A simplified PyTorch implementation of "SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient." (Yu, Lantao, et al.)
642
stars
149
forks
source link
Why doesn't it work if I use Adam for the discriminator? #12
Closed
Mhzzzzz closed 4 years ago
It's pretty fragile from what I remember, and even small changes in the setup might destabilise training. Try tinkering with a few hyperparameters.