llama.cpp
arg : allow up to llama_max_devices() configs for --tensor-split and --fit-target (#13293)
#31435
Merged

Loading