PYNQ: PYTHON PRODUCTIVITY FOR ZYNQ

Deep learning algorithm with Pynq-Z1

Hi, for academic purposes, I need to work with any simple deep learning algorithm and compare the throughput between the CPU (for instance, the code written in Python/Jupyter) and the CPU+FPGA deployment. For FPGA, I should accelerate part of the code using Vivado HLS.

I am completely new in machine learning and I know that there is plenty of information which I am already reading, but I think that experienced user could guide me in the right direction, with suitable examples and information, in short, how can I start to develop this task as quick as possible.

I am using Pynq-Z1 board and a Windows 7 laptop. Specific information would be really appreciated.

Hi, I believed that the best way to enter a new skill is to follow the previous work.

In fact, there is some similar work on Github, such as this and this.
Read and understand their code can really help your develop.

If you are a novice to FPGA, and is troubled with understanding these task, since you need numerous thing to do this job. I guess you can follow the “ZYNQ Book,” UG902 and UG585, these three documents helps me a lot when I first learn the development of HLS.