Improve model card: Highlight Tequila paper and add metadata
#5
by nielsr HF Staff - opened
This PR significantly improves the model card for the AngelSlim toolkit, with a focus on documenting the Tequila quantization method presented in the paper Tequila: Trapping-free Ternary Quantization for Large Language Models.
Key changes include:
- Adding a new section dedicated to the Tequila paper, including its abstract, for better discoverability and context.
- Adding
pipeline_tag: text-generationto align with the model's primary use case. - Adding
library_name: transformersas evidenced by theconfig.jsonand usage examples. - Setting
license: apache-2.0, confirmed by the project's LICENSE file. - Adding relevant tags
quantizationandternary-quantizationto the metadata. - Updating the citation section to include the Tequila paper's BibTeX entry.
Please review and merge this PR.