AngelSlim

/

Qwen3-4B_eagle3

like 2

Follow

AngelSlim 37

This PR significantly enhances the model card by:

• Linking to the paper Tequila: Trapping-free Ternary Quantization for Large Language Models and the specific Tequila implementation on GitHub.
• Adding the pipeline_tag: text-generation for better discoverability.
• Specifying library_name: transformers to enable the automated inference widget and reflect the model's compatibility.
• Updating the license to apache-2.0 as per the project's GitHub repository.
• Adding relevant tags such as llama, tequila, quantization, ternary-quantization, and speculative-decoding to accurately describe the model's characteristics and methods.
• Including a "Sample Usage" section with a Python code snippet for eagenerate (speculative decoding) directly from the AngelSlim GitHub README, demonstrating how to use the model with transformers.AutoTokenizer.
• Updating the "Latest Updates" section to include the Tequila implementation release.
• Adjusting the Table of Contents for the new "Sample Usage" section.

These improvements provide more comprehensive and structured information for users.