Name		Name	Last commit message	Last commit date
parent directory ..
README.md		README.md
opusTCv20210807+pft_transformer-big.yml		opusTCv20210807+pft_transformer-big.yml
opusTCv20210807+pft_transformer-big_2022-03-14.yml		opusTCv20210807+pft_transformer-big_2022-03-14.yml
opusTCv20210807+xb+bt+pft_transformer-big.yml		opusTCv20210807+xb+bt+pft_transformer-big.yml
opusTCv20210807+xb+bt+pft_transformer-big_2022-09-15.yml		opusTCv20210807+xb+bt+pft_transformer-big_2022-09-15.yml

README.md

opusTCv20210807+pft_transformer-big_2022-03-14.zip

dataset: opusTCv20210807+pft
model: transformer-big
source language(s): bel bel_Latn rus ukr
target language(s): dan isl nno nob swe
raw source language(s): bel rus ukr
raw target language(s): dan isl nno nob swe
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+pft_transformer-big_2022-03-14.zip
test set translations: opusTCv20210807+pft_transformer-big_2022-03-14.test.txt
test set scores: opusTCv20210807+pft_transformer-big_2022-03-14.eval.txt

testset	BLEU	chr-F	#sent	#words	BP
Tatoeba-test-v2021-08-07.bel-dan	29.1	0.50560	2	14	0.846
Tatoeba-test-v2021-08-07.bel-nor	63.1	0.72846	9	85	1.000
Tatoeba-test-v2021-08-07.multi-multi	52.6	0.69357	5108	37401	0.970
Tatoeba-test-v2021-08-07.rus-dan	59.3	0.74124	1713	11746	0.981
Tatoeba-test-v2021-08-07.rus-isl	7.0	0.28775	139	969	0.868
Tatoeba-test-v2021-08-07.rus-nno	72.8	0.80844	2	14	1.000
Tatoeba-test-v2021-08-07.rus-nob	45.9	0.66229	1277	11669	0.959
Tatoeba-test-v2021-08-07.rus-nor	45.9	0.66245	1279	11683	0.959
Tatoeba-test-v2021-08-07.rus-swe	52.6	0.69190	1282	8447	0.947
Tatoeba-test-v2021-08-07.ukr-dan	86.7	0.92688	10	54	1.000
Tatoeba-test-v2021-08-07.ukr-nor	59.3	0.73907	670	4378	1.000
Tatoeba-test-v2021-08-07.ukr-swe	9.5	0.46055	4	20	1.000

dataset: opusTCv20210807+xb+bt+pft
model: transformer-big
source language(s): bel bel_Latn rus ukr
target language(s): dan isl nno nob swe
raw source language(s): bel rus ukr
raw target language(s): dan isl nno nob swe
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+xb+bt+pft_transformer-big_2022-09-15.zip
test set translations: opusTCv20210807+xb+bt+pft_transformer-big_2022-09-15.test.txt
test set scores: opusTCv20210807+xb+bt+pft_transformer-big_2022-09-15.eval.txt

testset	BLEU	chr-F	#sent	#words	BP
flores101.ukr-dan	31.0	0.58310	1012	24643	1.000
flores101.ukr-swe	29.5	0.58309	1012	23135	0.995
Tatoeba-test-v2021-08-07.multi-multi	52.6	0.69303	5108	37401	0.967