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.

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

BUG: Save model parameters without DataParallel module wrapper #17

Merged
merged 1 commit into from
Aug 8, 2023

Conversation

carterbox
Copy link
Collaborator

Avoid mismatched model parameters when training and inference are conducted on different number of GPUs. The mismatched parameters because the model is wrapped in a DataParallel module when multiple GPUs are used which moves the parameters down one level.

@carterbox carterbox changed the title REF: Save model parameters without DataParallel module wrapper BUG: Save model parameters without DataParallel module wrapper Aug 7, 2023
Copy link
Collaborator

@stevehenke stevehenke left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code changes make sense. Training and inference operations both work with these changes. Thank you and nice job!

@carterbox carterbox merged commit 3834954 into mcherukara:package Aug 8, 2023
@carterbox carterbox deleted the consistent-dataparallel branch August 8, 2023 15:02
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants