forked from wylab/llama.cpp
7cc2d2c889
* ggml : move AMX to the CPU backend --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>