Open the box on Real-World Dexterous Manipulation Object-Unseen Setting
50Success RateBORA-Full
Evaluation Results
| Method | Links | |
|---|---|---|
| BORA-FullLearning Paradigm=Reinforcement Learning (RL) Fine-tuning2026.05 | 50 | |
| BORA-OfflineLearning Paradigm=Reinforcement Learning (RL) Fine-tuning2026.05 | 35 | |
| CP BaseLearning Paradigm=Pure Imitation Learning (IL)2026.05 | 15 | |
| VITRA (Diffusion)Learning Paradigm=Pure Imitation Learning (IL)2026.05 | 10 | |
| CP + Decoupled CriticLearning Paradigm=Reinforcement Learning (RL) Fine-tuning2026.05 | 5 |