Goal-conditioned multi-object manipulation on PandaPush 1 Cube
94.8Success RateEC Diffuser
Evaluation results
| Method | Links | |
|---|---|---|
| EC DiffuserPolicy Training=Separate policies per task2026.03 | 94.8 | |
| VQ-BeTPolicy Training=Separate policies per task2026.03 | 93 | |
| LPWMPolicy Training=Single policy across all tasks, Observation Mode=Multi-view2026.03 | 92.7 | |
| EIT+BCPolicy Training=Separate policies per task2026.03 | 89 | |
| EC Diffusion PolicyPolicy Training=Separate policies per task2026.03 | 88.7 | |
| DiffuserPolicy Training=Separate policies per task2026.03 | 36.7 |