llama.cpp

llama-quant : fix `--tensor-type` when default `qtype` is overriden

#22572

Merged