2U35

/

MIRG-RL-7B

like 0

This PR enhances the model card for the MIRG-RL model by:

• Adding library_name: transformers to the metadata, enabling the automated "How to use" widget on the Hugging Face Hub, as indicated by transformers_version and Qwen2VLForConditionalGeneration in the config files.
• Adding pipeline_tag: image-text-to-text to the metadata, making the model discoverable for multi-modal tasks involving image input and text output, consistent with "Multi-Image Reasoning and Grounding with Reinforcement Learning".
• Including a direct link to the paper MIRG-RL: Multi-Image Reasoning and Grounding with Reinforcement Learning.
• Providing a link to the official GitHub repository for the project (https://github.com/ZEUS2035/MIRG-RL).
• Integrating key information from the GitHub README regarding datasets, model details, training, and evaluation to provide a more comprehensive overview of the model.
• Removing the internal "File information" section from the model card.