Robotic Manipulation on Robomimic Transport mh 2D state-based
45Success Rate (Max Performance)DP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DPNFE=100, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 45 | 34 | |
| MIPNFE=2, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 44 | 35 | |
| IDPNFE=1, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 41 | 29 | |
| FlowNFE=9, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 33 | 23 | |
| Naive DriftingNFE=1, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 30 | 22 | |
| MP1NFE=1, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 24 | 14 | |
| CPNFE=1, Observation Modality=2D state-based, Network Architecture=Transformer, CNN (average)2026.05 | 1 | 1 |