search

williamSYSU / TextGAN-PyTorch

TextGAN is a PyTorch framework for Generative Adversarial Networks (GANs) based text generation models.
MIT License
892 stars 205 forks source link

A problem in seqgan #14

Closed maz0318 closed 4 years ago

maz0318 commented 5 years ago

Hello,Thanks for your sharing your code. But In your implement of seqgan, I have a problems: That is computing reward(in uitls.rollout), why you use a mean Q-value(reward) to replace every Q-value? I think it should be every Q-value multiply log(P(y_t|Y1:Y{t-1})).

williamSYSU commented 5 years ago

Thanks for your reviews.

I recently update my codes according to the mistakes you mentioned, please check out the latest codes. In the latest codes, by the way, the torch.mean() function only execute in the last dimension, which represents the rollout number of the SeqGAN.

maz0318 commented 5 years ago

Ok!Thanks for your reply.I will check out it.