xpolicylab.github.io

Project website for XPolicyLab — a unified standard and open ecosystem for robot policy evaluation and deployment.

Local preview

python3 -m http.server 8080

Open http://localhost:8080.

Structure

├── index.html
├── contributors.html          # redirects to index
├── css/styles.css
├── css/leaderboard.css        # RoboTwin / RoboDojo Top 10 widgets
├── js/main.js
├── js/leaderboard/widget.js   # live-sync board mounts
├── data/boards.json           # registry of remote board URLs
└── assets/
    ├── logo.png
    ├── booth_transparent.png  # real-eval cell
    ├── partners/              # optional partner marks
    └── paper/
        ├── XPolicyLab.pdf
        ├── fig1.png           # policy zoo overview
        ├── fig2.png           # infrastructure
        └── fig3.png           # cross-platform eval

Live leaderboards

The homepage mounts three Top 10 widgets:

• RoboTwin 2.0 (live): https://robotwin-platform.github.io/data/robotwin_leaderboard.json
• RoboDojo Sim (official 2026-09-28 snapshot, 51 models + type tags): js/leaderboard/robodojo_sim.js (fallback assets/leaderboards/robodojo_sim.json)
• RoboDojo Real (live): https://robotwin-platform.github.io/data/robodojo_real_leaderboard.json

The robotwin-platform Sim mirror (robodojo_leaderboard.json, dated 2026-09-05) is behind the official board at robodojo-benchmark.com/leaderboard.

Drop another mount anywhere:

<div data-xpl-lb="robotwin" data-top="10" data-refresh="300"></div>
<div data-xpl-lb="robodojo" data-top="10" data-refresh="300"></div>
<div data-xpl-lb="robodojo-real" data-top="10" data-refresh="300"></div>

data-refresh is seconds between refetches (omit to load once). Override data-url / data-page / data-top per mount if a later page needs a different source.

Links

• Code: XPolicyLab/XPolicyLab
• Contribute: CONTRIBUTING.md