README.md

opus1m+bt-2021-05-01.zip

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test.bod-eng	1.1	0.127	38	227	1.000
Tatoeba-test.brx-eng	0.8	0.131	12	57	1.000
Tatoeba-test.brx_Latn-eng	3.1	0.202	3	24	0.819
Tatoeba-test.cjy_Hans-eng	20.9	0.390	2	18	0.751
Tatoeba-test.cjy_Hant-eng	16.0	0.191	1	4	1.000
Tatoeba-test.cmn-eng	44.6	0.582	15	82	1.000
Tatoeba-test.cmn_Hans-eng	31.7	0.517	4195	36106	0.929
Tatoeba-test.cmn_Hant-eng	35.3	0.540	4418	34780	0.953
Tatoeba-test.dng-eng	0.7	0.060	6	30	1.000
Tatoeba-test.gan-eng	35.5	0.461	1	9	1.000
Tatoeba-test.lzh-eng	4.3	0.166	98	933	0.713
Tatoeba-test.lzh_Hans-eng	6.5	0.116	3	42	1.000
Tatoeba-test.multi-eng	28.8	0.476	10000	80668	0.985
Tatoeba-test.mya-eng	15.4	0.382	216	1659	1.000
Tatoeba-test.nan-eng	5.5	0.094	2	10	1.000
Tatoeba-test.nst-eng	0.4	0.113	805	4720	1.000
Tatoeba-test.wuu-eng	16.2	0.357	203	1625	1.000
Tatoeba-test.yue_Hans-eng	12.9	0.334	630	5838	1.000
Tatoeba-test.yue_Hant-eng	19.4	0.377	431	3350	0.978
Tatoeba-test.zho-eng	30.8	0.501	10000	82801	0.947
tico19-test.mya-eng	10.5	0.352	2100	56848	0.982

dataset: opus4m+btTCv20210807
model: transformer
source language(s): bod brx cjy cmn cnh dng dzo gan hak hsn iii kac ksw lus lzh mni mya nan new nst pck wuu yue
target language(s): eng
model: transformer
pre-processing: normalization + SentencePiece (spm32k,spm32k)
download: opus4m+btTCv20210807-2021-10-01.zip
test set translations: opus4m+btTCv20210807-2021-10-01.test.txt
test set scores: opus4m+btTCv20210807-2021-10-01.eval.txt

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test-v2021-08-07.multi-eng	27.9	0.472	10000	80747	1.000
Tatoeba-test-v2021-08-07.multi-multi	27.9	0.472	10000	80747	1.000
tico19-test.mya-eng	22.6	0.501	2100	56848	1.000