Search papers, labs, and topics across Lattice.
1
0
3
7
A single RL policy trained on procedurally generated tools in simulation can achieve zero-shot dexterous manipulation of diverse real-world tools, rivaling task-specific policies.
