Dynamic Push-T Task Evaluation
100Success RateOnline RL (DDIM)
Evaluation Results
| Method | Links | |
|---|---|---|
| Online RL (DDIM)Method Category=RL-100 (ours), Policy Architecture=DDIM, Learning Stage=Online RL2025.10 | 100 | |
| Online RL (CM)Method Category=RL-100 (ours), Policy Architecture=one-step CM policy, Learning Stage=Online RL2025.10 | 100 | |
| Iterative Offline RLMethod Category=RL-100 (ours), Learning Stage=Iterative offline fine-tuning stage2025.10 | 90 | |
| DP3Method Category=Imitation baselines2025.10 | 64 | |
| DP-2DMethod Category=Imitation baselines2025.10 | 40 |