A 3-layer neural network to differentiate handwritten digits in MNIST database
The network utilises MNIST data (in CSV format) to train itself and test data is feed in afterwards to show the network's ability to differentiate new and not-seen-before digit images.
MNIST data uploaded along with the .py file. They are downloaded from mnist-in-csv (https://pjreddie.com/projects/mnist-in-csv/). The original MNIST data in, .ubyte format, can be found here: http://yann.lecun.com/exdb/mnist/.
This project is made possible by the guidance on building a 3-layer neural network as illustrated by Mr. Tariq Rashid in his work Make Your Own Neural Network. (https://www.amazon.com/gp/product/B01EER4Z4G)