How can you get out the model/ formula TensorFlow created for you?

I can understand that if you write python codes of an ML process first you train the model, then you test the model and finaly you write some code to make predictions. My question is how I can get the model out of the codes? I would use the model in a python program that would predect the same thing the code in the ML process does. How can I get the result of the whole ML process namely the “formula” or the model? I would use the model (formula) in a web app:
Thank you.