Goal-Conditioned Reinforcement Learning (Manipulation) on puzzle-3x3-play state-based v0 (test)
14Success RateVIB
Evaluation Results
| Method | Links | |
|---|---|---|
| VIBDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 14 | |
| OrigDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 5 | |
| TRADownstream algorithm=GCIVL, Observation type=State-based2025.10 | 5 | |
| Dual Goal RepresentationsDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 5 | |
| VIPDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 3 | |
| BYOL-γDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 0 |