Use llama.cpp(convert.py) to run inference on this model