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

How much images should be used to compute the is? #15

Open
destructive-observer opened this issue Sep 9, 2021 · 1 comment
Open

How much images should be used to compute the is? #15

destructive-observer opened this issue Sep 9, 2021 · 1 comment

Comments

@destructive-observer
Copy link

Dear sbarratt:
Thanks for your job,i have a problem with the is. With cifar for example , if i used the total cifa dataset to compute the is score ,i will get the 9.5 score.But if i used a few batch images generate by GAN e.g (1000 images), i only get 2.5 score.So i am confused by how many images i should used to compute the IS score as to judge the quality of the generator images.Thanks for your help.

@BridgeQZH
Copy link

@destructive-observer Did you find out the answer by yourself? I also met this question.

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

2 participants