3.6	0.191	98	933	0.755
Tatoeba-test.lzh_Hans-eng	6.0	0.170	3	42	0.900
Tatoeba-test.nan-eng	6.1	0.179	2	10	1.000
Tatoeba-test.wuu-eng	17.4	0.400	203	1625	1.000
Tatoeba-test.yue_Hans-eng	21.1	0.407	630	5838	0.946
Tatoeba-test.yue_Hant-eng	25.2	0.428	431	3350	0.949
Tatoeba-test.zho-eng	36.0	0.546	9999	82822	0.946