sigmoid / WScaleLayer - Githubissues

zhangqianhui / progressive_growing_of_gans_tensorflow

Tensorflow implementation of PROGRESSIVE GROWING OF GANS FOR IMPROVED QUALITY, STABILITY, AND VARIATION

MIT License

149 stars 71 forks source link

Closed elbaro closed 6 years ago

elbaro commented 6 years ago

Hi, I am glad to find the implementation using wgan-gp.

Can you explain the differences from the reference implementation:

zhangqianhui commented 6 years ago

I don't use the sigmoid for the loss function. You can check it.
I implement WScalelayer for learning rate equalization , but get the bad performance. If you implement it with good performance, please make a PR. Thank you.

elbaro commented 6 years ago

I implement WScalelayer for learning rate equalization , but get the bad performance.

That's good to know. Thanks!