Skip to main content

SPIN: An Empirical Evaluation on Sharing Parameters of Isotropic Networks

  • Conference paper
  • First Online:
Computer Vision – ECCV 2022 (ECCV 2022)

Part of the book series: Lecture Notes in Computer Science ((LNCS,volume 13671))

Included in the following conference series:

Abstract

Recent isotropic networks, such as ConvMixer and Vision Transformers, have found significant success across visual recognition tasks, matching or outperforming non-isotropic Convolutional Neural Networks. Isotropic architectures are particularly well-suited to cross-layer weight sharing, an effective neural network compression technique. In this paper, we perform an empirical evaluation on methods for sharing parameters in isotropic networks (SPIN). We present a framework to formalize major weight sharing design decisions and perform a comprehensive empirical evaluation of this design space. Guided by our experimental results, we propose a weight sharing strategy to generate a family of models with better overall efficiency, in terms of FLOPs and parameters versus accuracy, compared to traditional scaling methods alone, for example compressing ConvMixer by \(1.9\times \) while improving accuracy on ImageNet. Finally, we perform a qualitative study to further understand the behavior of weight sharing in isotropic architectures. The code is available at https://github.com/apple/ml-spin.

C-Y. Lin and A. Prabhu—Equal contribution.

C.-Y. Lin—Work done while interning at Apple.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 89.00
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 119.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

Notes

  1. 1.

    We interchangeably use the terms parameter and weight sharing throughout this paper.

References

  1. Battash, B., Wolf, L.: Adaptive and iteratively improving recurrent lateral connections. CoRR abs/1910.11105 (2019). http://arxiv.org/abs/1910.11105

  2. Dehghani, M., Gouws, S., Vinyals, O., Uszkoreit, J., Kaiser, L.: Universal transformers. ArXiv abs/1807.03819 (2019)

    Google Scholar 

  3. Dosovitskiy, A., et al.: An image is worth 16x16 words: transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

  4. Guo, Q., Yu, Z., Wu, Y., Liang, D., Qin, H., Yan, J.: Dynamic recursive neural network. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 5142–5151 (2019)

    Google Scholar 

  5. Han, S., Pool, J., Tran, J., Dally, W.J.: Learning both weights and connections for efficient neural network. ArXiv abs/1506.02626 (2015)

    Google Scholar 

  6. He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition (2015)

    Google Scholar 

  7. Howard, A.G., et al.: Searching for mobilenetv3. In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 1314–1324 (2019)

    Google Scholar 

  8. Jastrzebski, S., Arpit, D., Ballas, N., Verma, V., Che, T., Bengio, Y.: Residual connections encourage iterative inference. CoRR abs/1710.04773 (2017). http://arxiv.org/abs/1710.04773

  9. Kim, D., Kang, W.: Learning shared filter bases for efficient convnets. CoRR abs/2006.05066 (2020). https://arxiv.org/abs/2006.05066

  10. Kornblith, S., Norouzi, M., Lee, H., Hinton, G.E.: Similarity of neural network representations revisited. ArXiv abs/1905.00414 (2019)

    Google Scholar 

  11. Kubilius, J., et al.: Aligning artificial neural networks to the brain yields shallow recurrent architectures (2018)

    Google Scholar 

  12. Kubilius, J., et al.: Brain-like object recognition with high-performing shallow recurrent ANNs. CoRR abs/1909.06161 (2019). http://arxiv.org/abs/1909.06161

  13. Lan, Z., Chen, M., Goodman, S., Gimpel, K., Sharma, P., Soricut, R.: Albert: a lite BERT for self-supervised learning of language representations. In: ICLR. OpenReview.net (2020). http://dblp.uni-trier.de/db/conf/iclr/iclr2020.htmlLanCGGSS20

  14. Leroux, S., Molchanov, P., Simoens, P., Dhoedt, B., Breuel, T.M., Kautz, J.: IamNN: iterative and adaptive mobile neural network for efficient image classification. CoRR abs/1804.10123 (2018). http://arxiv.org/abs/1804.10123

  15. Liu, Z., Mao, H., Wu, C.Y., Feichtenhofer, C., Darrell, T., Xie, S.: A convnet for the 2020s. arXiv preprint arXiv:2201.03545 (2022)

  16. Rastegari, M., Ordonez, V., Redmon, J., Farhadi, A.: XNOR-Net: ImageNet classification using binary convolutional neural networks. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9908, pp. 525–542. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46493-0_32

    Chapter  Google Scholar 

  17. Sandler, M., Howard, A.G., Zhu, M., Zhmoginov, A., Chen, L.C.: MobileNetV 2: inverted residuals and linear bottlenecks. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4510–4520 (2018)

    Google Scholar 

  18. Shen, Z., Liu, Z., Xing, E.P.: Sliced recursive transformer. CoRR abs/2111.05297 (2021). https://arxiv.org/abs/2111.05297

  19. Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition (2015)

    Google Scholar 

  20. Spoerer, C.J., Kietzmann, T.C., Mehrer, J., Charest, I., Kriegeskorte, N.: Recurrent networks can recycle neural resources to flexibly trade speed for accuracy in visual recognition. bioRxiv (2020). https://doi.org/10.1101/677237, https://www.biorxiv.org/content/early/2020/03/26/677237

  21. Takase, S., Kiyono, S.: Lessons on parameter sharing across layers in transformers. CoRR abs/2104.06022 (2021). https://arxiv.org/abs/2104.06022

  22. Tan, M., Chen, B., Pang, R., Vasudevan, V., Le, Q.V.: MnasNet: platform-aware neural architecture search for mobile. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2815–2823 (2019)

    Google Scholar 

  23. Tan, M., Le, Q.: EfficientNet: rethinking model scaling for convolutional neural networks. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning. Proceedings of Machine Learning Research, vol. 97, pp. 6105–6114. PMLR (2019)

    Google Scholar 

  24. Tolstikhin, I.O., et al.: MLP-mixer: An all-MLP architecture for vision. CoRR abs/2105.01601 (2021). https://arxiv.org/abs/2105.01601

  25. Trockman, A., Kolter, J.Z.: Patches are all you need? CoRR abs/2201.09792 (2022). https://arxiv.org/abs/2201.09792

  26. Wang, G., Zhao, Y., Tang, C., Luo, C., Zeng, W.: When shift operation meets vision transformer: an extremely simple alternative to attention mechanism. CoRR abs/2201.10801 (2022). https://arxiv.org/abs/2201.10801

  27. Zhai, S., Talbott, W., Srivastava, N., Huang, C., Goh, H., Zhang, R., Susskind, J.M.: An attention free transformer. CoRR abs/2105.14103 (2021). https://arxiv.org/abs/2105.14103

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Chien-Yu Lin .

Editor information

Editors and Affiliations

1 Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (pdf 1482 KB)

Rights and permissions

Reprints and permissions

Copyright information

© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Lin, CY. et al. (2022). SPIN: An Empirical Evaluation on Sharing Parameters of Isotropic Networks. In: Avidan, S., Brostow, G., Cissé, M., Farinella, G.M., Hassner, T. (eds) Computer Vision – ECCV 2022. ECCV 2022. Lecture Notes in Computer Science, vol 13671. Springer, Cham. https://doi.org/10.1007/978-3-031-20083-0_33

Download citation

  • DOI: https://doi.org/10.1007/978-3-031-20083-0_33

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-031-20082-3

  • Online ISBN: 978-3-031-20083-0

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics