Goal-conditioned multi-object manipulation on PandaPush 2 Cubes
91.7Success RateEC Diffuser
Evaluation results
| Method | Links | |
|---|---|---|
| EC DiffuserPolicy Training=Separate policies per task2026.03 | 91.7 | |
| LPWMPolicy Training=Single policy across all tasks, Observation Mode=Multi-view2026.03 | 74 | |
| EC Diffusion PolicyPolicy Training=Separate policies per task2026.03 | 38.8 | |
| EIT+BCPolicy Training=Separate policies per task2026.03 | 14.6 | |
| VQ-BeTPolicy Training=Separate policies per task2026.03 | 5.2 | |
| DiffuserPolicy Training=Separate policies per task2026.03 | 1.3 |