Numerical Gradient Checking: Debug Your Autograd Engine
What is numerical gradient checking? The central difference formula, why you never train with finite differences, and a grad checker that catches real bugs.
What is numerical gradient checking? The central difference formula, why you never train with finite differences, and a grad checker that catches real bugs.
How does backpropagation work? Build a working autograd engine from scratch in ~80 lines of Python: computation graphs, chain rule, reverse-mode autodiff.