Goal-Conditioned Reinforcement Learning (Navigation) on pointmaze-large-navigate state-based v0 (test)
52Success RateOrig
Evaluation Results
| Method | Links | |
|---|---|---|
| OrigDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 52 | |
| VIBDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 50 | |
| Dual Goal RepresentationsDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 46 | |
| BYOL-γDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 22 | |
| TRADownstream algorithm=GCIVL, Observation type=State-based2025.10 | 1 | |
| VIPDownstream algorithm=GCIVL, Observation type=State-based2025.10 | 0 |