A personal project from high school aimed to understand CUDA kernels and Neural Networks.
This is implemented in C++ and features a Matrix library supporting both CPU and GPU operations. This library was then used as a backend for training multi layer perceptron networks. As well as Convolutional Neural Networks (At a later date)
The GPU CUDA code is the most interesting.
This library was then used in my first ever dissertation on the "Detection of Parkinson's within mouse data"