Hardware for Quantized Mixed-Precision Deep Neural Networks | IEEE Conference Publication | IEEE Xplore