zzz66

/

AnyTalker-1.3B

like 12

This PR significantly enhances the model card for AnyTalker by:

• Adding the pipeline_tag: audio-to-video to the metadata, ensuring the model is discoverable under this category on the Hugging Face Hub.
• Adding explicit badges and links for the official Hugging Face paper: AnyTalker: Scaling Multi-Person Talking Video Generation with Interactivity Refinement.
• Adding an explicit badge and link to the GitHub repository: https://github.com/HKUST-C4G/AnyTalker.
• Integrating comprehensive content from the project's GitHub README, including:
  • The concise "TL; DR" summary.
  • Visual demonstrations with corrected raw GitHub URLs for images and videos.
  • "Latest News" and "Todo List" for project updates.
  • A detailed "Quick Start" guide with installation instructions, model preparation, and a full "Quick Inference" code snippet as sample usage.
  • Hyper-parameter descriptions and an illustration of the "pad" mode.
  • A "Benchmark" section.
  • BibTeX citation information.
  • The detailed Apache 2.0 license text.
• Removing the irrelevant "File information" section from the original model card.

These updates make the model card much more informative and user-friendly, providing essential details and runnable code for anyone interested in using AnyTalker.