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

[FSDP] Allow training to work with FSDP in ZeRO stage-3 #201

Open
RobotSail opened this issue Sep 12, 2024 · 0 comments
Open

[FSDP] Allow training to work with FSDP in ZeRO stage-3 #201

RobotSail opened this issue Sep 12, 2024 · 0 comments
Assignees
@JamesKunstle

Comments

@RobotSail
Copy link
Member

We currently have the ap/fsdp-port branch which houses an implementation of training with FSDP, but it currently isn't using stage-3 parameter offloading. We need to update the implementation to do parameter offloading and successfully save checkpoints.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@JamesKunstle JamesKunstle

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@JamesKunstle @RobotSail