add colqwen2_5 #27
Merged
add colqwen2_5 #27
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
Blaizzy
reviewed
May 27, 2025
Blaizzy
requested changes
May 27, 2025
- Removed the MODEL_REMAPPING dictionary as it was no longer needed. - Updated the quantize_model function to accept skip_vision as a parameter. - Refactored ColQwen2_5Config to ModelArgs for better compatibility with utils.py. - Enhanced from_dict method to cleanly extract and convert text and vision configurations.
Contributor
Author
|
All done, everything should be fixed now @Blaizzy :) |
Owner
|
Thanks! Are you able to load the quantized models? If not it's because the predicate for quants has to be applied in the load_model function as well. |
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Outdated suggestions cannot be applied.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.
Hello @Blaizzy , please check out this PR for ColQwen2.5, which brings multimodal document retrieval to MLX Embeddings :)
However, I would recommend to wait for this PR to get merged first by HuggingFace to transformers so that we have a proper support for the AutoProcessor :)
Please let me know if there should be any additional. Looking forward to get this merge :D