Week 1 : Regularization

Looks like the test pass when I use rand instead of randn when creating the dropout layers. Considering both generate numbers between (0,1) , why do we use rand and not randn ?

Do they?

rand is a uniform distribution.
randn is a normal distribution.