Hey!, i have a question in this topic. When you make the dense function using vectorization, when getting Z now you get a matrix instead of a scalar. When having this, do you need to re-write the sigmoid function in order for it to work given a matrix as an input?

Hello @CoolingCoder,

Usually we implement a sigmoid that accepts array input, such as by using numpy or Tensorflow functions (those functions, such as `np.exp`

, will also accept scalar by the way).

So yes, we need to make sure it can work with an array!

Cheers,

Raymond

1 Like