For instance, if you're building a class called MLModel, using which you configure a machine learning model. When you run the model, you may want to make it callable, something like

model = MLModel()

# set up the model parameters

model()

Certainly, you can define a run method if you prefer, as an alternative.