Quantization training with two-level bit width

  • Kang, Hansung
  • Lee, Yongjoo
  • Cho, Dongbin
  • Lee, Jaeyoung
  • Kang, Mincheal
  • 외 2명
Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

As the DNN model becomes more complex, the number of parameters constituting the model increases and requires a large amount of computation. Recently, a quantization technique that reduces the memory of the model and enables efficient computation has been studied. In this paper, we propose Fake Single Precision Training (FST) to increase accuracy by using a high bit range for weight and a low bit range for activation output with a certain probability. FST improved the accuracy of the model by applying the features of Google's Quantization Aware Training and FaceBook's Quant Noise method.

키워드

Fake Single Precision TrainingQaunt NoiseQuantization Aware TrainingBit-WidthEfficient computationFacebookFake single precision trainingGoogle+Large amountsQaunt noiseQuantisationQuantization aware trainingSingle precision
제목
Quantization training with two-level bit width
저자
Kang, HansungLee, YongjooCho, DongbinLee, JaeyoungKang, MinchealKim, YounghoonSeo, Jiwon
DOI
10.1109/ICEIC54506.2022.9748737
발행일
2022-02
유형
Proceedings Paper
저널명
2022 INTERNATIONAL CONFERENCE ON ELECTRONICS, INFORMATION, AND COMMUNICATION (ICEIC)
페이지
1 ~ 4