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

Variable and Optimizer aggregation='mean' default is incorrect #20695

Open
rivershah opened this issue Dec 27, 2024 · 0 comments
Open

Variable and Optimizer aggregation='mean' default is incorrect #20695

rivershah opened this issue Dec 27, 2024 · 0 comments
Assignees
@sachinprasadhs

Comments

@rivershah
Copy link

The default policy of using aggregation='mean' is incorrect. This should be set to none. Backends do the grad reductions and var applications in distributed contexts and don't need this mean. The usage of aggregation='mean' breaks optimizer moment estimates. Also compare with keras==2.15.0 where default policy was equivalent to aggregation='none'
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@sachinprasadhs sachinprasadhs

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@rivershah @sachinprasadhs