guox18

/

Qwen2.5-7B-Instruct-IFDecorator

like 0

This PR significantly enhances the model card for the IFDECORATOR model, presented in IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards.

Key additions include:

• Relevant metadata: pipeline_tag: text-generation and library_name: transformers for improved discoverability and the correct display of the "how to use" widget on the Hub.
• Links to the paper, project page, and the inferred official GitHub repository (https://github.com/tianyilt/IFDecorator).
• A concise overview of the model's purpose, key innovations, and performance highlights, derived from the paper's abstract.
• A basic Python code snippet for quick inference using the transformers library, demonstrating its use with a chat template.
• A BibTeX citation for the paper.

These updates will provide users with essential information and a clear starting point for using the model effectively.