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

FID score for Mnist #12

Closed
zzqiuzz opened this issue May 13, 2019 · 2 comments
Closed

FID score for Mnist #12

zzqiuzz opened this issue May 13, 2019 · 2 comments

Comments

@zzqiuzz
Copy link

zzqiuzz commented May 13, 2019

Hi! Thanks for your work.May i have you instruction on how to calculate FID score for Mnist dataset which has only 1 channels ,no idea to solve it now.Thanks.

@mseitzer
Copy link
Owner

Generally, if inputting grayscale images into RGB networks, you would just copy the grayscale channel three times. However, it is unclear how FID behaves in this case, as the original Inception network was probably not trained on grayscale images.

Also, I would advise against computing a FID score on Mnist. FID score was designed to work on larger images of natural scenes, not on 28x28 images of handwritten digits. I am guessing that the numbers you would get from that are pretty meaningless and do not reflect human judgement of quality.

@saurabheights
Copy link

I added more info here :- https://stackoverflow.com/questions/57183647/
FYI: Thanks for your repo. Helped me finish this task fast.

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

3 participants