andyc03

/

Qwen2-VL-PRISM-DPO

like 1

This PR adds a comprehensive model card for the PRISM model, ensuring it is properly documented on the Hugging Face Hub.

Specifically, this PR:

• Links the model to its paper: PRISM: Robust VLM Alignment with Principled Reasoning for Integrated Safety in Multimodality.
• Adds the license (MIT), library_name (transformers), and pipeline_tag (image-text-to-text) to the metadata, making the model discoverable on the Hub via relevant filters and enabling the automated 'How to use' widget.
• Includes the paper's abstract for a quick overview of the model's capabilities and purpose.
• Provides a direct link to the GitHub repository for easy access to the code and further details.

Please review and merge this PR to enhance the visibility and information quality of the PRISM model on the Hugging Face Hub.