nlp_xiaojiang/AugmentText/augment_seq2seq/code_seq2seq_char
2019-04-21 22:00:00 +08:00
..
__init__.py augment of seq2seq with webank datas 2019-04-17 20:07:33 +08:00
extract_char_webank.py augment of seq2seq with webank datas 2019-04-17 20:07:33 +08:00
predict_char_anti.py augment of seq2seq with webank datas 2019-04-17 20:07:33 +08:00
train_char_anti.py 将log(p(T))改为log(p(S/T)),就是让问句和同义句尽可能相似吧,不过收敛困难,效果也比较一般,哭 2019-04-21 22:00:00 +08:00