HIT-TMG

/

Uni-MoE-2.0-Omni

like 36

Follow

Lychee Team 90

This PR improves the model card for Uni-MoE 2.0-Omni by adding essential metadata and enhancing external links:

• pipeline_tag: any-to-any: This tag is added to accurately reflect the model's omnimodal capabilities, including multimodal understanding, reasoning, and generation (images, text, speech). This will improve its discoverability on the Hugging Face Hub under the appropriate pipeline filter.
• library_name: transformers: Evidence from the config.json and the provided Example Usage code snippet indicates compatibility with the Hugging Face transformers library. Adding this will enable the automated "how to use" widget on the model page.
• Improved External Links:
  • A dedicated badge for the Hugging Face paper page (https://huggingface.co/papers/2511.12609) has been added alongside the existing arXiv link for direct access to the paper on the Hugging Face platform.
  • A clear badge for the GitHub repository (https://github.com/HITsz-TMG/Uni-MoE) has been added to provide quick access to the source code.

These additions will significantly improve the model card's completeness, usability, and discoverability for the community.