Keras MobileDetectNet (Fast Object Detection on Jetson Nano)

What if I want to train a model that can detect two objects? Lets say a car and a motorcycle?

So instead of having four separate folders such as:

~/photos/car/train
~/photos/car/val
~/photos/motorcycle/train
~/photos/motorcycle/val

Can I just merge the car/train with the motorcycle/train folder and do the same for the val folders? Which would give the following:

~/photos/vehicles/train
~/photos/vehicles/val

And after running the training, I’ll get a model that will draw boxes with different labels for cars and motorcycles?