Welcome to MAITE’s documentation!#

The Modular AI Trustworthy Engineering (MAITE) library provides common interfaces, utilities, and tooling to streamline the development of test and evaluation (T&E) capabilities, support core evaluation use cases, and allow composition of T&E tools into custom workflows.

Installation#

The core of MAITE is lightweight: its only dependencies are typing-extensions and numpy. To install it, run:

$ pip install maite

If instead you want to try out the features in the upcoming version, you can install the latest pre-release of MAITE with:

$ pip install --pre maite

Learning about MAITE#

Our docs are divided into four sections: Tutorials, How-Tos, Explanations, and Reference.

If you want to get a bird’s-eye view of what MAITE is all about check out our Tutorials. For folks who have already used the MAITE library, our How-Tos and Reference materials can help acquaint you with the unique capabilities that are offered by MAITE. Finally, Explanations provide readers with taxonomies, design principles, recommendations, and other articles that will enrich their understanding of the MAITE library.

Contents:

Indices and tables#