--------INFO---------

This is my 'Hello World' project for machine learning. I used the MNIST data to train my model.

The training examples contain 28x28 pixel images of handwritten digits. I used tensorflow flow to build a neural network with 3 hidden layers and 1 output layer. The layers have 512, 256, and 128 neurons respectively. The output layer is 10 neurons with linear activation function. Softmax is applied when reading the model's results to make a final prediction. It is likely that a functioning model can be built with even smaller amount of nodes in each layer.

The GUI uses web-based technologies such as javascript, HTML and css along with electron that is hooked up to an executable program written with python.

--------Sources and Credits--------- Training data is taken from MNIST: https://www.kaggle.com/competitions/digit-recognizer Received help from gpt4 when building the front-end.