Robot Manipulation on Dexart
85Success RatePointNet-DP3
Evaluation results
| Method | Links | |
|---|---|---|
| PointNet-DP3input=privileged background-cropped 3D point clouds2026.01 | 85 | |
| KAN-We-Flowreproduced=false2026.02 | 85 | |
| FlowPolicyreproduced=true2026.02 | 81 | |
| MambaPolicyreproduced=true2026.02 | 80 | |
| DP3reproduced=true2026.02 | 78 | |
| Simple DP3reproduced=true2026.02 | 75 | |
| Diffusion Policyreproduced=false2026.02 | 69 | |
| Depth-Anything2026.01 | 66 | |
| X-Distill2026.01 | 63.5 | |
| X-DP3Horizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 59.4 | |
| DP3+UniAuxHorizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 59.3 | |
| DINOv22026.01 | 58 | |
| DP3+SepAuxHorizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 57.4 | |
| DP3+HybridHorizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 57.3 | |
| DP3+MBAHorizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 57 | |
| DP3Horizon (H)=16, Conditional Trajectory Length (N)=8, Number of trials=300, Checkpoint strategy=Final checkpoint2026.05 | 55.9 | |
| ResNet-scratch2026.01 | 54.5 | |
| BCRNNreproduced=false2026.02 | 29 | |
| Theia2026.01 | 24 | |
| IBCreproduced=false2026.02 | 1 |