Offline Reinforcement Learning on D4RL Walker Medium-Replay
62.9RewardDT
Evaluation Results
| Method | Links | |
|---|---|---|
| DTLayers=3, Hidden Channels=256, Heads=42022.02 | 62.9 | |
| Linear Trans.Layers=3, Hidden Channels=256, Heads=42022.02 | 62.1 | |
| FLOWFORMERLayers=3, Hidden Channels=256, Heads=42022.02 | 62 | |
| PerformerLayers=3, Hidden Channels=256, Heads=42022.02 | 61.3 | |
| cosFormerLayers=3, Hidden Channels=256, Heads=42022.02 | 60.5 | |
| ReformerLayers=3, Hidden Channels=256, Heads=42022.02 | 50.1 | |
| AWAC2022.02 | 27 | |
| BC2022.02 | 26 |