ShiftCNN: Generalized Low-Precision Architecture for Inference of Convolutional Neural Networks

Abstract

In this paper we introduce ShiftCNN, a generalized low-precision architecture for inference of multiplierless convolutional neural networks (CNNs). ShiftCNN is based on a power-of-two weight representation and, as a result, performs only shift and addition operations. Furthermore, ShiftCNN substantially reduces computational cost of convolutional layers by precomputing convolution terms. Such an optimization can be applied to any CNN architecture with a relatively small codebook of weights and allows to decrease the number of product operations by at least two orders of magnitude. The proposed architecture targets custom inference accelerators and can be realized on FPGAs or ASICs. Extensive evaluation on ImageNet shows that the state-of-the-art CNNs can be converted without retraining into ShiftCNN with less than 1% drop in accuracy when the proposed quantization algorithm is employed. RTL simulations, targeting modern FPGAs, show that power consumption of convolutional layers is reduced by a factor of 4 compared to conventional 8-bit fixed-point architectures.

References in corpus (10)

• Caffe: Convolutional Architecture for Fast Feature Embedding
• Fully Convolutional Networks for Semantic Segmentation
• Ternary Weight Networks
• Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weights
• Convolutional Neural Networks using Logarithmic Data Representation
• Training deep neural networks with low precision multiplications
• XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
• Hardware-oriented Approximation of Convolutional Neural Networks
• Fast Algorithms for Convolutional Neural Networks
• LCNN: Lookup-based Convolutional Neural Network

Cited by in corpus (10)

• A Survey on Methods and Theories of Quantized Neural Networks
• Hardware Acceleration of Sparse and Irregular Tensor Computations of ML Models: A Survey and Insights
• Deep Neural Network Approximation for Custom Hardware: Where We've Been, Where We're Going
• Compacting Deep Neural Networks for Internet of Things: Methods and Applications
• MWQ: Multiscale Wavelet Quantized Neural Networks
• Fully Learnable Group Convolution for Acceleration of Deep Neural Networks
• Recent Advances in Efficient Computation of Deep Convolutional Neural Networks
• Training a Binary Weight Object Detector by Knowledge Transfer for Autonomous Driving
• A System-Level Solution for Low-Power Object Detection
• $S^3$: Sign-Sparse-Shift Reparametrization for Effective Training of Low-bit Shift Networks