Search papers, labs, and topics across Lattice.
Sichuan University
1
0
3
Action chunk utilization triples and physical execution steps drop by over 50%, resulting in a 5.83x speedup in VLA model deployment without sacrificing performance.
