axon

A neural network implemented from scratch designed in a way to make using it straight forward

Features

• Calculating outputs based on inputs, weights, and activation functions
• Performing gradient descent to learn
• Loading and saving weights to a file

Requirements

• C++23 compiler
• CMake 3.14+
• Git

Instalation

Clone from github:

git clone https://github.com/Boxur/axon external/axon

CMake

Add to your project in CMake

add_subdirectory(external/axon)
 
target_link_libraries(project PRIVATE axon)

Usage

Create an implementation of the NetworkData class in include/axon/NetworkData.hpp. Create a Network class instance and pass in the implementation as a template argument of Network::Data<T>().

This can be see in the fully working example in example/.

Contributing

Contributions, bug reports, and suggestions are welcome.

Before submitting a pull request, please ensure that the project builds successfully

License

This project is licensed under the MIT License. See LICENSE for details.