Should you inherit from keras.Model even if you don't need the fit function?

Question

I am working on a project where I need to create a custom TensorFlow model but I don\'t need the built-in fit function, or basically anything else except the