cmnemoi-learn - Machine Learning from scratch by Charles-Meldhine Madi Mnemoi

Repository in which I will implement some of the machine learning models described in Elements Of Statistical Learning by Hastie, Tibshirani and Friedman from scratch (using only numpy) in form of a Python package.

The implementations will be unit tested against popular implementation (Scikit-learn, PyTorch) with pytest.

The quality of the code will be checked using black, pylint and mypy at each commit through a GitHub Action CI pipeline.

The package will be published on PyPI at each push to the main branch through a GitHub Action CD pipeline.

Install the package

pyenv and poetry

pyenv install 3.11
pyenv local 3.11
poetry install

Anaconda or Miniconda

conda create -n cmnemoi-learn python=3.11 -y
conda activate cmnemoi-learn
pip install cmnemoi-learn

Contributing

Clone the repo :

git clone https://github.com/cmnemoi/cmnemoi-learn.git
cd cmnemoi-learn

Then install dependencies.

If you run Poetry and pyenv:

pyenv install 3.11
pyenv local 3.11
poetry install --with=dev,test

If you run Miniconda or Anaconda:

conda create -n cmnemoi-learn-dev python=3.11 -y
conda activate cmnemoi-learn-dev
pip install -r requirements-dev.txt

License

MIT License