Brief Announcement: PRESERVE: Prefetching Model Weights and KV-Cache in Distributed LLM Serving.

Ahmet Caner Yzgler, Jiawei Zhuang, Lukas Cavigelli

Browse the full SPAA paper archive.