ConvNetCpp

ConvNetCpp is a C++ implementation of convolutional neural networks with integration of llama.cpp for AI model processing capabilities.

Prerequisites

• C++17 compatible compiler (clang++ or g++)
• Git
• CMake (required for building dependencies)
• Make or Ninja build system

Getting Started

Clone the Repository with Submodules

Since this project uses git submodules, you need to either clone with submodules:

git clone --recurse-submodules https://github.com/username/ConvNetCpp.git

Or if you've already cloned the repository, initialize and update submodules separately:

git submodule update --init --recursive

Build the Project

The project requires building its dependencies. Follow the build instructions in the respective directories:

1. Build llama.cpp submodule:

cd src/llama.cpp
mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j$(nproc)

The above commands will:

• Create a build directory inside llama.cpp
• Configure the build using CMake with Release flags for optimized performance
• Build all necessary binaries and libraries with parallel compilation using all CPU cores
• Place the resulting executables and libraries in src/llama.cpp/build/bin

After building, the llama.cpp directory will contain:

• Essential libraries (libggml.so, libllama.so, etc.)
• Executables (llama-cli, llama-server, llama-bench, etc.)
• Test utilities and example programs

2. Build the main ConvNetCpp project: Instructions for building the main project are available in the build documentation.

Configuration Notes

The project uses a CLANG.bm configuration file that references the built llama.cpp binaries and libraries. Ensure that the path in CLANG.bm matches where you built the project. By default, it expects the binaries in:

• Path: /home/sblo/Dev/ConvNetCpp/src/llama.cpp/build/bin
• Include: /home/sblo/Dev/ConvNetCpp/src/llama.cpp/ggml/include;/home/sblo/Dev/ConvNetCpp/src/llama.cpp/include
• Library: /home/sblo/Dev/ConvNetCpp/src/llama.cpp/build/bin

Project Structure

• src/ConvNet: Main ConvNet implementation
• src/ConvNetCtrl: ConvNet control components
• src/llama.cpp: AI model processing submodule
• examples/: Example implementations
• docs/: Documentation
• tutorial/: Tutorial materials
• upptst/: Unit/performance tests

Git Submodules

This project uses the following git submodules:

• src/llama.cpp: A C/C++ library for AI models, specifically optimized for LLM inference

Working with Submodules

After cloning or when updating the repository, always ensure submodules are up to date:

# Update all submodules to their latest commits
git submodule update --remote --merge

# If you want to update submodules to the latest commits and also pull latest changes
git pull --recurse-submodules

# To initialize and update submodules (if they are not initialized)
git submodule update --init --recursive

Contributing

1. Fork the repository
2. Create your feature branch
3. Ensure submodules are properly updated if needed
4. Commit your changes
5. Push to the branch
6. Create a new Pull Request

License

This project is licensed under the terms found in the LICENSE file.