Online Reinforcement Learning on FingerTurnHard DMControl (final)
884.59Normalized ReturnGoRL(Diff)
Evaluation results
| Method | Links | |
|---|---|---|
| GoRL(Diff)Seeds=5, Decoder=Diffusion2025.12 | 884.59 | |
| GoRL(FM)Seeds=5, Decoder=Flow-Matching2025.12 | 860.83 | |
| FPOSeeds=52025.12 | 752.08 | |
| PPOSeeds=52025.12 | 738.7 | |
| DPPOSeeds=52025.12 | 633.84 |