Use fake_mode when constructing ShardingOptimizer #70
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
This is particularly important for constructor nodes and for the flop estimation, otherwise they could materialize massive tensors in memory, leading to OOM. This shows up in DeepSeek.
meta-cla
bot
added
the
CLA Signed
wconstab
approved these changes
Aug 4, 2025
|
|
||
| from torch._subclasses.fake_tensor import unset_fake_temporarily | ||
|
|
||
| with unset_fake_temporarily(): |
Contributor
There was a problem hiding this comment.
Do you happen to know why this was here originally? Seems like it would have been logical to use fake mode all along. Maybe it was just to work around the one case with reshape below?
Contributor
Author
There was a problem hiding this comment.
Yeah, it was there because in one of the shardings we were calling copy.deepcopy , which is not allowed under fake_mode.
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Outdated suggestions cannot be applied.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 is particularly important for constructor nodes and for the flop estimation, otherwise they could materialize massive tensors in memory, leading to OOM. This shows up in DeepSeek, and I'm splitting this up from #29