Robot Manipulation Policy Inference on MimicGen (D0/D1/D2 Breakdown)
3.79Coffee Success Rate (D2)SDP
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SDPSampling Solver=DDPM2026.03 | 3.79 | 3.46 | 3.67 | 3.98 | 3.71 | 3.85 | 3.84 | 3.56 | 3.73 | |
| EquiDiffInput Representation=img2026.03 | 2.44 | 2.07 | 2.36 | 3.68 | 2.34 | 2.39 | 2.41 | 2.38 | 2.51 | |
| EquiBot2026.03 | 2.06 | 2.13 | 1.92 | 1.91 | 1.8 | 2.07 | 2.16 | 2.17 | 2.03 | |
| EquiDiffInput Representation=voxel2026.03 | 1.1 | 1.11 | 1.11 | 1.1 | 1.09 | 1.11 | 1.09 | 1.1 | 1.1 | |
| DP2026.03 | 0.95 | 0.89 | 0.94 | 1.01 | 0.94 | 0.95 | 0.95 | 0.94 | 0.95 | |
| E3Flow2026.03 | 0.51 | 0.51 | 0.51 | 0.52 | 0.49 | 0.54 | 0.5 | 0.53 | 0.51 | |
| SDPSampling Solver=DDIM2026.03 | 0.49 | 0.44 | 0.48 | 0.46 | 0.45 | 0.46 | 0.46 | 0.45 | 0.46 | |
| DP32026.03 | 0.103 | 0.112 | 0.105 | 0.113 | 0.102 | 0.112 | 0.112 | 0.111 | 0.109 |