mlcore

A core Machine Learning library implemented in C++ with Python bindings via pybind11.

Overview

This repository is designed as a learning resource to understand how to implement basic ML algorithms and utilities in C++, and expose them to Python for easy experimentation. It demonstrates:

• Writing efficient C++ code for ML tasks
• Creating Python bindings with pybind11
• Using CMake to build C++ extension modules for Python

Prerequisites

• C++17 compatible compiler
• CMake >= 4.0
• Python 3.x
• pybind11 (can be installed via pip: pip install pybind11)

Getting Started

1. Clone the repository:

git clone https://github.com/yourusername/mlcore.git
cd mlcore

2. Create a build directory:

mkdir -p cpp/build
cd cpp/build

3. Build the project using CMake:

cmake ..
make

This will generate the Python bindings (mlcore_cpp) in the build directory.

4. Install Python dependencies:

pip install -r requirements.txt

5. Use in Python:

import sys
sys.path.append("cpp/build")  # path to the compiled module

import mlcore_cpp

# Example usage
log_reg = mlcore_cpp.LogisticRegression(alpha=0.01)

Notes

• You can use the provided alias to simplify building:

alias cmk='cd cpp/build && cmake .. && make'
cmk

• Environment variables (like build paths) are stored in .env to avoid hardcoding them in scripts.

• For development, you can regenerate Python stubs for IDE auto-completion:

bash generate_stubs.sh

Contributing

Contributions are welcome! Please submit pull requests or open issues for bugs, suggestions, or improvements.