Aravind08/dataforge-dual-hand-test

A LeRobot dataset for the SO101, built by DataForge from human demonstration video. 1 episodes, 2805 frames at 30 fps.

Columns

Column	What it is
observation.images.top	The operator's video, re-encoded
observation.state	Joint positions, LeRobot degrees for joints 1-5 and 0-100 for the gripper
action	The commanded joint positions, same units

Two arms, one per hand of the person: 12 values, the left arm's six (shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper) then the right arm's, named left_arm_shoulder_pan.pos … right_arm_gripper.pos as a two-arm SO-101 (bi_so101_follower) records them. Joints 1-5 are -100 to 100 of each joint's calibrated range (use_degrees=False), the gripper 0-100. The left arm is the hand that worked on the left of the picture. Both arms share one clock. The arm units are LeRobot degrees, as SO101FollowerConfig(use_degrees=True) reports them; the gripper is 0-100 of its own travel, 0 closed.

How it was made

Nobody teleoperated a robot to record this. Each episode is a phone video of a person doing the task with their own hand, retargeted onto the SO-101 by hand2robot-1.0: a 3D hand reconstruction, then differential IK onto the SO-101 URDF, then retiming inside the servos' speed and acceleration limits.

Episodes are on the robot's clock, which runs about 1.6 to 2 times slower than the human's, so video frames repeat where the arm moves slowly. The travel to and from the calibration pose is not included.

Only clips that passed both an automatic content check and a trajectory quality gate are here. The gate rejects, among other things, clips where the manipulated object was tracked in too few frames, where too much of the hand path lies outside the arm's reach.

What to know before you train on it

• observation.state is not measured. It is the previous commanded position: the IK trajectory, assuming the servos track it perfectly. No physical robot executed these trajectories.
• The gripper closes when the object is held, timed from the steps the automatic check named ("pick up ...", "place ..."). A hand around an object keeps thumb and index apart, so it is not read off the fingers. Episodes whose steps do not say follow the thumb-index distance.
• The geometry is uncalibrated. The camera focal length is approximated from the frame size and the hand scale is a population average, so expect a few centimetres of absolute error.
• 0 of 1 episodes were solved against a real lerobot-calibrate file. Episodes without one assume symmetric joint ranges: fine to train on, wrong to replay on a particular arm.
• Open loop. The object is where the video said it was. A real arm needs the same layout relative to its base, or its own perception.