Vision fix for non-clip models

[michaelfeil]

Closes #437

[greptile-apps]

PR Summary

This PR enhances the TIMM class in torch_vision.py for improved compatibility and performance with various vision model types, accompanied by corresponding updates to the test suite.

• Modified encode_core method in TIMM class to handle different processing paths based on is_colipali flag
• Updated tokenize_lengths method to ensure correct data types for input tensors
• Introduced parameterization in test_clip_like_model function to test multiple model types
• Replaced specific CLIP imports with generic AutoModel and AutoProcessor in test file
• Updated main block in test file to use new image sample URL and test both CLIP-like and ColPali models

_{2 file(s) reviewed, 2 comment(s)
Edit PR Review Bot Settings | Greptile}

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 60.00000% with 2 lines in your changes missing coverage. Please review.

Project coverage is 79.18%. Comparing base (02480ab) to head (6118a77).

Files with missing lines	Patch %	Lines
...mb/infinity_emb/transformer/vision/torch_vision.py	60.00%	2 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #439      +/-   ##
==========================================
- Coverage   79.22%   79.18%   -0.05%     
==========================================
  Files          41       41              
  Lines        3245     3248       +3     
==========================================
+ Hits         2571     2572       +1     
- Misses        674      676       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.