super_model

This MNIST model uses a Multi-Layer Perceptron with 669,706 parameters. It takes 784 inputs, passes through two 512-unit ReLU layers, and outputs 10 classes, enabling effective handwritten digit classification.