Iterative Deep Model Compression and Acceleration in the Frequency Domain

Yao Zeng (Nanjing University); Xusheng Liu (State Grid Customer Service Center ); Lintan Sun (State Grid Customer Service Center ); Wenzhong Li (Nanjing University)*; Yuchu Fang (Nanjing University); Qingning Lu (Nanjing University); Sanglu Lu (NJU)
PMLR Page

Abstract

Deep Convolutional Neural Networks (CNNs) are successfully applied in many complex tasks, but their storage and huge computational costs hinder their deployment on edge devices. CNN model compression techniques have been widely studied in the past five years, most of which are conducted in the spatial domain. Inspired by the sparsity and low-rank properties of weight matrices in the frequency domain, we propose a novel frequency pruning framework for model compression and acceleration while maintaining high-performance. We firstly apply Discrete Cosine Transform (DCT) on convolutional kernels and train them in the frequency domain to get sparse representations. Then we propose an iterative model compression method to decompose the frequency matrices with a sampled-based low-rank approximation algorithm, and then fine-tune and recompose the low-rank matrices gradually until a predefined compression ratio is reached. We further demonstrate that model inference can be conducted with the decomposed frequency matrices, where model parameters and inference cost can be significantly reduced. Extensive experiments using well-known CNN models based on three open datasets show that the proposed method outperforms the state-of-the-arts in reduction of both the number of parameters and floating-point operations (FLOPs) without sacrificing too much model accuracy.