llama.cpp

5d67acd4 - ggml : check int overflow in ggml_new_tensor_impl and ggml_new_object