Fix: Use attn_implementation='eager' for MPS compatibility #78
Conversation
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Outdated suggestions cannot be applied.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.
Description:
When running the colqwen2 model on an MPS device (Apple Silicon), the default attention implementation causes compatibility issues.
The fix ensures that attn_implementation="eager" is used when the device is MPS, allowing stable execution.
Steps to Reproduce:
1. Run the model on a MacBook Pro with an MPS device.
# Initialize RAGMultiModalModel model = RAGMultiModalModel.from_pretrained( "vidore/colqwen2-v0.1", device="mps",)2. Observe that the attention implementation may cause an error
Error
IndexError: Dimension out of range (expected to be in range of [-3, 2], but got 3)Proposed Fix:
Update the from_pretrained method to set attn_implementation="eager" when running on MPS.
attn_implementation = "eager" if device == "mps" or ( isinstance(device, torch.device) and device.type == "mps" ) else NoneFile: byaldi/colpali.py
Location: Inside the from_pretrained method.
Rationale:
• MPS devices currently face issues with the default attention implementation.
• "eager" mode provides a stable alternative.