Search papers, labs, and topics across Lattice.
1
0
3
Achieve 30x computational cost reduction in end-to-end autonomous driving MLLMs without sacrificing performance using a novel supervised token reduction framework.
