GR00T-N1.7-CubeStack-synth-abs

GR00T-N1.7-CubeStack-synth-abs

GR00T N1.7 (nvidia/GR00T-N1.7-3B) finetuned on 16 synthetic Isaac Lab episodes of the Unitree G1 (Dex3 hands) cube-stacking task, head camera only (cam_left_high). Arm and hand actions are trained with absolute joint-position targets (ActionRepresentation.ABSOLUTE for all action groups).

Part of the G1 Dex3 Cube Stacking Masterprojekt (Fichtl00/Masterprojekt_G1_CubeStack).

• Base model: nvidia/GR00T-N1.7-3B
• Training data: Fichtl00/Cube_Stacking_synth_groot_abs (16 episodes, GR00T-LEAPP unitree_g1 registry hand-joint convention)
• Modality config: action groups left_arm, right_arm, left_hand, right_hand, all absolute
• Training steps: 3000, global batch size 8, lr 1e-4
• Final train_loss: ~0.48

Sibling checkpoints:

• GR00T-N1.7-CubeStack-synth-rel — same data, relative arm actions
• GR00T-N1.7-CubeStack-cotrain-abs — + 50 real episodes, absolute
• GR00T-N1.7-CubeStack-cotrain-rel — + 50 real episodes, relative

Status

Closed-loop evaluation in Isaac Lab sim (JointSpaceV2 env, head camera only): 0/5 successful stacking episodes across multiple actuator stiffness/damping settings tested. Reach behavior is plausible (hand moves toward the correct cube) but grasp execution fails to produce a stable stacking outcome. Provided as-is for research/reproduction purposes.