The MNIST dataset comprises 60,000 training examples and 10,000 test examples of the handwritten digits 0–9, formatted as 28x28-pixel monochrome images. Inspired by tensorflow's tutorial for the MNIST dataset, I compare the accuracies of a simple neural network versus a Convolutional Neural Network for recognizing handwritten digits in a figure.
Model | Accuracy |
---|---|
Simple Neural Network | 98.17% |
Convolutional Neural Network | 98.95% |