fix: VLMModelWrapper PEFT isinstance compatibility for TRL validation#252
Merged
Conversation
TRL's validate_quantization_for_training() uses isinstance(model, PeftModel) to check for adapters. The wrapper hid the PeftModel, causing: "You cannot perform fine-tuning on purely quantized models." Fix: dynamically create a combined class inheriting from both VLMModelWrapper and the wrapped model's type. This makes isinstance() pass while keeping our forward/generate/cache methods via MRO. Tests added: - test_peft_attributes_delegated: peft_config accessible through wrapper - test_hasattr_peft_config: hasattr() works for TRL's checks - test_isinstance_peft_model: isinstance(wrapper, PeftModel) == True - test_wrapper_passes_peft_validation (e2e): full TRL validation sim - test_wrapper_preserves_trainable_parameters (e2e): optimizer setup Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
TRL's
validate_quantization_for_training()usesisinstance(model, PeftModel)to check for adapters. The VLMModelWrapper hid the PeftModel, causing: "You cannot perform fine-tuning on purely quantized models."Fix: Dynamically create a combined class inheriting from both
VLMModelWrapperand the wrapped model's type.isinstance()passes, ourforward/generate/cachemethods take priority via MRO.Tests added
test_peft_attributes_delegated— peft_config accessibletest_hasattr_peft_config— hasattr() works for TRLtest_isinstance_peft_model— isinstance(wrapper, PeftModel) == Truetest_wrapper_passes_peft_validation(e2e) — simulates TRL validationtest_wrapper_preserves_trainable_parameters(e2e) — optimizer setup worksTest plan
🤖 Generated with Claude Code