llama.cpp

ggml : change ggml_scale to take a float instead of tensor

#4573

Merged