3D Diffuser Actor

Reported on 6 benchmarks across 2 tasks · 1 paper · 3 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Robots5 results

Robot ManipulationonRLBench
Succ. Rate (18 tasks, 100 demo/task)· 2024-02-16
81.3
best: 89.4 (EquAct)
SOTA
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885
Robot ManipulationonRLBench
Training Time (A100 x hour)· 2024-02-16
936
SOTA
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885
Robot ManipulationonRLBench
Input Image Size· 2024-02-16
256
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885
Robot ManipulationonRLBench
Training Time (V100 x 8 x day)· 2024-02-16
8
best: 16 (PerAct (Evaluated in RVT))
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885
Robot ManipulationonThe COLOSSEUM
Average decrease average across all perturbations· 2024-02-16
-15.6
best: -4.3 (SAM2Act)
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885

Methodology1 result

Zero-shot GeneralizationonCALVIN
Avg. sequence length· 2024-02-16
3.27
best: 4.04 (GR-MG)
SOTA
3D Diffuser Actor: Policy Diffusion with 3D Scene Representations arXiv:2402.10885