llama.cpp

Using abort_callback from ggml to interrupt llama computation

#5409

Merged