Search papers, labs, and topics across Lattice.
1
0
3
Collaborative edge-cloud LLM inference can achieve 275% higher throughput by dynamically synchronizing heterogeneous requests to batch their common cloud-side suffix.