Correctness fix PP+ZeRO for gradient accumulation #1264
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jeffra
requested review from
RezaYazdaniAminabadi,
ShadenSmith,
awan-10,
cli99,
conglongli,
eltonzheng,
minjiaz,
niumanar,
samyam and
tjruwase
as code owners
jeffra
changed the title
samyam
approved these changes
Jul 29, 2021
tjruwase
reviewed
Jul 30, 2021
|
|
||
| def allreduce_gradients(self, bucket_size=MEMORY_OPT_ALLREDUCE_SIZE): | ||
| # Pass (PP) gas boundary flag to optimizer (required for zero) | ||
| self.optimizer.is_gradient_accumulation_boundary = self.is_gradient_accumulation_boundary( |
Contributor
There was a problem hiding this comment.
Could self.optimizer be an fp32 optimizer, our fp16 optimizer, or client optimizer? Is it safe to assume self.optimizer.is_gradient_accumulation_boundary?
|
This PR fixes this issue: #1217 |
Collaborator
|
@jeffra, could you please replay to @big-science? Thank you! |
Collaborator
Author
|
I believe this is already replayed on big-science branch: f93e22b I think this commit was mixed with a few other items as well though. |
Collaborator
|
Indeed. Thank you for finding it, Jeff! I was just comparing the titles of the commits. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
No description provided.