Multi-task Robotic Manipulation on LIBERO standard suites
87.5Goal Success RateDGPO
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DGPOInput modality=State2026.06 | 87.5 | 94.8 | 99.4 | 59.1 | 85.2 | |
| MT-PPOInput modality=State2026.06 | 69.9 | 20.8 | 59.6 | 19.8 | 42.5 | |
| DGPOInput modality=Visual2026.06 | 66.1 | 86.9 | 85.1 | 41.2 | 69.8 | |
| MT-RLPDInput modality=State2026.06 | 50 | 0 | 79.4 | 0.6 | 32.5 | |
| MT-RFCLInput modality=State2026.06 | 49.4 | 0 | 85 | 19.4 | 38.4 | |
| MT-PPOInput modality=Visual2026.06 | 47.1 | 24.6 | 79.9 | 29.3 | 45.2 | |
| MT-DAPGInput modality=Visual2026.06 | 18.9 | 9.5 | 47.7 | 0 | 19 | |
| MT-DAPGInput modality=State2026.06 | 18.1 | 1.5 | 41.6 | 0 | 15.3 | |
| MT-DeepMimicInput modality=State2026.06 | 10 | 0 | 0 | 0 | 2.5 | |
| MT-DeepMimicInput modality=Visual2026.06 | 9.1 | 0 | 0 | 0 | 2.3 |