Improve model card: Add metadata, paper/code links, abstract, and usage examples
#2
by
nielsr
HF Staff
- opened
This PR significantly enhances the model card for Llama3.2-Mamba2-3B-dpo by:
- Adding the
pipeline_tag: text-generation, making the model discoverable on the Hugging Face Hub (e.g., at https://huggingface.co/models?pipeline_tag=text-generation). - Specifying
library_name: transformers, which correctly indicates its compatibility with the Transformers library'sAutoTokenizerand its underlying architecture. - Including a direct link to the paper: The Mamba in the Llama: Distilling and Accelerating Hybrid Models.
- Adding a link to the official GitHub repository: https://github.com/jxiw/MambaInLlama.
- Integrating the paper's abstract for quick context.
- Providing a clear Python usage example for direct inference with this specific model.
- Updating the performance evaluation section with more comprehensive tables and visuals from the original GitHub repository.
These additions will greatly improve the model's visibility, usability, and documentation for the community.