Search papers, labs, and topics across Lattice.
2
0
4
SAC achieves a staggering 2.1x throughput increase by fetching only the essential KV entries for sparse attention models, revolutionizing memory efficiency in LLM inference.
DisagFusion unlocks up to 20x higher throughput for diffusion model serving by intelligently splitting the workload across heterogeneous GPUs and dynamically adapting to workload shifts.
