I'm learning ML and I'm exploring object detection and classification. I discovered Yolo few months ago and it's impressively efficient and accurate. There are several pre-trained Yolo models on the web that can handle most of everyday objects.
Now, I'd like to train a Yolo model with custom objects. However... I want to keep already learned labels in the model. I would like to add new labels in the model, not to replace them. There are lot of tutorials that explain very well how to use custom labels from a pre-trained model but I did not find anything about increasing the number of classes inside a pre-trained model.
I barely found a discussion on a GitHub about this and the only solution was to train again all existing labels along with the new ones. I think this is quite tedious. I've read there is a mechanism called transfer learning but I'm not sure how this can be applied to add new labels in a model.