Fix device_map computation part 2 #42290
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
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
SunMarc
approved these changes
Nov 19, 2025
Member
SunMarc
left a comment
SunMarc
left a comment
There was a problem hiding this comment.
Indeed, this was indeed strange that in this version of compute_module_sizes it returned two values. Thanks for fixing this nasty bug !
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Outdated suggestions cannot be applied.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.
What does this PR do?
Follow up to #42264.
Since #42043,
compute_module_sizesis also used in_init_infer_auto_device_map. However, there it expects to have all the modules inside, but also the leaves parameters and buffers, which was not the case. Because we returndefaultdictinstance, we would not have anyKeyErrorbut instead very nasty silent bugs because the values found would always be 0 (the default of the defaultdicts.This PR fixes it by adding the option for adding them in the dict. Doing
compute_module_sizes(model, only_modules=False)is now equivalent toaccelerate.compute_module_sizes(model), which is not the case without the flag (leaves are not in the dict)cc @SunMarc