robot-learning-vislab/fr3-pick-tomato

20 demonstrations of picking up a toy tomato and holding it above the table.

Teleoperated with a GELLO leader arm through robot-stack on a Franka FR3 under a cartesian impedance controller, recorded from a ZED 2i left camera.

episodes	20
frames	2620
rate	10 Hz
robot	franka_crisp
camera	1280x720, uncropped
task string	pick up the tomato and lift it above the table

Each episode starts from a randomised pose (±2 cm in xy, ±15° in yaw about a measured start pose) that the arm drives to by itself, so the starts are drawn from the same distribution a policy would be reset to. Between episodes the operator puts the object back on its mark by hand while the arm holds still.

State is 14-D (fr3_joint*.pos, ee.x..ee.rz as a base-frame position and rotation vector, gripper.pos normalised by an 80 mm opening). Actions are 7-D: six pose deltas plus a gripper command, which uses the leader trigger's own hysteresis (shuts below 0.45, opens above 0.55).

Notes

• Uncropped: the whole 1280x720 ZED frame is kept so a crop can be chosen per downstream task.
• The toy tomato is soft and compresses under the 20 N grasp, so the fingers settle at 1-7 mm of an 80 mm opening. Frames confirm the tomato is gripped and lifted clear of the table; the narrow reading is the object yielding, not an empty gripper.
• Every episode ends with the tomato still held, so there is one gripper transition per episode and no release.