-
Notifications
You must be signed in to change notification settings - Fork 2.2k
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
Support backward_passes_per_step > 1 for TF Keras Eager Execution #2371
Conversation
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
PR is currently failing: |
Hey @aaron276h, landed a fix for the Python 3.8n issue, please rebase and force push. |
Signed-off-by: aaron276h <aaron@determined.ai>
Signed-off-by: aaron276h <aaron@determined.ai>
Thanks @tgaddair! |
Signed-off-by: aaron276h <aaron@determined.ai>
Signed-off-by: aaron276h <aaron@determined.ai>
Unit Test Results 395 files 395 suites 2h 47m 29s ⏱️ results for commit 6a08ba9 |
This comment has been minimized.
This comment has been minimized.
Signed-off-by: aaron276h <aaron@determined.ai>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
Checklist before submitting
Description
This PR is a follow up #2346. This PR adds support for backward_passes_per_step > 1 for TF Keras optimizers executing in eager mode. This is one of the features that we have built into Determined AI's fork of Horovod that we would like to upstream.