llama.cpp

ggml-cuda: remove unneccesary prints on ggml_cuda_init

#18502

Merged