Inside vLLM: Anatomy of a High-Throughput LLM Inference System
From paged attention, continuous batching, prefix caching, specdec, etc. to multi-GPU, multi-node dynamic serving at scale.
Topic
#cpt5
1 篇笔记
Topic
1 篇笔记
From paged attention, continuous batching, prefix caching, specdec, etc. to multi-GPU, multi-node dynamic serving at scale.