1 repository on SrcLog
Ultra-fast LLM inference engine — Vulkan backend, no CUDA required, AMD/Intel/NVIDIA