llama.cpp

llama-server: fix duplicate HTTP headers in multiple models mode

#17698

Merged