YOLOv5s and YOLOv7-tiny

[fizzgo]

l want to run my own YOLOv5s and YOLOv7-tiny model on Raspberry PI 4B with NCS2.But I can't find the right tutorial on the Internet, so I ask for your help. Thanks.

[Hairul_Intel]

Hi Fizzgo,

Thank you for reaching out to us.

For your information, OpenVINO does provide Object Detection Demo which is available from Open Model Zoo. However, do note that Yolov5s and Yolov7-tiny models are not listed in the Supported Models of the demo.

I would recommend you to try out this Object Detection & YOLOs demo available from GitHub repository: bethusaisampath/YOLOv5_Openvino. Hope this helps.

Regards,

Hairul

[fizzgo]

I want to use python to run the YOLOv5s and YOLOv7-tiny with openvino. And my openvino version is 2021.4.752, l will try your recommendation. Thanks a lot.

[fizzgo]

I'm confused that the Github link you provided me didn't work successfully on Raspberry Pi. Thank you anyway. It didn't give any outputs or window, just nothing.

[Hairul_Intel]

Hi Fizzgo,

For your information, here are the Model Optimizer commands that I used to convert yolov5s model into Intermediate Representation (IR) format using openvino-dev[onnx] pip package:

mo --input_model yolov5s.onnx --output /model.24/m.0/Conv,/model.24/m.1/Conv,/model.24/m.2/Conv --data_type FP16 --scale_values=images[255] --input_shape=[1,3,640,640] --input=images

To find out your --input, --input_shape, and --output, please refer to this thread for more information.

On another note, you may use the following command in Object Detection Python Demo to do inference on MYRIAD with YOLOv5s model:

python3 object_detection_demo.py -at yolov4 -d MYRIAD -m yolov5s.xml -i <path_to_input>

Regards,

Hairul

[Hairul_Intel]

Hi fizzgo,

This thread will no longer be monitored since we have provided a solution. If you need any additional information from Intel, please submit a new question.

Regards,

Hairul