CoChat

Training Quantized Neural Networks With a Full-Precision Auxiliary Module

Open access
Zhuang, Bohan, Liu, Lingqiao, Tan, Mingkui, Shen, Chunhua, Reid, Ian · 2020 · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Cited by 75

The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.

Abstract

In this paper, we seek to tackle a challenge in training low-precision networks: the notorious difficulty in propagating gradient through a low-precision network due to the non-differentiable quantization function. We propose a solution by training the low-precision network with a full-precision auxiliary module. Specifically, during training, we construct a mix-precision network by augmenting the original low-precision network with the full precision auxiliary module. Then the augmented mix-precision network and the low-precision network are jointly optimized. This strategy creates additional full-precision routes to update the parameters of the low-precision model, thus making the gradient back-propagates more easily. At the inference time, we discard the auxiliary module without introducing any computational complexity to the low-precision network. We evaluate the proposed method on image classification and object detection over various quantization approaches and show consistent performance increase. In particular, we achieve near lossless performance to the full-precision model by using a 4-bit detector, which is of great practical value.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/cvpr42600.2020.00156

Cites

23 of 23 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 23 references

Research works printed but not linked: 35

  1. redmon, 2018, YOLOv3 An Incremental Improvement
  2. 10.1109/CVPR.2017.690 · DOI · Paper page
  3. 10.1109/CVPR.2019.00934 · DOI · Paper page
  4. mishra, 0, Apprentice: Using knowledge distillation techniques to improve low-precision network accuracy, Proc Int Conf on Learning Rep
  5. louizos, 0, Relaxed quantization for discretized neural networks, Proc Int Conf on Learning Rep
  6. liu, 0, Bi-real net: Enhancing the performance of 1-bit cnns with improved representational capability and advanced training algorithm, Proc Eur Conf Comp Vis, 722
  7. pham, 0, Efficient neural architecture search via parameter sharing, Proc Int Conf Mach Learn, 4092
  8. 10.1109/CVPR.2019.00409 · DOI · Paper page
  9. liu, 0, Circulant binary convolutional networks: Enhancing the performance of 1-bit dcnns with circulant back propagation, Proc IEEE Conf Comp Vis Patt Recogn, 2691
  10. liu, 0, Ssd: Single shot multibox detector, Proc Eur Conf Comp Vis, 21
  11. lee, 2015, Deeply-supervised nets, Artificial Intelligence and Statistics, 562
  12. 10.1109/ICCV.2017.282 · DOI · Paper page
  13. lin, 0, Microsoft coco: Common objects in context, Proc Eur Conf Comp Vis, 740
  14. lin, 0, Focal loss for dense object detection, Proc IEEE Int Conf Comp Vis, 2980
  15. wu, 2019, Detectron 2
  16. 10.1109/CVPR.2019.00050 · DOI · Paper page
  17. 10.1109/CVPR.2018.00826 · DOI · Paper page
  18. zhuang, 2019, Effective training of convolutional neural networks with low-bitwidth weights and activations
  19. zhu, 0, Trained ternary quantization, Proc Int Conf on Learning Rep
  20. 10.1109/CVPR.2017.660 · DOI · Paper page
  21. zhang, 0, Lq-nets: Learned quantization for highly accurate and compact deep neural networks, Proc Eur Conf Comp Vis, 365
  22. 10.1109/ICCV.2017.322 · DOI · Paper page
  23. hinton, 0, Distilling the knowledge in a neural network, Proc Adv Neural Inf Process Syst Workshops
  24. hou, 0, Loss-aware binarization of deep networks, Proc Int Conf on Learning Rep
  25. 10.1007/978-3-030-01231-1_21 · DOI
  26. krizhevsky, 2009, Learning multiple layers of features from tiny images
  27. bethge, 2018, Training competitive binary neural networks from scratch
  28. chen, 2018, Joint neural architecture search and quantization
  29. chen, 0, Learning efficient object detection models with knowledge distillation, Proc Adv Neural Inf Process Syst, 742
  30. 10.1109/CVPR.2014.81 · DOI · Paper page
  31. 10.1109/ICCV.2015.169 · DOI · Paper page
  32. wang, 0, Haq: Hardware-aware automated quantization, Proc IEEE Conf Comp Vis Patt Recogn
  33. wei, 0, Quantization mimic: Towards very tiny cnn for object detection, Proc Eur Conf Comp Vis, 267
  34. 10.1007/s11263-015-0816-y · DOI · Paper page
  35. 10.1109/CVPR.2015.7298594 · DOI · Paper page

Cited by

See all 75

Metadata

Authors
Zhuang, B., Liu, L., Tan, M., Shen, C., & Reid, I.
Venue
2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher
IEEE
Published
2020-06
Pages
1485-1494
ISSN
1063-6919 (ISSN-L)
OpenAlex
W2991002474
Cited by
75
References
23
FWCI
4.70
Type
conference paper
Language
en
Open access
green

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Training Quantized Neural Networks With a Full-Precision Auxiliary Module. https://app.cochat.ai/papers/10.1109/cvpr42600.2020.00156
The paper
Zhuang, B., Liu, L., Tan, M., Shen, C., & Reid, I. (2020). Training Quantized Neural Networks With a Full-Precision Auxiliary Module. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). https://doi.org/10.1109/cvpr42600.2020.00156

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.