Fix FSDP error #1196
Merged
Fix FSDP error #1196
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
Fixes error when `loss` field of model output is non-empty, and indexing as [0] returns loss instead of logits. Can happen with FSDP.
|
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. |
mgerstgrasser
force-pushed
the
fix_reward_trainer_FSDP_error
branch
from
c9ed225 to
989f4a4
Compare
force return_dict Co-authored-by: Younes Belkada <[email protected]>
Yes, that works! Thanks for spotting that. |
younesbelkada
approved these changes
Jan 9, 2024
There was a problem hiding this comment.
Thanks again for all your contributions @mgerstgrasser ! You are on fire!
lapp0
pushed a commit
to lapp0/trl
that referenced
this pull request
May 10, 2024
* Fix FSDP error Fixes error when `loss` field of model output is non-empty, and indexing as [0] returns loss instead of logits. Can happen with FSDP. * Apply suggestions from code review force return_dict Co-authored-by: Younes Belkada <[email protected]> --------- Co-authored-by: Younes Belkada <[email protected]>
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Outdated suggestions cannot be applied.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 fixes an error in reward training when the
lossfield of model output is non-empty, andmodel(...)[0]returns loss instead of logits. That can happen with FSDP. This PR changes it so the trainer explicitly grabsmodel(...)["logits"]instead.Closes #1195