Opoca

Opoca library aims to drastically speed up producing proof of concepts (PoC) for machine learning projects.

We define proof of concept as a small, quick and (not) dirty projects that results in:

• exploratory data analysis
• log of experiments along with models
• deployable best model
• demo (jupyter notebook/streamlit app etc.)
• short report including results analysis

There are several challenges that ML Engineer faces given a task to build new PoC project:

• it's not easy to track and reproduce experiments
• it's not easy to version and share data
• it's not easy to schedule jobs and not burn much money on training
• there's a lot of code that can be reused between different PoCs such as:
  • training logic for similar problems
  • evaluation logic
  • plotting
  • hyperparameters search
  • generic feature engineering transformations

Those are just few and a list is not complete without a doubt.

Prerequisites

Before you begin, ensure you have met the following requirements:

• You have installed the latest version of poetry

Installing Opoca

Opoca is installable from PyPi by executing:

pip install opoca

One may also use docker to build image:

docker build -t opoca -f Dockerfile .

And run bash session interactively by executing:

docker run -it --rm -v $PWD:/home -w /home opoca bash

Contributing to Opoca

To contribute to Opoca, follow these steps:

1. Fork this repository.
2. Create a branch: git checkout -b <branch_name>.
3. Make your changes and commit them: git commit -m '<commit_message>'
4. Push to the original branch: git push origin <project_name>/<location>
5. Create the pull request.

Alternatively see the GitHub documentation on creating a pull request.

Contributors

Thanks to the following people who have contributed to this project:

• @plazowicz
• @pedrito87
• @mjmikulski
• @AdrianMaciej
• @dkosowski87

Contact

If you want to contact me you can reach me at ml-team@netguru.com.

License

This project uses the following license: Apache License, Version 2.0.