A simple neural network implementation for handwritten digit classification using the MNIST dataset. This Python repository features a single-layer perceptron with ReLU activation, Softmax output, and cross-entropy loss. Notably, the implementation is entirely self-contained, intentionally avoiding external libraries. Dive into the basics of neural networks, implemented from scratch for hands-on learning without dependencies.
GraphicsMonster/MNIST-single-layer-perceptron
MNIST digit classifier with a single-layer perceptron. Pure Python implementation for hands-on learning, no external libraries used.
Jupyter Notebook