English

LANCE: Efficient Low-Precision Quantized Winograd Convolution for Neural Networks Based on Graphics Processing Units

Computer Vision and Pattern Recognition 2020-07-29 v3 Machine Learning Neural and Evolutionary Computing

Abstract

Accelerating deep convolutional neural networks has become an active topic and sparked an interest in academia and industry. In this paper, we propose an efficient low-precision quantized Winograd convolution algorithm, called LANCE, which combines the advantages of fast convolution and quantization techniques. By embedding linear quantization operations into the Winograd-domain, the fast convolution can be performed efficiently under low-precision computation on graphics processing units. We test neural network models with LANCE on representative image classification datasets, including SVHN, CIFAR, and ImageNet. The experimental results show that our 8-bit quantized Winograd convolution improves the performance by up to 2.40x over the full-precision convolution with trivial accuracy loss.

Keywords

Cite

@article{arxiv.2003.08646,
  title  = {LANCE: Efficient Low-Precision Quantized Winograd Convolution for Neural Networks Based on Graphics Processing Units},
  author = {Guangli Li and Lei Liu and Xueying Wang and Xiu Ma and Xiaobing Feng},
  journal= {arXiv preprint arXiv:2003.08646},
  year   = {2020}
}

Comments

Accepted by ICASSP 2020

R2 v1 2026-06-23T14:19:48.072Z