Reinforcement Learning on D4RL HalfCheetah no thighs (medium)
3,910Mean ReturnVGDF + BC
Evaluation results
| Method | Links | |
|---|---|---|
| VGDF + BCtarget interaction steps=10^52023.05 | 3,910 | |
| H2Otarget interaction steps=10^52023.05 | 3,023 | |
| Symmetric samplingtarget interaction steps=10^52023.05 | 2,211 | |
| Offline onlybase algorithm=CQL, target interaction steps=10^52023.05 | 361 |