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

RGB Image or Gray Scale Image #40

Closed
FeiMiBa opened this issue Aug 20, 2020 · 5 comments
Closed

RGB Image or Gray Scale Image #40

FeiMiBa opened this issue Aug 20, 2020 · 5 comments

Comments

@FeiMiBa
Copy link

FeiMiBa commented Aug 20, 2020

Thanks for your great work!

Did you train/test all these on RGB images? I have a task to train/test on gray scale images, do you think i will suffer from performance degradation in case I train from scratch with all your default settings? Any suggestions? Thank you!

@cavalleria
Copy link
Owner

Yes, all images are rgb. I think the performance will degrade if you train with gray image.

@xsacha
Copy link
Contributor

xsacha commented Aug 20, 2020

Probably change the input channels to 1 so you at least get a performance improvement in using grayscale, if that is your requirement.

@FeiMiBa
Copy link
Author

FeiMiBa commented Aug 21, 2020

Yes, all images are rgb. I think the performance will degrade if you train with gray image.

Have you cleaned the ms1m-retinaface dataset, or you just used it anyway?

@cavalleria
Copy link
Owner

Yes, all images are rgb. I think the performance will degrade if you train with gray image.

Have you cleaned the ms1m-retinaface dataset, or you just used it anyway?

yes. slightly cleaned

@FeiMiBa
Copy link
Author

FeiMiBa commented Aug 24, 2020

Yes, all images are rgb. I think the performance will degrade if you train with gray image.

Have you cleaned the ms1m-retinaface dataset, or you just used it anyway?

yes. slightly cleaned

Awesome, thank you!

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