Quick start to Edge-AI on Ultra96-V2

I wanted to create an easy guide for beginners using the Ultra96-V2. The guide is super-short and should take less than five minutes to set up.

Step1: Prepare the SD card

a. Get an SD card of at least 16Gb.

b. Format the SD card. Hint: I like to use SD Card Formatter (https://www.sdcardformatter.com/)

c. Etch the latest Pynq Image on the SD card. Tip: You can find it on the PYNQ official page (http://www.pynq.io/board.html)

d. Insert the SD card into the Ultra96-V2.

Step 2: Install PYNQ-DPU

a. Connect your Ultra96-V2 to power.

b. Connect your Ultra96-V2 to PC via Port J7.

c. Open a browser of your choice and type in 192.168.3.1. A Jupyter-notebook login page similar to the below image should pop up. Enter the password: <xilinx>.

d. Open the wifi.ipynb notebook found in the common folder.

e. Connect the Ultra96-V2 board to wifi. Ensure you enter the correct SSID and password.

f. Open a terminal and type the following commands sequentially.

sudo apt-get update
sudo apt-get install gnutls-bin
pip3 install pynq-dpu --no-build-isolation
cd $PYNQ_JUPYTER_NOTEBOOKS
pynq get-notebooks pynq-dpu -p .

Congratulations, you have completed the hardest part. You are now ready to carry out inference at the edge.

g. Confirm installation is correct by running the dpu_mnist_classifier.ipynb in the pynq-dpu folder.

Next, we would look at the inference of a custom model on the Ultra96-v2.