Should model or guide functions have return at the end?

[Bayes]

I see that in tutorials some examples end with a return statements other examples don’t. Am I missing something?

Here’s an example where there is no return statement. Sounds arbitrary to me.

[jpchen]

since models are just stochastic functions that can generate data, it is useful sometimes to run the model/guide and generate a sample or make predictions, but no it is not necessary.