Value-aware Quantization for Training and Inference of Neural Networks | Read Paper on Bytez