llama.cpp

0d7c1435 - convert: add error message for mistral3 quantized weight