Why is Theano.ifelse.ifelse executing both tensor functions?

[Ines Ayed]

I looked up a theano variant for keras.backend.switch, beause I did not want both operations to be executed and I found this :

https://github.com/Theano/Theano/blob/master/theano/ifelse.py

Here it says that (lazy) ifelse executes only the branch corresponding to the condition and not both like switch. I tested it like this: 

    import theano

    def function1():

        print("function 1 is executed")

        return a

    

    def function2():

        print("function 2 is executed")

        return b

    a = 2

    b = 10

    result = theano.ifelse.ifelse(T.gt(b,a), function1(), function2())

But when I run this both messages are printed which means that both branches are executed. This is confusing since the description of ifelse says that it should not. Am I missing something here?

[Brandon T. Willard]

What you’re seeing is most likely a property of Python and not Theano. In your example, function1 and function2 are evaluated when the call to theano.ifelse.ifelse is made. See Calls, which says:

All argument expressions are evaluated before the call is attempted.

​