Distinguish between dist and DDP
#201
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
jbloxham
requested review from
ravi-mosaicml,
Averylamp,
ajaysaini725 and
anisehsani
ravi-mosaicml
approved these changes
Jan 5, 2022
added 2 commits
January 6, 2022 19:47
coryMosaicML
pushed a commit
to coryMosaicML/composer
that referenced
this pull request
Feb 23, 2022
* dist, not ddp * simplify ClosureGradScaler * formatting * formatting and more fixes * that did not save * small fixes * dont need to worry about circular dependencies any longer * dumb pyright fix * woops
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Outdated suggestions cannot be applied.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.
The distributed runtime and the DDP engine are distinct entities, but our code has been treating them almost as synonyms. This is already causing some confusion in parts of the DeepSpeed integration, and it will only get worse if we experiment with other parallelism techniques like model and pipeline parallelism. The purpose of this PR is to separate out DDP-specific code from anything that just deals with the distributed runtime in general.
This is purely a refactor of something that was making me unhappy.