llama.cpp

1cbd80f8 - examples : support encoder-decoder models in the simple example (#16002)