Support special mapping of dtypes when preparing device map#1179
Merged
Support special mapping of dtypes when preparing device map#1179
Conversation
|
The documentation is not available anymore as the PR was closed or merged. |
younesbelkada
approved these changes
Mar 12, 2023
Contributor
younesbelkada
left a comment
There was a problem hiding this comment.
Thanks a mile for this!
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 PR will go along with [PR to come in Transformers] that fixes big model inference for T5 models. T5 models now have some weights that are kept in float32, which interferes with the computation done by
infer_auto_device_map. On my setup, it gets me OOM when trying to load any big T5 model.To fix that, on the Accelerate side we now accept a new argument
special_dtypeswhich maps weight names to dtypes to use (and for the others we use thedtypeprovided).