My first attempt at writing a machine learning algorithm. Relies on a variant of the MNIST dataset created by Chris Crawford and can be found for free online. Uses a straightforward and basic structure to determine if a given image is any digit 0-9.
You'll first need to download the extended MNIST (EMNIST) dataset created by Chris Crawford, which can be found here: https://www.kaggle.com/crawford/emnist. You only need to download the emnist-digits-train.csv and the emnist-digits-test.csv files. Be sure to add these files to the digits_from_scratch folder. Then go to line 10 of the python file and update the path variable with the path of the folder containing these fileds. The code should run from there.
Thank you to Yann LeCun, Corinna Cortes, and Christopher J.C. Burges for putting together the original MNIST dataset and for making it available to anyone who wants to practice with it and to Chris Crawford for being as generous with his work. Also, thank you to Ajit Jairaj for helping me put this project together.