llama.cpp

6e264a90 - docker : add GGML_CPU_ARM_ARCH arg to select ARM architecture to build for (#11419)