PYNQ-TinyYoloV3 ON THE PYNQ

Today,You can use python to develop FPGA,it is PYNQ.

The brief specification of PYNQ Z2 is as follows, with some parameters I am concerned about bolded:

ZYNQ XC7Z020-1CLG400C • 650MHz dual-core Cortex-A9 processor • DDR3 memory controller with 8 DMA channels and 4 High Performance AXI3 Slave ports • High-bandwidth peripheral controllers: 1G Ethernet,  　　USB 2.0, SDIO • Low-bandwidth peripheral controller: SPI, UART, CAN, I2C • Programmable from JTAG, Quad-SPI flash, and MicroSD card • Programmable logic equivalent to Artix-7 FPGA • 13,300 logic slices, each with four 6-input LUTs and 8 flip-flops • 630 KB of fast block RAM • 4 clock management tiles, each with a phase locked loop (PLL) and mixed-mode clock manager (MMCM) • 220 DSP slices • On-chip analog-to-digital converter (XADC) Memory • 512MB DDR3 with 16-bit bus @ 1050Mbps • 16MB Quad-SPI Flash with factory programmed 48-bit globally unique EUI-48/64™ compatible identifier • MicroSD slot Power • Powered from USB or 7V-15V external power source

USB and Ethernet • Gigabit Ethernet PHY • Micro USB-JTAG Programming circuitry • Micro USB-UART bridge • USB 2.0 OTG PHY (supports host only) Audio and Video • HDMI sink port (input) • HDMI source port (output) • I2S interface with 24bit DAC with 3.5mm TRRS jack • Line-in with 3.5mm jack Switches, Push-buttons and LEDs • 4 push-buttons • 2 slide switches • 4 LEDs • 2 RGB LEDs Expansion Connectors • Two standard Pmod ports • 16 Total FPGA I/O (8 shared pins with  　　　Raspberry Pi connector) • Arduino Shield connector • 24 Total FPGA I/O • 6 Single-ended 0-3.3V Analog inputs to XADC • Raspberry Pi connector • 28 Total FPGA I/O (8 shared pins with Pmod A port)

About  ZYNQ and PYNQ ,you can refer to this"http://www.zynqbook.com/"

Now,i will introduce PYNQ,you can find the official website and documentation of the project can be found at" http://www.pynq.io/".you can use the Jupyter notebooks to develop PYNQ.

• PYNQ provides a Python interface that allows you to control overlays in PL through Python running in PS.
• FPGA design is a professional task that requires hardware engineering knowledge and professional knowledge.
• PYNQ's Overlay layer is created by hardware designers and is included in this PYNQ Python API.
• Software developers can then use Python interfaces to program and control proprietary hardware overlays without having to design their own.

•  This paper briefly introduces the principle of Tiny YOLO target detection algorithm, and finally records the experimental results.

started

• 

 (1)According to the official guide, first set the jump cap on the board, burn and write the TF card image

(2)Power on

(3)Connect PYNQ to the LAN

(4)PYNQ hostname:PYNQ,you can input the http://PYNQ ,you can use Jupyter notebook

(5)you can refer to the docment,https://pynq.readthedocs.io

(6)download QNN-PYNQ :https://github.com/Xilinx/QNN-MO-PYNQ

(7)install QNN;

The first method:sudo pip3 install git+http://github.com/Xilinx/QNN-MO-PYNQ.git 

The Second method:

https://github.com/Xilinx/QNN-MO-PYNQ,Then,Create a Shared folder，Copy the *.zip to the qnn folder.

python setup.py build and python setup.py install.

(8)open tiny-yolo-image-loop.ipynb,you can see the code:

(9)test

the result :

More refer to:

人工智能
BNN：

https://github.com/Xilinx/BNN-PYNQ/

QNN：
https://github.com/Xilinx/QNN-MO-PYNQ

LSTM：
https://github.com/tukl-msd/LSTM-PYNQ

CNN：
https://github.com/awai54st/PYNQ-Classification

ZIPML：
https://github.com/fpgasystems/ZipML-PYNQ

iSmart：
https://github.com/onioncc/iSmartDNN

spooNN：
https://github.com/fpgasystems/spooNN

accelerator：
https://mp.weixin.qq.com/s/ic2OcbAto_daylAVh7-0HQ

物联网
PYNQ-BOT：
https://github.com/Xilinx/PYNQ-BOT

SPYN-IIOT：
https://github.com/Xilinx/IIoT-SPYN

机器视觉
部分可重配：
https://github.com/AEW2015/PYNQ_PR_Overlay

软GPU：
https://github.com/malkadi/FGPU_IPython/

Video Processing：
https://github.com/Pieter-Berteloot/PYNQ_Projects

板卡移植
参考资料：

Zynq-7000移植指南：
https://gist.github.com/PeterOgden/9a5054a06408d2bd711d6de563281930

Zedboard移植参考：
https://superuser.blog/pynq-linux-on-zedboard/

官方文档：
https://pynq.readthedocs.io/en/latest/pynq_sd_card.html#retargeting-to-a-different-board

Minized移植参考：
https://github.com/ticktronaut/PYNQ

UltraZed移植参考：
https://github.com/tchoyt/pynq_dev

Pynq部分补充教程
HLS filter example：
https://github.com/drichmond/PYNQ-HLS/blob/master/tutorial/notebooks/stream/0-How-To-HLS.ipynb

Custom HLS adder IP：
http://www.fpgadeveloper.com/2018/03/create-a-custom-pynq-overlay-for-pynq-z1.html

Accelerate FIR software function：
http://www.fpgadeveloper.com/2018/03/how-to-accelerate-a-python-function-with-pynq.html

Add existing IP to a PYNQ overlay：
https://www.youtube.com/watch?v=UBsCNPWudww

Control custom IP using GPIO：
https://www.youtube.com/watch?v=LomArt-hi4M

the source code:updating.