lehome/dataset_challenge_real
Viewer β’ Updated β’ 61 β’ 121
This is my policy for the real-world final of the LeHome Challenge 2026 (ICRA 2026), a competition on bimanual garment folding with a physical LeRobot SO-ARM101 setup. It placed 2nd.
It is a Ο0.5-based vision-language-action policy, trained as behavior cloning on a mix of organizer teleop data, my own teleop / DAgger corrections, and heavily-augmented simulation replays. It builds on the simulation-round policy.
For how it works and how to run it, please refer to the code repository and the tech report.