Are there are some open machine translation data include candidate and reference data for calculating BLEU scores. 100 rows data of candidate and reference data are ok for me to implement BLEU in python. Thanks!
Asked
Active
Viewed 60 times
-2
-
Try downloading this: https://github.com/nltk/nltk_data/blob/gh-pages/packages/models/wmt15_eval.zip Unzip and there's a readme inside. – alvas Oct 17 '17 at 07:34
1 Answers
1
You can take any parallel text corpora. Plenty of open datasets are available in this regard. Have a look at this: http://www.statmt.org/europarl/

Unni
- 5,348
- 6
- 36
- 55
-
thanks, but I also want to the candidate data , not just reference data. – tktktk0711 Sep 25 '17 at 07:47