Pure-MNIST

The problem

Frameworks make backpropagation easy to use and easy to never understand. The point of this build was to be unable to hide behind loss.backward().

Approach

A 784-128-10 network implemented as a modular layer API in NumPy only: DenseLayer owns the weights and the Y = XW + b forward pass, ReLU handles the dead-neuron gradient mask, and Softmax is fused with cross-entropy for numerical stability. Weights start from He initialization; every gradient is derived by hand from the chain rule and applied through vectorized matrix updates. No autograd anywhere.

Results

Final training accuracy: 87.86% after 100 epochs. That's training accuracy, not held-out test accuracy. I haven't run a separate test split on this one yet.