Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Help on Parameter tuning #4

Open
kickbox opened this issue Nov 20, 2017 · 1 comment
Open

Help on Parameter tuning #4

kickbox opened this issue Nov 20, 2017 · 1 comment

Comments

@kickbox
Copy link

kickbox commented Nov 20, 2017

Thanks for the awesome work! I used pix2pix in tensorflow and had some good success. I tried the same set of images with your model, however the initial training doesn't seem go as expected.

Below is the model measures;
capture

Since I have only CPU my training times are typically longer(>3 days on 20 cores).

I have set another run with learning rate of 0.00002 for both D & G. Is there any particular hyper-parameter that needs adjustment?

@kickbox
Copy link
Author

kickbox commented Nov 21, 2017

Update: All my runs with different values of gamma (.3,.5,.7) and lr(.0002,.00002) produce all black images and the measures look just like above...

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant