I've test OAK-D Pro PoE with tirandazi/depthai-yolov8-segment code.
I created a separate model that learns only two classes from the basic YOLOv8s_seg model, created onnx and blob files, and ran this model after modifying some shape and class information using the github code mentioned above, but the following error occurred.
```
PS C:\Users\great\OneDrive\Documents\ITman\OAK_D\depthai-yolov8-segment-main>
& C:/Users/great/anaconda3/envs/py310/python.exe c:/Users/great/OneDrive/Documents/ITman/OAK_D/depthai-yolov8-segment-main/main.py
Color cam resolution: (1280, 800)
Left cam resolution: (1280, 800)
Right cam resolution: (1280, 800)
[18443010E103F7F400] [192.168.0.102] [6.110] [ColorCamera(4)] [warning] Unsupported resolution set for detected camera IMX378/214, needs 1080_P / 4_K / 12_MP. Defaulting to 1080_P
[18443010E103F7F400] [192.168.0.102] [6.161] [NeuralNetwork(3)] [warning] Network compiled for 6 shaves, maximum available 10, compiling for 5 shaves likely will yield in better performance
[18443010E103F7F400] [192.168.0.102] [6.670] [StereoDepth(5)] [error] Stereo rectification error: 1, trying to recover.
[18443010E103F7F400] [192.168.0.102] [6.670] [StereoDepth(5)] [error] Stereo alignment error: 1, trying to recover.
[18443010E103F7F400] [192.168.0.102] [1732086905.791] [host] [warning] Monitor thread (device: 18443010E103F7F400 [192.168.0.102]) - ping was missed, closing the device connection
[2024-11-20 16:17:22.877] [depthai] [error] Device with id 18443010E103F7F400 has crashed. Crash dump logs are stored in: C:\Users\great\OneDrive\Documents\ITman\OAK_D\depthai-yolov8-segment-main\.cache\depthai\crashdumps\b4f6dd7b65d6361d04e08fb19ae9f6f2814ae31d\crash_dump.json - please report to developers.
So I use OAK-D Pro with usb-c type interface for comparison and I don't find any problem log at all.
I need to use PoE, how should I do it?
