Allow scalar all gather #5797
Merged
Allow scalar all gather #5797
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
Nov 15, 2023
JackCaoG
approved these changes
Nov 15, 2023
|
Ping on this PR, the failing tests seem unrelated to this change. |
|
sorry my bad.. should of merge it earlier. Do you mind resolve the merge conflict and rebase? Do you need this to backport to 2.2 release? |
|
Resolved the conflict and no need to backport. |
|
sg, I will merge it once all ci passed. |
|
Thanks! |
chunnienc
pushed a commit
to chunnienc/xla
that referenced
this pull request
Dec 14, 2023
golechwierowicz
pushed a commit
that referenced
this pull request
Jan 12, 2024
bhavya01
pushed a commit
that referenced
this pull request
Apr 22, 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Outdated suggestions cannot be applied.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.
This change allows all gather to work for scalars. This is supported by the normal pytorch distributed API, but if you attempt to all gather scalars under xla, a low level error is emitted as the xla gather axis, 0, is not less than the rank, also 0, of the inputs.
This is fixed by reshaping the scalar tensors to shape (1,) and then reshaping the sliced outputs back to scalars.
This was tested manually on Cloud TPU.