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

[Training] Support for 2 fwd graphs #992

Open
vladimirjovanovicTT opened this issue Dec 31, 2024 · 0 comments
Open

[Training] Support for 2 fwd graphs #992

vladimirjovanovicTT opened this issue Dec 31, 2024 · 0 comments
Assignees
@vladimirjovanovicTT
Labels
enhancement New feature or request

Comments

@vladimirjovanovicTT
Copy link
Contributor

When training, we need 2 forward passes in case layers (like BatchNorm, Dropout and some lesser known layers) have different behavior dependent on whether the model is in training mode or not.
Support for 2 forward graphs (compilation, switching) needs to be added to FFE.

This issue will be updated with API details, to be discussed offline.
@vladimirjovanovicTT vladimirjovanovicTT added the enhancement New feature or request label Dec 31, 2024
@vladimirjovanovicTT vladimirjovanovicTT self-assigned this Dec 31, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@vladimirjovanovicTT vladimirjovanovicTT

Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@vladimirjovanovicTT