llama.cpp

63fd76fb - Reduce model loading time (#43)

Reduce model loading time (#43) * Use buffering * Use vector * Minor --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>