text-generation-inference

e14ae3b5 - feat(server): support quantization for flash models (#200)