markendo

/

llava-extract-qwen3-1.7B

like 0

This PR significantly enhances the model card for markendo/llava-extract-qwen3-1.7B by providing crucial metadata and comprehensive usage instructions.

Key improvements include:

• Adding pipeline_tag: image-text-to-text to accurately reflect its multimodal capabilities (image + text input, text output) and improve discoverability.
• Specifying library_name: transformers based on config.json and usage patterns, enabling the automated "how to use" widget on the Hub.
• Including descriptive tags (e.g., multimodal, vision-language-model, reasoning, small-language-model) and base_model information for better context and searchability.
• Linking directly to the associated paper, project page, and GitHub repository for easy access to more information and code.
• Adding a comprehensive "Usage" section with setup instructions and code snippets directly from the GitHub README, demonstrating how to use the Extract+Think framework for visual extraction and reasoning.

These additions will make the model more discoverable and user-friendly on the Hugging Face Hub.