A cool collection of research papers that build on top of vLLM and optimize it further:

→ View original post on X — @julien_c