Betelgeuse-AI/Betelgeuse

A custom fork for frontier-scale mixture-of-experts inference on an RTX 3060 with 12 GB VRAM, 16 GB system RAM and local NVMe.** Betelgeuse keeps dense computation on the GPU and brings routed experts into a bounded VRAM pool only when needed.

语言：C++ · GitHub 星标：69 · Fork：1

在 GitHub 上查看原仓库