NOTICE: The program have finished about 80%. I am interning in Hitachi now, and I am studying on the zero-shot learning recently, I will do my best to complete the program when I finish the project of zero-shot learning.
TensorFlow： 0.12.0
Python： 2.7
本项目实现论文Deep Reinforcement Learning for Dialogue Generation，论文地址：https://arxiv.org/abs/1606.01541
the start file ：
grl_train.py
run step :
reference funciton main() in file grl_train.py
TODO
implement seq2seq model with mutual information based on beam search