Skip to content

Flash Next on CUDA: serve NVIDIA's NVFP4 checkpoint on one GB10 or at --tp 2 - #1771

Merged
ashhart merged 14 commits into
ashhart:mainfrom
laudney:pr-fn-nvfp4
Nov 3, 2026
Merged

ashhart merged 14 commits into
ashhart:mainfrom
laudney:pr-fn-nvfp4

Commits

  1. Commits on Nov 3, 2026