The split of train and test.

Hello, authors:

Great work. Could you upload the train and test split text as your paper showed. Cause in this github, I'm not sure how to split the train and test.
Further, if possible, could you give me more helpful scripts to easy process/use the dataset? 

Thanks.
