llama.cpp

server : support multiple model aliases via comma-separated --alias

#19926

Merged