Offline Reinforcement Learning on D4RL Antmaze-Diverse
9,240AntMaze-MediumDiffuserLite
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DiffuserLiteBackbone=D, Runtime per action (second)=0.022024.01 | 9,240 | 6,800 | 7,760 | |
| DiffuserLiteBackbone=R1, Runtime per action (second)=0.0142024.01 | 8,920 | 8,040 | 8,250 | |
| DiffuserLiteBackbone=R2, Runtime per action (second)=0.0082024.01 | 8,760 | 7,520 | 8,030 | |
| IQL2024.01 | 7,380 | 3,030 | 5,300 | |
| CQL2024.01 | 6,730 | 2,050 | 4,360 | |
| BC2024.01 | 80 | 0 | 20 | |
| DiffuserRuntime per action (second)=0.7492024.01 | 0 | 0 | 0 | |
| DDRuntime per action (second)=2.4352024.01 | 0 | 0 | 0 |