llama.cpp

Fix build in baby-llama after ggml_rope change

#2016

Merged