JetBrains-Research

/

PIPer-8B-SFT-only

like 0

Follow

JetBrains Research 123

This PR significantly enhances the model card for JetBrains-Research/PIPer-8B by:

• Updating Metadata: Adding the code-generation tag to improve discoverability, as the model specializes in generating Bash scripts for environment setup.
• Adding Prominent Links: Including explicit links to the paper (PIPer: On-Device Environment Setup via Online Reinforcement Learning), the GitHub repository (https://github.com/JetBrains-Research/PIPer), and the project homepage (https://jb.gg/PIPer) at the top of the content section for easy access.
• Introducing Sample Usage: Adding a "Sample Usage" section with a Python code snippet extracted from the GitHub README's "Play with the Model Yourself" instructions. The policy_lm argument in the snippet has been adjusted to point directly to JetBrains-Research/PIPer-8B to ensure immediate usability for users. This provides a practical example of how to interact with the model's agent.

All other existing valuable information and the overall structure of the model card have been preserved.