search

akbartus / Yolov8-Object-Detection-on-Browser

Example of YOLOv8 object detection on browser. It is powered by ONNX and TFJS and served through JavaScript without any frameworks. It demonstrates live web camera detection.
https://yolov8-object-detection.glitch.me/
MIT License
30 stars 13 forks source link

Custom YOLO model #3

Closed darich10 closed 4 months ago

darich10 commented 7 months ago

Hello,

Thanks for your contribution. I am trying to implement a custom model with a different number of classes but I'm not sure where to apply my changes. Could you help me with this process?

Thanks in advance.

akbartus commented 4 months ago

@darich10 thank you for your kind question. You can learn custom model implementation here: https://blog.roboflow.com/how-to-train-yolov8-on-a-custom-dataset/. Basically you need to generate model, indicate path to it, then replace classes in this code.

Alternatively I added tensorflow.js implementation. The only difference is that you convert your custom model to tfjs model.

I hope this helps.