llama.cpp

26a48ad6 - ggml : prevent integer overflow in gguf tensor size calculation (#14595)