Canonical, a leader in open source and maker of Ubuntu, today announced Data Science Stack (DSS), an innovative solution that simplifies and accelerates the creation and management of data science environments. DSS enables businesses, researchers and developers to efficiently set up data environments with tools specifically designed to meet the specific needs of machine learning, artificial intelligence (AI) and data science.
Fully open source, free, and native to Ubuntu, it is also available on other Linux distributions, on Windows using Windows Subsystem Linux (WSL), and on macOS with Multipass. By default, DSS includes access to Jupyter Notebook for model development, MLflow for experiment tracking and model registry, and ML frameworks such as Pytorch and Tensorflow. However, users can customize Data Science Stack and add new libraries based on their use case.
DSS can be set up with just three commands, allowing for quick initial exploration on AI workstations. All that’s required is to configure the container orchestration layer, install the DSS CLI, and initialize the Data Science Stack to access the environment. This can be done in 10 to 30 minutes, depending on the technician’s experience level. DSS also provides migration paths, helping them scale their AI initiatives as projects mature.
To get early access to performance improvements and capabilities like Intel GPU support before they arrive upstream, you can access ITEX and IPEX, Intel’s distributions of PyTorch and Tensorflow. IPEX and ITEX improve optimization performance based on hardware, leveraging Advanced Vector Extensions (AVX), Vector Neural Network Instructions (VNNI), and Advanced Matrix Extensions (AMX). By integrating these extensions, in addition to GPU acceleration, DSS gains acceleration for common operations in AI use cases, reducing model training time and speeding up the experimentation phase of ML projects.
Canonical provides security maintenance for all packages included in the solution, enabling timely patching of vulnerabilities and protecting both the software and the artifacts created. The offering also includes simplified management of software dependencies and versions, reducing the technical challenges often faced by data scientists when deploying AI and machine learning models. Canonical has placed particular emphasis on optimizing cloud infrastructures for the Data Science Stack. Thanks to the integration with Kubernetes and native support for Ubuntu, it becomes easier to deploy and scale in hybrid or multi-cloud environments, benefiting from a robust and secure infrastructure.
“This removes the burden of managing package dependencies or configuring compute resources, with simple commands that AI practitioners can run,” said Chris Schnabel, Silicon Alliance Ecosystem Lead at Canonical. “By default, DSS includes access to Jupyter Notebook for model development, MLflow for experiment tracking and model registry, and ML frameworks like Pytorch and Tensorflow. However, users can customize Data Science Stack and add new libraries based on their use case.”
Related News :