Gallery · LIBERO-Pro

Put the wine bottle on the rack

libero_goal_object, task 9 · seed 2. DeepSeek-V4.1-Flash, code mode: 10 decisions, 13 motions, solved.

2× · 87 s of motion
  1. 1 of 10 · look_rack3.py
  2. 2 of 10 · look_rack4.py
  3. 3 of 10 · look_rack5.py
  4. 4 of 10 · grasp_bottle.py
  5. 5 of 10 · grasp3.py
  6. 6 of 10 · grasp4.py
  7. 7 of 10 · grasp5.py
  8. 8 of 10 · place1.py
  9. 9 of 10 · insert.py
  10. 10 of 10 · place_top.py
  1. move_pose
  2. move_pose
  3. move_pose
  4. set_gripper
  5. move_pose
  6. move_pose
  7. move_pose
  8. set_gripper
  9. move_pose
  10. move_pose
  11. move_pose
  12. move_pose
  13. move_pose
Decisions

The recording is the robot's own cameras at 2×; the time between motions, while the model thinks, is not filmed. The 3D replay is the camera's picture in depth after each decision, with the paths it planned, to turn and step through.