Offline Reinforcement Learning on D4RL HalfCheetah Med-Expert v2
105.9Avg Normalized ReturnMBOP
Evaluation results
| Method | Links | |
|---|---|---|
| MBOP2021.06 | 105.9 | |
| Trajectory Transformerdiscretization=quantile2021.06 | 95 | |
| Onestep2022.02 | 93.4 | |
| CQL2021.06 | 91.6 | |
| CQL2022.02 | 91.6 | |
| TD3+BC2022.02 | 90.7 | |
| SPOT2022.02 | 86.9 | |
| Decision Transformer2021.06 | 86.8 | |
| DT2022.02 | 86.8 | |
| IQL2022.02 | 86.7 | |
| Behavior Cloning2021.06 | 59.9 | |
| BC2022.02 | 55.2 | |
| AWAC2022.02 | 42.8 | |
| BRAC2021.06 | 41.9 | |
| Trajectory Transformerdiscretization=uniform2021.06 | 40.8 |