here's my favorite, the explanation of vLLM prefix caching: http://
docs.vllm.ai/en/v0.8.5/desi
gn/automatic_prefix_caching.html
…

→ View original post on X — @jxmnop