llama.cpp

d583cd03 - ggml : Fix compilation issues on ARM platform when building without fp16 (#10811)