Generated by data/utils/convert_piper_hdf5_to_lerobot.py from challenge_deploy HDF5.
Repo id: ZhaoRunyi/Piper_traffic_light_water_0612
Input dir: /home/edemlab/challenge_ws/data/challenge_hdf5/PourBeaker_Long
FPS: 10
Visual backend: video
State/action order: per arm joint6, gripper01, ee_pos3, ee_rot6d6.
observation.state and action use identical 32D names:
left_joint_waist, left_joint_shoulder, left_joint_elbow, left_joint_forearm_roll… See the full description on the dataset page:
https://huggingface.co/datasets/ZhaoRunyi/Piper_traffic_light_water_0612.