Offline Reinforcement Learning (Robotic Manipulation) on Franka Kitchen partial D4RL (pt)
80.5Expert Normalized ReturnDMM
Evaluation results
| Method | Links | |
|---|---|---|
| DMMArchitecture=SSM-based2026.02 | 80.5 | |
| DCArchitecture=Transformer-based2026.02 | 73.6 | |
| WTArchitecture=Transformer-based2026.02 | 63.8 | |
| DMArchitecture=SSM-based2026.02 | 61.7 | |
| DS4Architecture=SSM-based2026.02 | 52.9 | |
| CQLArchitecture=Value-based2026.02 | 49.8 | |
| IQLArchitecture=Value-based2026.02 | 46.3 | |
| TD3+BCArchitecture=Value-based2026.02 | 38.3 | |
| QLDTArchitecture=Transformer-based2026.02 | 36.9 | |
| DTArchitecture=Transformer-based2026.02 | 31.4 |