shawnshekari/strixite

LLM inference engine written from scratch in HIP for AMD Strix Halo (gfx1151) and Qwen3.8-Flash-Next - ~50 tokens/s out to 492k context

语言：C++ · GitHub 星标：91 · Fork：11

话题：amd · hip · llm-inference · local-llm · qwen · rocm · strix-halo

在 GitHub 上查看原仓库