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

What is the FID-30K test subset? #131

Open
songtianhui opened this issue Jul 5, 2023 · 2 comments
Open

What is the FID-30K test subset? #131

songtianhui opened this issue Jul 5, 2023 · 2 comments

Comments

@songtianhui
Copy link

Hi, thanks for your awesome work!
I want to know what is the FID-30K from MSCOCO val set, since as the paper says that the 30k samples are drawn randomly. Are the evaluation result is aligned with the original Imagen paper?

@Mywifi
Copy link

Mywifi commented Jul 31, 2023

Hi, thanks for your awesome work! I want to know what is the FID-30K from MSCOCO val set, since as the paper says that the 30k samples are drawn randomly. Are the evaluation result is aligned with the original Imagen paper?

COCO FID-30K is a metric used to evaluate the quality of face generative models. It stands for:

COCO Face Identification at 30K

It measures how different 30,000 generated faces from a generative model are from 30,000 real human faces.
https://cocodataset.org/#home

@songtianhui
Copy link
Author

As far as I know, COCO FID-30k is MSCOCO val2014. I think that current diffusion models are not designed and tested fro face identification task?

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