EUVIMEDEUVIMED entdecken
Crossref

UNIQ

Chaim Baskin, Natan Liss, Eli Schwartz, Evgenii Zheltonozhskii, Raja Giryes, Alex M. Bronstein, Avi Mendelson

ACM Transactions on Computer Systems · 30.11.2019 · 37 (1-4) · 1-15

DOI: 10.1145/3444943 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Computer Systems, online erschienen am 26.03.2021. Verfasst von Chaim Baskin und 6 weiteren Autor:innen. Die Arbeit wurde bisher 92-mal zitiert (Crossref) und nennt 35 Literaturangaben.

AbstractBibliografische AngabenZitieren

Abstract (Originalsprache)

We present a novel method for neural network quantization. Our method, named UNIQ , emulates a non-uniform k -quantile quantizer and adapts the model to perform well with quantized weights by injecting noise to the weights at training time. As a by-product of injecting noise to weights, we find that activations can also be quantized to as low as 8-bit with only a minor accuracy degradation. Our non-uniform quantization approach provides a novel alternative to the existing uniform quantization techniques for neural networks. We further propose a novel complexity metric of number of bit operations performed (BOPs), and we show that this metric has a linear relation with logic utilization and power. We suggest evaluating the trade-off of accuracy vs. complexity (BOPs). The proposed method, when evaluated on ResNet18/34/50 and MobileNet on ImageNet, outperforms the prior state of the art both in the low-complexity regime and the high accuracy regime. We demonstrate the practical applicability of this approach, by implementing our non-uniformly quantized CNN on FPGA.

Zitieren

APA

Baskin, C., Liss, N., Schwartz, E., Zheltonozhskii, E., Giryes, R., Bronstein, A. M., & Mendelson, A. (2019). UNIQ. ACM Transactions on Computer Systems, 37(1-4), 1-15. https://doi.org/10.1145/3444943

Vancouver

Baskin C, Liss N, Schwartz E, Zheltonozhskii E, Giryes R, Bronstein AM, et al. UNIQ. ACM Transactions on Computer Systems. 2019;37(1-4):1-15. doi:10.1145/3444943

BibTeX
@article{Baskin2019,
  author = {Baskin, Chaim and Liss, Natan and Schwartz, Eli and Zheltonozhskii, Evgenii and Giryes, Raja and Bronstein, Alex M. and Mendelson, Avi},
  title = {UNIQ},
  journal = {ACM Transactions on Computer Systems},
  year = {2019},
  volume = {37},
  number = {1-4},
  pages = {1--15},
  doi = {10.1145/3444943}
}
Kostenlos starten →