llama.cpp

model-conversion : add qat-q4 quantization targets

#15588

Merged