Skip to content

[FSDP] Make env var and dataclass flag consistent for cpu_ram_efficient_loading #3307

Merged
SunMarc merged 1 commit intohuggingface:mainfrom
SumanthRH:fsdp-ram-efficient-loading
May 12, 2025
Merged

[FSDP] Make env var and dataclass flag consistent for cpu_ram_efficient_loading #3307
SunMarc merged 1 commit intohuggingface:mainfrom
SumanthRH:fsdp-ram-efficient-loading

Conversation

@SumanthRH
Copy link
Contributor

@SumanthRH SumanthRH commented Dec 19, 2024

What does this PR do?

Makes environment variable and dataclass consistent for cpu_ram_efficient_loading. Currently, there's an inconsistency when the FSDP plugin is initialized in a standalone fashion: the cpu_ram_efficient_loading flag can be set to True but if the user doesn't use accelerate launch , then the environment variable FSDP_CPU_RAM_EFFICIENT_LOADING would not be set. This environment variable is actually used by transformers to initialize parameters efficiently in the process group (See: #2100 (comment) for details on how this happens)

I can see a previous PR #2998 that tried to make all env vars optional, but looks like this was left.

Fixes # (issue)

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

cc @muellerzr

Signed-off-by: SumanthRH <sumanthrh@anyscale.com>
@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@github-actions
Copy link
Contributor

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

@SumanthRH
Copy link
Contributor Author

not stale

@github-actions
Copy link
Contributor

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

@SumanthRH
Copy link
Contributor Author

not stale.

@muellerzr would appreciate a review on this!

@github-actions
Copy link
Contributor

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

@SumanthRH
Copy link
Contributor Author

not stale

@SunMarc
Copy link
Member

SunMarc commented Apr 8, 2025

cc @S1ro1 if you can take a look as you were working on that for fsdpv2 also

@github-actions
Copy link
Contributor

github-actions bot commented May 2, 2025

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

@S1ro1
Copy link
Contributor

S1ro1 commented May 2, 2025

not stale, gonna take a look!

@SunMarc SunMarc requested a review from S1ro1 May 6, 2025 15:51
Copy link
Member

@SunMarc SunMarc left a comment

Choose a reason for hiding this comment

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

Thx !

@SunMarc SunMarc merged commit f76208f into huggingface:main May 12, 2025
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.

4 participants