yuuoniy's picture
ggml : prevent integer overflow in gguf tensor size calculation (llama/14595)
31f34e7