Robotic Manipulation on Push-T
100Success RateRL-100
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RL-100Type=Robustness against disturbances2025.10 | 100 | — | |
| TASPolicy Architecture=DP2025.11 | 96 | 98 | |
| TASPolicy Architecture=CVAE2025.11 | 95.3 | 97 | |
| DCDPReplanning Frequency (H)=82026.03 | 92.5 | — | |
| Open-LoopReplanning Frequency (H)=82026.03 | 88.4 | — | |
| TASPolicy Architecture=MLP2025.11 | 88.2 | 92.1 | |
| Close-LoopReplanning Frequency (H)=12026.03 | 84.6 | — | |
| Temporal EnsembleReplanning Frequency (H)=12026.03 | 81 | — | |
| TASPolicy Architecture=VQ-BeT2025.11 | 78.1 | 82.9 | |
| EMAPolicy Architecture=DP2025.11 | 74.1 | 91.9 | |
| BID-OPolicy Architecture=DP2025.11 | 73.3 | 93.4 | |
| Vanilla-OPolicy Architecture=DP2025.11 | 72.6 | 93.8 | |
| BID-CPolicy Architecture=DP2025.11 | 65.8 | 87.7 | |
| Vanilla-CPolicy Architecture=DP2025.11 | 64.9 | 88.4 | |
| π0+FPGenerative Mode=Flow Matching, Model Type=VLA & VA2025.10 | 62.25 | — | |
| π0Generative Mode=Flow Matching, Model Type=VLA2025.10 | 57.69 | — | |
| Florence-Policy-F+FPGenerative Mode=Flow Matching, Model Type=VLA & VA2025.10 | 56.06 | — | |
| Flow Policy (FP)Generative Mode=Flow Matching, Model Type=VA2025.10 | 54.25 | — | |
| Vanilla-OPolicy Architecture=VQ-BeT2025.11 | 46 | 70.4 | |
| EMAPolicy Architecture=CVAE2025.11 | 43.8 | 78 | |
| BID-OPolicy Architecture=VQ-BeT2025.11 | 43.4 | 67.6 | |
| BID-OPolicy Architecture=CVAE2025.11 | 43.1 | 79.3 | |
| Florence-Policy-D+DPGenerative Mode=Diffusion, Model Type=VLA & VA2025.10 | 43.06 | — | |
| EMAPolicy Architecture=VQ-BeT2025.11 | 42.6 | 65.1 | |
| Florence-Policy-D+MPGenerative Mode=Diffusion, Model Type=VLA & VA2025.10 | 40.88 | — | |
| Florence Policy-DGenerative Mode=Diffusion, Model Type=VLA2025.10 | 40 | — | |
| Vanilla-OPolicy Architecture=CVAE2025.11 | 39.5 | 77.6 | |
| Florence Policy-FGenerative Mode=Flow Matching, Model Type=VLA2025.10 | 39.38 | — | |
| EMAPolicy Architecture=MLP2025.11 | 33.2 | 67.8 | |
| Vanilla-OPolicy Architecture=MLP2025.11 | 33 | 66.2 | |
| Vanilla-CPolicy Architecture=VQ-BeT2025.11 | 30.2 | 56 | |
| DP+MPGenerative Mode=Diffusion, Model Type=VA & VA2025.10 | 23.63 | — | |
| BID-CPolicy Architecture=VQ-BeT2025.11 | 23.1 | 48.7 | |
| Diffusion Policy (DP)Generative Mode=Diffusion, Model Type=VA2025.10 | 21.75 | — | |
| Mamba Policy (MP)Generative Mode=Diffusion, Model Type=VA2025.10 | 12.06 | — | |
| BID-CPolicy Architecture=CVAE2025.11 | 5.5 | 40.7 | |
| Vanilla-CPolicy Architecture=CVAE2025.11 | 3.7 | 36.1 | |
| FineTunePolicy Architecture=CVAE2025.11 | 2.3 | 86.1 | |
| FineTunePolicy Architecture=MLP2025.11 | 2.1 | 79.2 | |
| Vanilla-CPolicy Architecture=MLP2025.11 | 0.2 | 22.2 |