.png){kind=icon}
.png){kind=icon}
Return
Dynamic Dataflow Scheduling and Computation Mapping Techniques for Efficient Depthwise Separable Convolution Acceleration
.png){kind=icon}
PRE
AI
DOI：10.1109/TCSI.2021.3078541.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Depthwise separable convolution (DSC) has become one of the essential structures for lightweight convolutional neural networks. Nevertheless, its hardware architecture has not received much attention. Several previous hardware designs incur either high off-chip memory traffic or large on-chip memory usage, and hence have deficiency in terms of hardware efficiency as well as performance. This paper proposes two efficient dynamic design techniques, i.e. adaptive row-based dataflow scheduling and adaptive computation mapping, to achieve a much better trade-off between hardware efficiency and performance for DSC-based lightweight CNN accelerator. The effectiveness and efficiency of the proposed dynamic design techniques have been extensively evaluated using six DSC-based lightweight CNNs. Compared with the reference architectures, the simulation results show the proposed architectural techniques can at least reduce on-chip buffer size by 50.4% and improve the performance of convolution calculation by 1.18x while maintaining the minimum off-chip memory traffic. MobileNetV2 is implemented on Zynq UltraScale+ ZCU102 SoC FPGA, and the results show the proposed accelerator can achieve 381.7 frames per second (fps), which is 1.43x of the reference design, and it can save about 36.3% on-chip buffer size compared with the reference design, while maintaining the same off-chip memory traffic.
Keywords:
Convolutional neural network
depthwise separable convolution
adaptive row-based dataflow scheduling
adaptive computation mapping
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
5.2
Papers:
9.7K
Citations:
2.2W