Inference engines

The inference engine that much of the local AI ecosystem is built on. A C/C++ implementation that runs GGUF models efficiently across CPU, CUDA, Metal, Vulkan, and more, with fine control over how inference behaves.