server: skip device enumeration in router mode to avoid creating CUDA…#23137

allozaur merged 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

ServeurpersoCom:server/avoid-cuda-ctx-on-routerServeurpersoCom/llama.cpp:server/avoid-cuda-ctx-on-routerCopy head branch name to clipboard