Question 2/10 v4 lecture 7

You are using the sigmoid trick and the highest value you want your model to be able to predict is 5. What max value do you need to set? (in fastai, the name of the parameter passed to a function that enables this is y_range).


It's impossible for a sigmoid to get all the way to the top or all the way to the bottom due to its asymptotic nature, no matter how big the x is. If you want to be actually able to predict a value of five then you need to use something higher than 5 as your maximum.

Relevant part of lecture