Offline Behavior Distillation on D4RL Hopper medium
56.4Normalized ReturnOffRL (Whole)
Evaluation Results
| Method | Links | |
|---|---|---|
| OffRL (Whole)Dataset=Whole offline dataset, Algorithm=Cal-QL2025.12 | 56.4 | |
| BC (Whole)Dataset=Whole offline dataset2025.12 | 50.2 | |
| SDWNsyn=256, tau=0.12025.12 | 38.4 | |
| DBCNsyn=2562025.12 | 37.8 | |
| Av-PBCNsyn=2562025.12 | 32.5 | |
| Rand (Dreal)Selection=Random, Source=Dreal, Nsyn=2562025.12 | 29.1 | |
| PBCNsyn=2562025.12 | 25.1 | |
| Rand (Doff)Selection=Random, Source=Doff, Nsyn=2562025.12 | 19.2 |