Search papers, labs, and topics across Lattice.
1
0
3
Achieve 10% higher success rates in robotic manipulation tasks while speeding up inference by 1.5-1.8x by intelligently pruning visual tokens in multi-view Vision-Language-Action models.
