fix(reward_classifier): remove leftover normalization calls from predict_reward#2355
Open
TheWisp wants to merge 2 commits intohuggingface:mainfrom
Open
fix(reward_classifier): remove leftover normalization calls from predict_reward#2355TheWisp wants to merge 2 commits intohuggingface:mainfrom
TheWisp wants to merge 2 commits intohuggingface:mainfrom
Conversation
…ict_reward Remove normalize_inputs and normalize_targets calls from the predict_reward method that were accidentally left behind during the processor pipeline refactoring in PR huggingface#1452. This completes the migration where normalization was moved from internal policy methods to external processor pipelines. The __init__ and forward methods were updated in commit 78b8661, but predict_reward was missed. Fixes AttributeError: 'Classifier' object has no attribute 'normalize_inputs'
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Removes
normalize_inputsandnormalize_targetscalls from thepredict_rewardmethod that were accidentally left behind during the processor pipeline refactoring in PR #1452.Details
__init__andforwardmethods were properly updated in commit 78b8661 to remove normalizationpredict_rewardmethod was missed, leaving calls to non-existent methodsFixes
AttributeError: 'Classifier' object has no attribute 'normalize_inputs'
Testing
Verified by checking git history that all other policies (ACT, Diffusion, SAC) had their normalization removed in the same way during PR #1452.
Related
predict_reward