Cnn with multiple outputs

Lets say that I have a picture, with zero turkeys, or a turkey at some position (x1,y1), or two turkeys at position (x0,y0),(x1,y1).
I want the neural network to tell me the turkeys and their locations.
I’m not sure how to train the nn for multiple outputs.
For example it seems to me that there is now some order to the list. Would it declare a wrong answer if it outputs two turkeys in a different order, and how would it know the correct order.