YoloV3 training own data

1

How do I train YoloV3 to indentify my own custom objects?
All I can find is tutorials on how to train the COCO dataset with pretrained weights.

I use this implementation of YoloV3: https://github.com/ayooshkathuria/pytorch-yolo-v3

This is an example of things i have to detect:
CLEVR_train_011356

Is there anywhere a topic on how to train completely new data? Can’t seem to find it anywhere.

Thanks in advance

2

Check out @andy-yun’s repo.
He added some training routines to the YOLO implementation.