Offline Reinforcement Learning on D4RL Walker Medium-Expert
108RewardFLOWFORMER
Evaluation Results
| Method | Links | |
|---|---|---|
| FLOWFORMERLayers=3, Hidden Channels=256, Heads=42022.02 | 108 | |
| DTLayers=3, Hidden Channels=256, Heads=42022.02 | 107.7 | |
| BC2022.02 | 107.5 | |
| cosFormerLayers=3, Hidden Channels=256, Heads=42022.02 | 100.5 | |
| AWAC2022.02 | 74.5 | |
| PerformerLayers=3, Hidden Channels=256, Heads=42022.02 | 72.6 | |
| ReformerLayers=3, Hidden Channels=256, Heads=42022.02 | 71.4 | |
| Linear Trans.Layers=3, Hidden Channels=256, Heads=42022.02 | 67.2 |