Set parallelism_config in constructor due to Trainer reset of State #3713
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This is a fix for how Trainer by default resets the PartialState shared state, which means we can't read the ParallelismConfig from PartialState anymore. Since we can infer the ParallelismConfig from the env vars like we do all the other
ACCELERATE_*options (e.g. FSDP, Deepspeed,etc), we can repopulate this inside the constructor.Addtionally, there is a fix to the indent level of setting the fsdp_plugin, as if it the conditional it currently is in would bypass setting self.fsdp_plugin if the env var was set to allow standalone CP, which is independent of FSDP.
We'll need to add some fixes down the line in Trainer, but this will get us to a working release.
Fixes # (issue)
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
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.