jusjinuk

/

Llama-3.1-70B-Instruct-4bit-GuidedQuant-QTIP

like 0

This PR significantly improves the model card for the GuidedQuant quantized Llama-3.1-70B-Instruct model by:

• Adding Metadata: Including pipeline_tag: text-generation and library_name: transformers to enhance discoverability on the Hub and enable the usage widget.
• Updating Paper Link: Replacing the generic arXiv link with the direct Hugging Face Papers page link (https://huggingface.co/papers/2505.07004).
• Adding Explicit Links: Providing prominent links to the project page (https://jusjinuk.me/blog/guidedquant/) and the official GitHub repository (https://github.com/snu-mllab/GuidedQuant).
• Including Abstract: Adding the paper's abstract to provide a quick and comprehensive overview of the research.
• Providing Sample Usage: Integrating a clear and runnable Python code snippet for quick-start inference with the model, adapted to use the correct model identifier for this repository.
• Improving Structure: Reorganizing model details and consolidating references into a dedicated "Citation" section.

These changes make the model card more informative, accessible, and user-friendly for researchers and practitioners on the Hugging Face Hub.