ReLU, Vectorised

ReLU is the activation function that made deep networks trainable, and it is almost embarrassingly simple: keep positive numbers, flatten everything else to zero.

relu(x) = max(0, x)

Task: write relu(x) where x is a list of numbers. Return a list where every negative entry has become 0.0 and everything else is unchanged.

• Do it elementwise across the whole list at once — reach for np.maximum rather than a Python loop.
• An empty list in means an empty list out.