Generative Low-bitwidth Data Free Quantization

Shoukai Xu, Haokun Li, Bohan Zhuang, Jing Liu, Jiezhang Cao, Chuangrun Liang, Mingkui Tan

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

Abstract

Neural network quantization is an effective way to compress deep models and improve their execution latency and energy efficiency, so that they can be deployed on mobile or embedded devices. Existing quantization methods require original data for calibration or fine-tuning to get better performance. However, in many real-world scenarios, the data may not be available due to confidential or private issues, thereby making existing quantization methods not applicable. Moreover, due to the absence of original data, the recently developed generative adversarial networks (GANs) cannot be applied to generate data. Although the full-precision model may contain rich data information, such information alone is hard to exploit for recovering the original data or generating new meaningful data. In this paper, we investigate a simple-yet-effective method called Generative Low-bitwidth Data Free Quantization (GDFQ) to remove the data dependence burden. Specifically, we propose a knowledge matching generator to produce meaningful fake data by exploiting classification boundary knowledge and distribution information in the pre-trained model. With the help of generated data, we can quantize a model by learning knowledge from the pre-trained model. Extensive experiments on three data sets demonstrate the effectiveness of our method. More critically, our method achieves much higher accuracy on 4-bit quantization than the existing data free quantization method. Code is available at https://github.com/xushoukai/GDFQ.

Original languageEnglish
Title of host publicationComputer Vision – ECCV 2020
Subtitle of host publication16th European Conference Glasgow, UK, August 23–28, 2020 Proceedings, Part XII 123
EditorsAndrea Vedaldi, Horst Bischof, Thomas Brox, Jan-Michael Frahm
Place of PublicationCham Switzerland
PublisherSpringer
Pages1-17
Number of pages17
ISBN (Electronic)9783030586102
ISBN (Print)9783030586096
DOIs
Publication statusPublished - 2020
EventEuropean Conference on Computer Vision 2020 - Glasgow, United Kingdom
Duration: 23 Aug 202028 Aug 2020
Conference number: 16th
https://link.springer.com/book/10.1007/978-3-030-58452-8 (Proceedings)
https://eccv2020.eu (Website)

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
Volume12357
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Computer Vision 2020
Abbreviated titleECCV 2020
CountryUnited Kingdom
CityGlasgow
Period23/08/2028/08/20
Internet address

Keywords

  • Data free compression
  • Knowledge matching generator
  • Low-bitwidth quantization

Cite this