MMR1

/

MMR1-3B-SFT

like 0

Follow

MMR1 13

This PR significantly improves the model card for MMR1 by:

• Adding pipeline_tag: image-text-to-text to enhance discoverability for multimodal reasoning tasks.
• Adding library_name: transformers based on the model's configuration (Qwen2_5_VLForConditionalGeneration, Qwen2_5_VLProcessor, transformers_version) to enable the automated usage widget.
• Adding license: apache-2.0 as explicitly stated in the GitHub repository.
• Including the GitHub repository link for easy access to the code.
• Providing a comprehensive model description derived from the paper abstract and GitHub README, structured for clarity and incorporating key visuals.
• Retaining existing arXiv and Hugging Face paper badges.

Please review and merge if these improvements align with our goals.