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.

Sign up for GitHub

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

Jump to bottom

Why scaling it by 2.74? #3

Open
AmperiaWang opened this issue Aug 5, 2024 · 1 comment
Open

Why scaling it by 2.74? #3

AmperiaWang opened this issue Aug 5, 2024 · 1 comment

Comments

@AmperiaWang
Copy link

I noticed that you have used large amount of Scale(2.74) in your code. However, I didn't saw any descriptions or similar codes by others that can explain it. I wonder if they are some kind of feature or just I missed the related description. Can you explain it for me? I'll appreciate it.
@pkuxmq
Copy link
Owner

pkuxmq commented Aug 6, 2024

This is the scale for scaled weight standardization in NF-ResNet (instead of batch normalization). Please refer to Sec. 4.4 and Appendix C.1 in the paper for details.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@pkuxmq @AmperiaWang