Fix EtaLogitsWarper on fully masked logits#45413
Open
ezylopx5 wants to merge 1 commit intohuggingface:mainfrom
Open
Fix EtaLogitsWarper on fully masked logits#45413ezylopx5 wants to merge 1 commit intohuggingface:mainfrom
ezylopx5 wants to merge 1 commit intohuggingface:mainfrom
Conversation
Signed-off-by: HarshRathva <harshrathvaai@gmail.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
I ran into an edge case in eta sampling where
EtaLogitsWarpercrashes if a row is fully masked (scores == -inffor all tokens).The previous entropy computation used
Categorical(logits=scores).entropy(), which fails on that input. This change computes entropy from NaN-safe probabilities instead, so fully masked rows stay fully masked without raising.I also added a regression assertion to the existing eta warper test for the all-
-infcase.Local test run:
python -m pytest tests/generation/test_logits_process.py -x -q(45 passed).AI-assisted: I used AI help for drafting, but I personally reproduced the bug, reviewed the diff, and ran local tests.