FBNA: A Fully Binarized Neural Network Accelerator

被引:58
|
作者
Guo, Peng [1 ,2 ]
Ma, Hong [1 ]
Chen, Ruizhi [1 ,2 ]
Li, Pin [1 ]
Xie, Shaolin [1 ]
Wang, Donglin [1 ]
机构
[1] Chinese Acad Sci, Inst Automat, Beijing, Peoples R China
[2] Univ Chinese Acad Sci, Sch Comp & Control Engn, Beijing, Peoples R China
来源
2018 28TH INTERNATIONAL CONFERENCE ON FIELD PROGRAMMABLE LOGIC AND APPLICATIONS (FPL) | 2018年
关键词
CNN; BNN; FPGA; Accelerator;
D O I
10.1109/FPL.2018.00016
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
In recent researches, binarized neural network (BNN) has been proposed to address the massive computations and large memory footprint problem of the convolutional neural network (CNN). Several works have designed specific BNN accelerators and showed very promising results. Nevertheless, only part of the neural network is binarized in their architecture and the benefits of binary operations were not fully exploited. In this work, we propose the first fully binarized convolutional neural network accelerator (FBNA) architecture, in which all convolutional operations are binarized and unified, even including the first layer and padding. The fully unified architecture provides more resource, parallelism and scalability optimization opportunities. Compared with the state-of-the-art BNN accelerator, our evaluation results show 3.1x performance, 5.4x resource efficiency and 4.9x power efficiency on CIFAR-10.
引用
收藏
页码:51 / 54
页数:4
相关论文
共 50 条
  • [21] Customizable FPGA-based Accelerator for Binarized Graph Neural Networks
    Wang, Ziwei
    Que, Zhiqiang
    Luk, Wayne
    Fan, Hongxiang
    Proceedings - IEEE International Symposium on Circuits and Systems, 2022, 2022-May : 1968 - 1972
  • [22] High-Performance and Robust Binarized Neural Network Accelerator Based on Modified Content-Addressable Memory
    Choi, Sureum
    Jeon, Youngjun
    Seo, Yeongkyo
    ELECTRONICS, 2022, 11 (17)
  • [23] Customizable FPGA-based Accelerator for Binarized Graph Neural Networks
    Wang, Ziwei
    Que, Zhiqiang
    Luk, Wayne
    Fan, Hongxiang
    2022 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS 22), 2022, : 1968 - 1972
  • [24] Content Addressable Memory Based Binarized Neural Network Accelerator Using Time-Domain Signal Processing
    Choi, Woong
    Jeong, Kwanghyo
    Choi, Kyungrak
    Lee, Kyeongho
    Park, Jongsun
    2018 55TH ACM/ESDA/IEEE DESIGN AUTOMATION CONFERENCE (DAC), 2018,
  • [25] On Class-Incremental Learning for Fully Binarized Convolutional Neural Networks
    Basso-Bert, Yanis
    Guiequero, William
    Molnos, Anca
    Lemaire, Romain
    Dupret, Antoine
    2024 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, ISCAS 2024, 2024,
  • [26] ECG signal classification with binarized convolutional neural network
    Wu, Qing
    Sun, Yangfan
    Yan, Hui
    Wu, Xundong
    COMPUTERS IN BIOLOGY AND MEDICINE, 2020, 121
  • [27] FCA-BNN: Flexible and Configurable Accelerator for Binarized Neural Networks on FPGA
    Gao, Jiabao
    Yao, Yuchen
    Li, Zhengjie
    Lai, Jinmei
    IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS, 2021, E104D (08) : 1367 - 1377
  • [28] BitFlow-Net: Toward Fully Binarized Convolutional Neural Networks
    Wu, Lijun
    Jiang, Peiqing
    Chen, Zhicong
    Lin, Xu
    Lai, Yunfeng
    Lin, Peijie
    Cheng, Shuying
    IEEE ACCESS, 2019, 7 : 154617 - 154626
  • [29] Efficient Super Resolution Using Binarized Neural Network
    Ma, Yinglan
    Xiong, Hongyu
    Hu, Zhe
    Ma, Lizhuang
    2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW 2019), 2019, : 611 - 620
  • [30] An Efficient Channel-Aware Sparse Binarized Neural Networks Inference Accelerator
    Liu, Qingliang
    Lai, Jinmei
    Gao, Jiabao
    IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS, 2022, 69 (03) : 1637 - 1641