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 was last DiT block of Mochi Skipped in Sageattnetion Example? #96

Open
asahni04 opened this issue Feb 1, 2025 · 1 comment
Open

why was last DiT block of Mochi Skipped in Sageattnetion Example? #96

asahni04 opened this issue Feb 1, 2025 · 1 comment

Comments

@asahni04
Copy link

asahni04 commented Feb 1, 2025

why was last DiT block of Mochi Skipped in Sageattnetion Example? do you suggest to skip last blocks for better accuracy?
@jason-huang03
Copy link
Member

We empirically find that the last layer in mochi is quite sensitive so we leave it in full precision. We do not know whether this holds true for general DiT.

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
@jason-huang03 @asahni04