PRM-as-a-Judge
A dense evaluation paradigm that turns trajectory videos into a goldmine of progress signals. Moving beyond the "Pass/Fail" mirage in robotics.
News 🎉
- July 20, 2026 PRM-as-a-Judge user guide is now available — Follow a streamlined Quick Start and a detailed Advanced Guide from manifest input to progress curves, metrics, and interactive reports.
- June 30, 2026 PRM-as-a-Judge 1.5 is coming — PPL correction, higher-level trajectory diagnostics, and Orca collaboration
- March 30, 2026 RoboPulse benchmark page is now live on Hugging Face — Explore microscopic evaluator verification across 1,800 pairwise samples
- March 23, 2026 Introducing PRM-as-a-Judge — Read the main article for methodology, OPD framework, and interactive demos
- March 20, 2026 RoboChallenge task leaderboard now available — Check out rankings and submit your results
WeChat Community
Scan the QR code to join the WeChat community.
We welcome collaboration and feedback from the robotics community. If you have questions or would like to work with us on dense trajectory evaluation, please reach out at jiyuheng2023@ia.ac.cn.