https://store-images.s-microsoft.com/image/apps.34067.58340273-ba99-4f9d-8c85-8f5936b64f6c.3e9bfefa-b968-400e-9b83-ae76c3f0573a.2d974166-c52a-4f7c-8da4-a177fe05e05d

Data Science Virtual Machine - Windows 2019

Microsoft

Data Science Virtual Machine - Windows 2019

Microsoft

Development and modeling tools for AI, data science and analytics

The 'Data Science Virtual Machine (DSVM)' is a 'Windows Server 2019 with Containers' VM that has several popular tools for data exploration, analysis, modeling & development pre installed.

Highlights:

Operating System, Drivers and other base components

  • "Windows Server 2019 with Containers" image from Azure Marketplace
  • Nvidia drivers, CUDA Toolkit, cuDNN (when GPU machines are used)
  • Windows Subsystem for Linux (version 1)
  • Docker (Windows docker images only)
  • Anaconda ("conda")
  • Git

Authoring Tools

  • Visual Studio Code
  • Visual Studio 2019 Community Edition
  • PyCharm Community Edition
  • Jupyter, Jupyter Lab
  • Power BI Desktop
  • Weka

ML Framework

  • PyTorch, TensorFlow, scikit-learn
  • pyspark
  • Vowpal Wabbit

Other Notable Components

  • Azure CLI
  • Azure ML SDK for Python
  • Azure Storage Explorer
  • SQL Server 2019 Developer Edition
  • Microsoft 365 BYOL (Word, Excel, PowerPoint, Teams)

Users can access the DSVM via Remote Desktop or SSH. Optionally, the VM can be placed in a corporate network.

Users have full access to the DSVM. If needed, configurations can be adjusted, and additional frameworks can be installed like with any other virtual machine. The image provided here is a static VM image. Maintenance and protection against vulnerabilities of provisioned DSVMs is in the customer's responsibility. We recommend enabling "automatically download and install updates" in the Advanced options as this is not done by default.

There is also a DSVM on Ubuntu.

Note: Windows 2019 Gen 2 SKUs are currently not supported.

For more details, see the Data Science Virtual Machine documentation.