Offline Reinforcement Learning on D4RL Average
73.5Average RewardFLOWFORMER
Evaluation results
| Method | Links | |
|---|---|---|
| FLOWFORMERLayers=3, Hidden Channels=256, Heads=42022.02 | 73.5 | |
| DTLayers=3, Hidden Channels=256, Heads=42022.02 | 72.2 | |
| cosFormerLayers=3, Hidden Channels=256, Heads=42022.02 | 67.8 | |
| Linear Trans.Layers=3, Hidden Channels=256, Heads=42022.02 | 64.4 | |
| ReformerLayers=3, Hidden Channels=256, Heads=42022.02 | 63.9 | |
| PerformerLayers=3, Hidden Channels=256, Heads=42022.02 | 63.8 | |
| BC2022.02 | 51.9 | |
| AWAC2022.02 | 50.1 |