llama.cpp

quantize: fix minor logic flaw in --tensor-type

#14572

Merged