Modularize zero step function and make it customizable #7233
Conversation
This file contains 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
JackCaoG
reviewed
Jun 10, 2024
Comment on lines
331
to
337
| ¦ return [ | ||
| ¦ ¦ { | ||
| ¦ ¦ ¦ "scale_factor": 1.0, | ||
| ¦ ¦ ¦ "sharding_group": self.sharding_groups, | ||
| ¦ ¦ ¦ "group_size": self.local_world_size, | ||
| ¦ ¦ }, | ||
| ¦ ] |
There was a problem hiding this comment.
you seems to copy some formatter metadata in the pr.
jeffhataws
approved these changes
Jun 12, 2024
jeffhataws
pushed a commit
that referenced
this pull request
Oct 29, 2024
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Outdated suggestions cannot be applied.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.
Refactor and (slightly) generalize step function of the zero redundancy optimizer, by breaking it into the following three high-level steps:
_reduce_gradients_clip_grad_norm_update_parametersThis makes it easier to sub-class and override specific behaviors of the zero optimizer in different contexts.
In addition, we introduce an additional optional parameter
sharding_scheme, which allows us to customize steps 1 and 3 above, if needed.