1 paper
Liang Zhao, Kunming Shao, Fengshi Tian +3
Deploying mixed-precision neural networks on edge devices is friendly to hardware resources and power consumption. To support fully mixed-precision neural network inference, it is…