User Guide 4: Simplifying Instance Segmentation on a Hailo Device Using DeGirum PySDK

Guides
21

Thanks @lawrence,

Can I look at the har file and can you show me how to compile or a post that talked about this before?

I also tried compiling the yolov8n-seg model on my computer for about 6 hours, although in the onnx file it is not less accurate than the yolov11n-seg model that you have compiled, but when I run it it is not as good.

I don’t know if it’s because my CPU is weaker, can you give me some advice?

22

Hi @shashi im having a problem over here on infering yolov8n_seg model

image
image989×350 31.6 KB

this is my model zoo files and im sure im following you step but not sure why this causes

Blockquote
{
“ConfigVersion”: 10,
“DEVICE”: [
{
“DeviceType”: “HAILO8L”,
“RuntimeAgent”: “HAILORT”,
“SupportedDeviceTypes”: “HAILORT/HAILO8L,HAILORT/HAILO8”
}
],
“PRE_PROCESS”: [
{
“InputType”: “Image”,
“InputN”: 1,
“InputH”: 640,
“InputW”: 640,
“InputC”: 3,
“InputPadMethod”: “letterbox”,
“InputResizeMethod”: “bilinear”,
“InputQuantEn”: true
}
],
“MODEL_PARAMETERS”: [
{
“ModelPath”: “yolov8n_seg.hef”
}
],
“POST_PROCESS”: [
{
“OutputPostprocessType”: “SegmentationYoloV8”,
“LabelsPath”: “labels_coco.json”,
“OutputNumClasses”: 80,
“OutputConfThreshold”: 0.3,
“SigmoidOnCLS”: true
}
]
}

and the model i downloaded from the public model repo

https://github.com/hailo-ai/hailo_model_zoo/blob/master/docs/public_models/HAILO8L/HAILO8L_instance_segmentation.rst

23

okkay i dont know why after add checksum the infering works fine. Anyways thanks for the sdk support for hailo

24

Hi @SAN
Glad to see it working for you. The checksum is a field we use in our model zoos to track if the model got updated. While it makes sense only for AI hub model zoos, in our design it is a required parameter.

25

How did you obtain your yolo11nseg.hef? Could you provide the relevant tutorial link?