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

Feature request: Use Orbax for checkpointing. #80

Open
OhadRubin opened this issue Jul 25, 2023 · 2 comments
Open

Feature request: Use Orbax for checkpointing. #80

OhadRubin opened this issue Jul 25, 2023 · 2 comments

Comments

@OhadRubin
Copy link

wdyt about changing the current checkpointing to use orbax?
https://github.com/google/orbax

@young-geng
Copy link
Owner

I think the current streaming checkpointer is quite good and I'm not planning to switch to orbax. Specifically, the streaming checkpointer can save and load checkpoints with minimal memory and temporary disk usage, which means that you can checkpoint models that does not fit in the memory or local disk of a single machine. Also as a personally preference, I want to stay away from complicated Google libraries as much as possible.

@erfanzar
Copy link

obrax do not support load streaming and sharding data or array across devices with pjit so I think the current checkpointing method that is being used right now is a smart move :)

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

No branches or pull requests

3 participants