| title: RoboGen | |
| emoji: 🤖 | |
| colorFrom: purple | |
| colorTo: blue | |
| sdk: gradio | |
| sdk_version: 5.9.1 | |
| app_file: app.py | |
| pinned: false | |
| python_version: 3.11 | |
| # RoboGen — Synthetic Robotics Dataset Generator | |
| **Physics-accurate LeRobot-format datasets for SO-100, SO-101, and Koch arms.** | |
| Generate synthetic datasets with physically-plausible trajectories, automatic quality | |
| scoring against HaptalAI's misalignment benchmark, and one-click download. | |
| ## Features | |
| - Cubic spline trajectories with analytical velocity derivatives (not random walks) | |
| - Three failure modes: grasp slip, velocity spike, torque saturation | |
| - Quality scoring via calibrated MAD z-score thresholds (community-benchmarked) | |
| - Downloadable zip: parquet dataset + auto-generated README | |
| ## Contact | |
| aarav@haptal.ai | |