ggml

ff54fda8 - gpt-2 : loading Q4_0 quantized model