FAT: Frequency-Aware Transformation for Bridging Full-Precision and Low-Precision Deep Representations.
Sign Up to like & getrecommendations! Published in 2022 at "IEEE transactions on neural networks and learning systems"
DOI: 10.1109/tnnls.2022.3190607
Abstract: Learning low-bitwidth convolutional neural networks (CNNs) is challenging because performance may drop significantly after quantization. Prior arts often quantize the network weights by carefully tuning hyperparameters such as nonuniform stepsize and layerwise bitwidths, which are… read more here.
Keywords: aware transformation; frequency; low precision; quantization ... See more keywords