Abstract
Compression and acceleration of convolutional neural network (CNN) have raised extensive research interest in the past few years. In this paper, we proposed a novel channel-level pruning method based on gamma (scaling parameters) of Batch Normalization layer to compress and accelerate CNN models. Local gamma normalization and selection was proposed to address the over-pruning issue and introduce local information into channel selection. After that, an ablation based beta (shifting parameters) transfer, and knowledge distillation based fine-tuning were further applied to improve the performance of the pruned model. The experimental results on CIFAR-10, CIFAR-100 and LFW datasets suggest that our approach can achieve much more efficient pruning in terms of reduction of parameters and FLOPs, e.g., \(8.64\times \) compression and \(3.79\times \) acceleration of VGG were achieved on CIFAR, with slight accuracy loss.
Keywords
The work is supported by National Natural Science Foundation of China (Grant No. 61672357 and U1713214), and the Science and Technology Project of Guangdong Province (Grant No. 2018A050501014).
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. arXiv preprint: arXiv:1409.1556 (2014)
Szegedy, C., et al.: Going deeper with convolutions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1–9 (2015)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778 (2016)
Huang, G., Liu, Z., Van Der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: CVPR, vol. 1, p. 3 (2017)
Denil, M., Shakibi, B., Dinh, L., De Freitas, N., et al.: Predicting parameters in deep learning. In: Advances in Neural Information Processing Systems, pp. 2148–2156 (2013)
Liu, Z., Li, J., Shen, Z., Huang, G., Yan, S., Zhang, C.: Learning efficient convolutional networks through network slimming. In: 2017 IEEE International Conference on Computer Vision (ICCV), pp. 2755–2763. IEEE (2017)
Ba, J., Caruana, R.: Do deep nets really need to be deep? In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Weinberger, K.Q. (eds.) Advances in Neural Information Processing Systems 27, pp. 2654–2662. Curran Associates, Inc. (2014)
Hinton, G., Vinyals, O., Dean, J.: Distilling the knowledge in a neural network. arXiv preprint: arXiv:1503.02531 (2015)
Luo, P., Zhu, Z., Liu, Z., Wang, X., Tang, X., et al.: Face model compression by distilling knowledge from neurons. In: AAAI, pp. 3560–3566 (2016)
Mirzadeh, S.I., Farajtabar, M., Li, A., Ghasemzadeh, H.: Improved knowledge distillation via teacher assistant: bridging the gap between student and teacher. arXiv preprint: arXiv:1902.03393 (2019)
LeCun, Y., Denker, J.S., Solla, S.A.: Optimal brain damage. In: Advances in Neural Information Processing Systems, pp. 598–605 (1990)
Hassibi, B., Stork, D.: Second order derivaties for network prunning: optimal brain surgeon. In: Advances in NIPS 5, pp. 164–171 (1993)
Han, S., Pool, J., Tran, J., Dally, W.: Learning both weights and connections for efficient neural network. In: Advances in Neural Information Processing Systems, pp. 1135–1143 (2015)
Han, S., Mao, H., Dally, W.J.: Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding. arXiv preprint: arXiv:1510.00149 (2015)
Li, H., Kadav, A., Durdanovic, I., Samet, H., Graf, H.P.: Pruning filters for efficient convnets. arXiv preprint: arXiv:1608.08710 (2016)
He, Y., Zhang, X., Sun, J.: Channel pruning for accelerating very deep neural networks. In: International Conference on Computer Vision (ICCV), vol. 2 (2017)
Luo, J.H., Wu, J., Lin, W.: Thinet: A filter level pruning method for deep neural network compression. arXiv preprint: arXiv:1707.06342 (2017)
Ye, J., Lu, X., Lin, Z., Wang, J.Z.: Rethinking the smaller-norm-less-informative assumption in channel pruning of convolution layers. arXiv preprint: arXiv:1802.00124 (2018)
Ioffe, S., Szegedy, C.: Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint: arXiv:1502.03167 (2015)
foolwood: pytorch-slimming (2018). https://github.com/foolwood/pytorch-slimming
szagoruyko: cifar.torch (2014). https://github.com/szagoruyko/cifar.torch
He, K., Zhang, X., Ren, S., Sun, J.: Identity mappings in deep residual networks. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016, Part IV. LNCS, vol. 9908, pp. 630–645. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46493-0_38
Liu, W., Zhang, Y.M., Li, X., Yu, Z., Dai, B., Zhao, T., Song, L.: Deep hyperspherical learning. In: Advances in Neural Information Processing Systems, pp. 3950–3960 (2017)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Liu, Y., Jia, X., Shen, L., Ming, Z., Duan, J. (2019). Local Normalization Based BN Layer Pruning. In: Tetko, I., Kůrková, V., Karpov, P., Theis, F. (eds) Artificial Neural Networks and Machine Learning – ICANN 2019: Deep Learning. ICANN 2019. Lecture Notes in Computer Science(), vol 11728. Springer, Cham. https://doi.org/10.1007/978-3-030-30484-3_28
Download citation
DOI: https://doi.org/10.1007/978-3-030-30484-3_28
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-30483-6
Online ISBN: 978-3-030-30484-3
eBook Packages: Computer ScienceComputer Science (R0)