Depth-wise Decomposition for Accelerating Separable Convolutions in Efficient Convolutional Neural Networks
This work addresses efficiency and accuracy trade-offs in CNNs for applications like robotics and self-driving cars, but it is incremental as it builds on existing separable convolution methods.
The paper tackles the problem of high inference latency in deep convolutional neural networks by proposing a depth-wise decomposition approach based on SVD to expand regular convolutions into depth-wise separable convolutions, improving Top-1 accuracy of ShuffleNet V2 by ~2% while maintaining speed.
Very deep convolutional neural networks (CNNs) have been firmly established as the primary methods for many computer vision tasks. However, most state-of-the-art CNNs are large, which results in high inference latency. Recently, depth-wise separable convolution has been proposed for image recognition tasks on computationally limited platforms such as robotics and self-driving cars. Though it is much faster than its counterpart, regular convolution, accuracy is sacrificed. In this paper, we propose a novel decomposition approach based on SVD, namely depth-wise decomposition, for expanding regular convolutions into depthwise separable convolutions while maintaining high accuracy. We show our approach can be further generalized to the multi-channel and multi-layer cases, based on Generalized Singular Value Decomposition (GSVD) [59]. We conduct thorough experiments with the latest ShuffleNet V2 model [47] on both random synthesized dataset and a large-scale image recognition dataset: ImageNet [10]. Our approach outperforms channel decomposition [73] on all datasets. More importantly, our approach improves the Top-1 accuracy of ShuffleNet V2 by ~2%.