Action Prediction on Multi-Robot Manipulation
92Action AccuracyConcat-View
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Concat-Viewmulti-view conditioning strategy=Concat-View, number of camera views=22026.04 | 92 | 0.82 | |
| MultiWorldmulti-view conditioning strategy=Ours2026.04 | 88.7 | 1.52 | |
| Combomulti-view conditioning strategy=Combo2026.04 | 88.5 | 1.54 | |
| Standardmulti-view conditioning strategy=Standard2026.04 | 88.2 | 1.6 |