Press — Unitree G1 Humanoid Teleoperation: "Press the Target Interface" Demonstrations (LeRobot)

Multi-view humanoid teleoperation demonstrations of a contact-rich press-and-release manipulation task, captured on a Unitree G1 robot operated by a human teleoperator. Task: "Press the target interface until it is actuated, then release." Packaged as a HuggingFace LeRobot v2.1 dataset (3 episodes, 1,975 frames total) and QC-complete (Tier-1, 0 FAILs). Six synchronized video streams at ~30 fps, all H.264: head RGB (1920×1080), left + right color wrist/side RGB (1280×720), a head-depth visualization (1280×720), a LiDAR render (640×480), and a third-person scene overview (1280×578). The parquet carries dense proprioception — 26-DOF joint state, 29-DOF action, joint velocity/torque, Cartesian pose/velocity/wrench, fingertip positions, tactile (normal/tangential/proximity), and a 4,096-point LiDAR point cloud per frame — plus frame timestamps + indices. Useful for contact-rich manipulation, force/tactile-aware policy learning, multi-view representation learning, and VLA pretraining on humanoid teleoperation data.