The ADAPT package

The ADAPT package is an open source library providing numerous tools to perform Transfer Learning and Domain Adaptation.

Enlarge the image
©Adapt Team

Presentation of the package

The ADAPT package is open-source python library providing the implementation of several domain adaptation methods. These methods correct the differences that exist in machine learning between the distribution of the training data (source) and that of the application data (target).

The purpose of the ADAPT library is to facilitate the access to transfer learning algorithms for a large public, including industrial players by allowing to quickly compare several methods. ADAPT offers more than fifteen methods as well as a modular base facilitating the development of new algorithms.

ADAPT is specifically designed for Scikit-learn and Tensorflow users with a "user-friendly" approach and offers a detailed documentation with many examples.

Bibliography

 

Resources

Adapt on Github

Awesome Domain Adaptation Python Toolbox : an open-source python library providing the implementation of several domain adaptation methods.

The project ADAPT has been funded by  Michelin and the Industrial Data Analytics and Machine Learning chair.