Fune: An FPGA Tuning Framework for CNN Acceleration
Qingcheng Xiao, Yun Liang · IEEE Design and Test · 2019
This article describes methodologies and tools for the customization of convolutional neural networks (CNNs) using fine-grained reconfiguration of FPGAs. Tuning uses dynamic programming to identify reconfiguration opportunities as well as a search algorithm for configurations.