This dataset was created using LeRobot.
This was a test of a minimal "robot": teleop and robot are each a single digital channel. This caused lerobot training (using ACT policy) to fail, giving an error about multiplying matrices with incorrect dimensions. Conclusion: I don't think lerobot supports single degree of freedom robots.
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aux_channel_follower"… See the full description on the dataset page:
https://huggingface.co/datasets/primordial-spork/lerobot_teleop_aux_channel_max_100.
