Redundancy-Reduced MobileNet Acceleration on Reconfigurable Logic for ImageNet Classification

Autor: Peter Y. K. Cheung, Julian Faraone, David B. Thomas, Yiren Zhao, Philip H. W. Leong, Junyi Liu, Jiang Su
Rok vydání: 2018
Předmět:
Zdroj: Applied Reconfigurable Computing. Architectures, Tools, and Applications ISBN: 9783319788890
ARC
DOI: 10.1007/978-3-319-78890-6_2
Popis: Modern Convolutional Neural Networks (CNNs) excel in image classification and recognition applications on large-scale datasets such as ImageNet, compared to many conventional feature-based computer vision algorithms. However, the high computational complexity of CNN models can lead to low system performance in power-efficient applications. In this work, we firstly highlight two levels of model redundancy which widely exist in modern CNNs. Additionally, we use MobileNet as a design example and propose an efficient system design for a Redundancy-Reduced MobileNet (RR-MobileNet) in which off-chip memory traffic is only used for inputs/outputs transfer while parameters and intermediate values are saved in on-chip BRAM blocks. Compared to AlexNet, our RR-mobileNet has 25\(\times \) less parameters, 3.2\(\times \) less operations per image inference but 9%/5.2% higher Top1/Top5 classification accuracy on ImageNet classification task. The latency of a single image inference is only 7.85 ms.
Databáze: OpenAIRE