Write a tutorial about training a network in TF or PyTorch and then running it in OpenCV

TensorFlow (TF) and PyTorch are some of the most popular deep learning frameworks. TF exports networks in its own format; PyTorch supports export of the trained models into ONNX format, which is also very popular. OpenCV is becoming quite popular tool for Deep Learning Inference. It would be very useful to describe clearly how to train a model in TF or PyTorch so that it can be then run in OpenCV.