MLUTNet: A Neural Network for Memory Based Reconfigurable Logic Device Architecture

Neural networks have been widely used and implemented on various hardware platforms, but high computational costs low similarity of network structures relative to are often obstacles research. In this paper, we propose a novel neural in combination with the structural features recently proposed memory-based programmable logic device, compare it standard structure, test common datasets full binary precision, respectively. The experimental results reveal that new structured can provide almost consistent full-precision performance binary-precision ranging from 61.0% 78.8% after using sparser connections about 50% reduction size weight matrix.