Data Science Stack documentation

Data Science Stack (DSS) is a ready-to-run environment for Machine Learning (ML) and data science. It’s built on open-source tooling, including Canonical K8s, JupyterLab, and MLflow, and is usable on any Ubuntu/Snap-enabled workstation.

DSS provides a Command Line Interface (CLI) for managing containerised ML environment images such as PyTorch or TensorFlow, on top of Canonical K8s.

Typically, creating ML environments on a workstation involves complex and hard-to-reverse configurations. DSS solves this problem by providing accessible, production-ready, isolated, and reproducible ML environments that fully utilise a workstation’s GPUs.

Both ML beginners and engineers who need to build complex development and runtime environments will see set-up time reduced to a minimum, allowing them to get started with meaningful work within minutes.

In this documentation

Tutorial

Get started - a hands-on introduction to DSS for newcomers

How-to guides

Step-by-step guides covering key operations and common tasks with DSS

Explanation

Discussion and clarification of key topics

Project and community

Data Science Stack is an open-source project that values its community. We warmly welcome contributions, suggestions, fixes, and constructive feedback from everyone.