llama.cpp

gguf_convert_endian.py: skip MXFP4 data

#17523

Merged