mnist-digit-classification

MNIST handwritten digit: 10-class classification using Convolutional Neural Network (CNN)

Data

A large set of images of handwritten digits 0-9, Image size: 28 x 28 x 1

• Train: 48000 images
• Validation: 12000 images
• Test: 10000 images

Model

• 2D-CNN (Convolution, Max Pooling, and Fully Connected Layers)
• Implemented in tensorflow.keras