Three ReLU networks with different widths learning noisy step(x)

ReLU Explained: How This Activation Function Revolutionizes CNNs | Deep Learning Essential

ReLU Isn’t Always the Best Choice

The Hidden Problem with ReLU (And How Leaky ReLU Fixes It)

Three deep ReLU networks learn sin(10x) + 0.5sin(3x)

ReLU vs Leaky ReLU Explained

Three ReLU networks with different widths learning noisy sin

Three ReLU networks with different widths learning sign(sin⁡(x))

Three ReLU networks with different widths learning noisy parabola