liuyuemaicha / Adversarial-Learning-for-Neural-Dialogue-Generation-in-Tensorflow

Adversarial-Learning-for-Neural-Dialogue-Generation-in-Tensorflow
http://blog.csdn.net/liuyuemaicha/article/details/60581187
187 stars 75 forks source link

all_train里的decode部分 #9

Open JiaheChang opened 7 years ago

JiaheChang commented 7 years ago

Hi…打扰了…我又来问问题了…… 我在跑all_train 这部分的时候, if mc_search: train_query, train_answer, train_labels = decoder(gen_config.beam_size) 不是说对每个输入要输出beam_size个输出,然后来判断reward吗……………… 但是对于一样的输入来说……我的输出都是一样的……= =所以虽然它decode了beam_size次,但是完全没起到作用啊…… 所以这是发生了什么……