llama.cpp

make sure params --split and --merge are not specified at same time in gguf-split

#9619

Merged