Neural Network Framework

From May to July of 2024, I created a framework in C to train, run, and store neural networks. The project is fully open source, and the code can be found at my GitHub. Here is the link.

This project was intended to be an educational tool, both for myself and others. One of the biggest features of this framework is the backpropogation tool. This enables users to train their own neural networks of any size (when provided training data). Once trained, users can then store neural networks and send them to other users to run, no re-training required. In addition to this framework, I wrote a technical paper detailing the optimizations of backpropogation algorithms. You can find details on that here.

Another aspect of this project was testing. I used statistical analysis, particularly in calculating error. Using a Python script I wrote, I generated large data sets to test training large neural networks.