LLM4Binary

/

llm4decompile-9b-v2

like 19

This PR enhances the model card for LLM4Decompile-6.7B-v2 by:

• Adding pipeline_tag: text-generation to improve discoverability on the Hugging Face Hub, as the model generates C source code.
• Adding library_name: transformers to enable the automated "How to use" widget, as evidenced by the model's usage of AutoTokenizer and AutoModelForCausalLM.
• Including datasets: LLM4Binary/decompile-bench in the metadata, as the model is based on research introducing this dataset.
• Adding a clear model card title and an updated introduction that integrates information about the model's purpose and its connection to the paper Decompile-Bench: Million-Scale Binary-Source Function Pairs for Real-World Binary Decompilation.
• Adding the official citation for the paper, as provided in the project's GitHub README.
• Ensuring all existing code snippets retain their literal newline characters (\n) as per guidelines.

These changes provide more comprehensive and structured information, making the model easier to find and understand for users.