Abstract: Fixed-point quantization techniques have attracted considerable attention in deep neural network (DNN) inference acceleration. Nevertheless, they often require time-consuming fine-tuning or ...