Studio Lab pre-installed environments - Amazon SageMaker
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Studio Lab pre-installed environments

Amazon SageMaker Studio Lab uses conda environments to contain your packages (or libraries). An environment is a folder that contains the packages you have installed. You can interact with an environment by using the terminal or your JupyterLab notebook. To use an environment and the packages installed within, you must choose the corresponding kernel that contains the same name as the environment when opening your JupyterLab notebook. For a walkthrough on how to manage your environments, see Manage your environment. For more information on installing packages within your environment, see Customize your environment.

Studio Lab has various environments pre-installed for you. Any changes made to persistent memory environments will remain for your next session. Any changes to non-persistent memory environments will not remain for your next sessions, but the packages within will be updated and tested for compatability by Amazon SageMaker. You will typically want to use the sagemaker-distribution non-persistent memory environment if you want to use a fully managed environment that already contains many popular packages used by machine learning (ML) engineers and data scientists. Otherwise you can use the default environment if you want to significantly customize your environment.

In the following we list the pre-installed environments and their use cases. To view the packages installed in an environment, see Customize your environment.

  • sagemaker-distribution: Non-persistent memory environment that is regularly updated and tested for compatibility, fully managed by Amazon SageMaker. This environment contains popular packages used in ML, data science, and visualization. The sagemaker-distribution environment is closely related to the environment used in Amazon SageMaker Studio Classic, so after graduating from Studio Lab to Studio Classic the notebooks should run similarly. For information on exporting your environment from Studio Lab to Studio Classic, see Export an Amazon SageMaker Studio Lab environment to Amazon SageMaker Studio Classic.

  • default: Persistent memory environment with very few packages pre-installed. Any installed packages or changes to this environment will continue on your next session.

  • studiolab: Persistent memory environment where JupyterLab and other related packages are installed. This environment should only be used for JupyterLab and Jupyter server extensions, for configuring the JupyterLab user interface.

  • studiolab-safemode: Non-persistent memory environment. This environment is automatically activated when there is an issue while starting your project runtime. Used for troubleshooting. For information on troubleshooting, see Troubleshooting.

  • base: Non-persistent memory environment. This environment is only used for system tooling and should not be used by customers.

For information on SageMaker images and their versions, see Amazon SageMaker images available for use with Studio Classic.