A comprehensive C++ neural network library designed for building, training, and evaluating deep learning models. This project provides a high-performance, modular implementation with GPU acceleration, real-time visualization, and extensive documentation.
- 🚀 Modern C++: Built with C++17 for optimal performance and clean code
- 🔧 Modular Architecture: Easily extensible design for custom network architectures and activations
- ⚡ Neural Network Components: Complete support for sub-networks, activation functions, loss computations, and more
- 📦 CMake Ready: Seamless build system integration
- 🎯 Research Focused: Designed with experimentation and learning in mind
- 🧠 Real-time Visualization:
- Full Live Network: Real-time display of all layers, neurons, and activations for a complete live view of the model
- Status Only: Lightweight view showing just the current network status, ideal for fast debugging or monitoring
- 💾 Model Persistence: Save and load trained model parameters for reuse and deployment
- ⚙️ GPU Support: Leverages CUDA for accelerated training and inference on compatible NVIDIA GPUs, boosting performance for large-scale models
- 💻 Flexible Device Support: Simple to switch between CPU and GPU for training and inference
- Tensor Class: Multi-dimensional array with CPU/GPU support
- Activation Functions: ReLU, Leaky ReLU, Sigmoid, Tanh, Softmax
- Network Types: Fully Connected (FNN) and Convolutional (CNN) networks
- Optimizers: Various gradient descent optimizers - still in work
- Data Management: Flexible dataset loading and preprocessing
- Visualization: Real-time network visualization and monitoring
- Performance First: Optimized for both CPU and GPU execution
- Modularity: Clean separation of concerns for easy extension
- Documentation: Comprehensive inline documentation and examples
- C++17-compatible compiler (GCC 7+, Clang 5+, MSVC 2017+)
- CMake 3.28+ for build system
- CUDA Toolkit 11.0+ (optional, for GPU support)
- OpenGL (for visualization features)
# Clone the repository
git clone https://github.com/dredstone1/NeuralNetwork.git
# Navigate to project directory
cd NeuralNetwork
# Create and enter build directory
mkdir build && cd build
# Configure and build
cmake ..
makeNeuralNetwork/
├── 📂 include/ # Public API headers
│ ├── 📄 tensor.hpp # Core tensor class
│ ├── 📄 model.hpp # Main model interface
│ ├── 📄 dataBase.hpp # Dataset management
│ ├── 📄 Globals.hpp # Global constants and types
│ └── 📂 network/ # Network interface definitions
├── 📂 src/ # Implementation files
│ ├── 📂 model/ # Core model implementation
│ │ ├── 📄 tensor.cpp # Tensor implementation
│ │ ├── 📄 activations.cpp # Activation functions
│ │ ├── 📄 optimizers.cpp # Optimization algorithms
│ │ └── 📄 dataBase.cpp # Dataset handling
│ ├── 📂 networks/ # Network implementations
│ │ ├── 📂 fnn/ # Fully Connected Networks
│ │ └── 📂 cnn/ # Convolutional Networks
│ └── 📂 visualizer/ # Visualization components
├── 📂 tests/ # Unit tests and examples
│ ├── 📂 data/ # Test datasets
│ └── 📄 *.cpp # Test files
├── 📂 resources/ # Static assets
│ └── 📂 fonts/ # Font files for visualization
├── 📄 CMakeLists.txt # Build configuration
├── 📄 Doxyfile # Documentation configuration
└── 📖 README.md # This file
Contributions are more than welcome! Whether you're fixing bugs, adding features, or improving documentation:
- 🐛 Open issues for bugs or feature requests
- 🔀 Submit pull requests with your improvements
- 💡 Share ideas and suggestions
- 📖 Help improve documentation
This project is licensed under the MIT License. See the LICENSE file for complete details.
dredstone1 - GitHub Profile
- CUDA team for GPU acceleration support
- OpenGL community for visualization capabilities
- C++ community for modern language features
⭐ If you find this project interesting, consider giving it a star!