Neural Network Framework

This project was inspired by the book Neural Networks and Deep Learning by Michael Nielsen. It's an excellent resource if you want to start learning about neural networks. You can find the original online book here. Pre-trained Networks Included

This repository comes with pre-trained networks for:

Recognition of handwritten numbers from the MNIST dataset.
Recognition of "amoguses" (crewmembers from Among Us) and cat faces.

How to Run the Program

To run this program, execute framework.py. The available model names are:

amogus
catsAI
Numbers

The dataset names are:

amogus
cats
mnist

Program Options

There are three main options: 0 - Manage Datasets

Allows you to add or remove data points from the dataset. You can also add random data, which generates images with random pixel arrangements.

Manage Datasets 1 - Train Neural Network (trainNN)

Enables you to train the neural network.

Train Neural Network 2 - Draw Check

Allows you to draw an input for the neural network and check its performance.

Draw Check Additional Features

This framework also allows you to create custom networks and datasets.

Additional Features

Note: For "Draw Check" to work, the number of inputs should be exactly 784 (28x28).