PyHailoRT on raspberry pi AI kit

General
,
1

Hi, I am new to the Raspberry Pi AI Kit and the Hailo platform. I successfully converted my custom YOLOv11 model to the HEF format and ran it on raspberry pi OS. Now, I want to write my own Python code to process given image arrays and perform some actions based on the results.

From what I’ve found, it seems I can do most of what I want using this documentation. However, I’m not sure whether I can use PyHailoRT like this on the Raspberry Pi, since the HailoRT documentation only includes installation instructions for standard Ubuntu and Hailo have another documentation for specifically Raspberry Pi AI kit. So, I’m uncertain whether it’s fully supported on the Raspberry Pi AI Kit.

2

Hi @Burak_KCN
Welcome to the Hailo community. You can use DeGirum PySDK, a python package designed to simplify such development. You can find instructions and examples at: DeGirum/hailo_examples: DeGirum PySDK with Hailo AI Accelerators. For your specific use case, you can see our guide at: User Guide 3: Simplifying Object Detection on a Hailo Device Using DeGirum PySDK

3

Hello @shashi ,
Thanks for the instructions. I successfully used them and was able to run my custom model using DeGirum.

Now, I’m wondering about the best practices for working with video input. Should I use the same code that I use for processing a single image, or is there a different recommended approach for video?

Also, I’m currently trying to implement an RT-DETR model that I trained using the Ultralytics library. I couldn’t find any specific instructions on how to integrate it. Could you please guide me on this?

4

Hi @Burak_KCN
Working with videos is also straightforward. See hailo_examples/examples/001_quick_start.ipynb at main · DeGirum/hailo_examples for an example. Regarding Rt-DETR model, we have not compiled it for Hailo yet. If there is a compiled model, integrating it is straightforward. The instructions are same as in the above guide.