Running a custom U-Net network on the Raspberry Pi 5 AI Kit.

General
,
1

Hello everyone!

I currently have a U-Net network, and I have made some modifications based on the original U-Net network (as shown in the figure below).

I’m just changing some convolution algorithms, maybe pruning, reducing the number of layers, and of course changing the size of the input and middle layers，The entire project is built on Pytorch.

unet_v2.onnx
unet_v2.onnx1920×152 25.2 KB

I have already trained the model and converted it to the .hef format on DFC (only using Quick Optimization in the optimization part).

success
success954×555 172 KB

Now I want to deploy this network on the Raspberry Pi 5 AI Kit (rspi5 hailo8 26TOPS). I have set up the environment on my Raspberry Pi 5 according to the tutorial on GitHub and can run the Yolo example.

QQ20250306-202141
QQ20250306-2021411461×816 89.7 KB

As for running my own U-Net network, I have no idea at all and don’t know where to start. I would really appreciate it if someone could give me some guidance or relevant examples. :pleading_face:

2

Can someone give me an answer? Thank you very much :blush:

1 Like
3

You should use the PySDK, it’s easy to inference your Unet.

1 Like
4

Thank you very much. I will try

5

@user134
Thanks for suggesting PySDK.
@zhao_yj
Please let us know if you encounter any difficulties.