llama.cpp

examples : allow extracting embeddings from decoder contexts

#13797

Merged