Offline Behavior Distillation on D4RL Halfcheetah (medium)
47.6Normalized ReturnOffRL (Whole)
Evaluation Results
| Method | Links | |
|---|---|---|
| OffRL (Whole)Dataset=Whole offline dataset, Algorithm=Cal-QL2025.12 | 47.6 | |
| BC (Whole)Dataset=Whole offline dataset2025.12 | 42.3 | |
| SDWNsyn=256, tau=0.12025.12 | 39.5 | |
| Av-PBCNsyn=2562025.12 | 36.9 | |
| PBCNsyn=2562025.12 | 30.9 | |
| DBCNsyn=2562025.12 | 28.2 | |
| Rand (Dreal)Selection=Random, Source=Dreal, Nsyn=2562025.12 | 5.9 | |
| Rand (Doff)Selection=Random, Source=Doff, Nsyn=2562025.12 | 1.8 |