Perform Common Tasks in Studio - Amazon SageMaker

Perform Common Tasks in Studio

The following sections describe how to perform common tasks in Amazon SageMaker Studio. For an overview of the Studio interface, see Amazon SageMaker Studio UI Overview.

Upload files to Amazon SageMaker Studio

Amazon SageMaker Studio can open only files listed in the Studio file browser. The following procedure shows how to add files to the browser.

To upload files

  1. In the left sidebar, choose the File Browser icon ( ) to display the file browser.

  2. In the File Browser, choose the Upload Files icon ( ) to open the dialog.

  3. In File Upload, browse to the files, choose the files you want to upload, and then choose Open.

  4. Double-click a file to open the file in a new tab in SageMaker Studio.

Clone and Connect to a Git Repository

For this example, we connect to a local clone of the awslabs/amazon-sagemaker-examples repository (repo). Studio can't connect to a remote repo.

To connect to a repo

  1. On the top menu, choose File > New > Terminal.

  2. At the command prompt, run the following command and wait for the repo to finish cloning.

    git clone

  3. In the left sidebar, choose the Git icon ( ).

  4. Choose Go find a repo.

  5. In File Browser, the cloned repo appears in the list. Double click the repo to open it.

  6. Choose the Git icon to view a list of Git tools.

Stop a Training Job

When you stop a training job, its status changes to Stopping. An algorithm can delay termination in order to save model artifacts after which the job status changes to Stopped. For more information, see the StopTrainingJob API.

To stop a training job

  1. Follow the View and Compare Experiments, Trials, and Trial Components procedure on this page until you open the Describe Trial Component tab.

  2. At the upper-right side of the tab, choose Stop training job. The Status at the top left of the tab changes to Stopped.

  3. To view the training time and billing time, choose AWS Settings.

Manage Your Storage Volume

The first time a user on your team opens Amazon SageMaker Studio, Amazon SageMaker creates a domain for the team. The domain includes an Amazon Elastic File System (Amazon EFS) volume with home directories for each of your users. Notebook files and data files are stored in these directories. Users can't access each other's home directories.


Don't delete the Amazon EFS volume. If you delete it, the domain will no longer function and all of your users will lose their work.

Find Your Amazon EFS Storage Volume

To find your Amazon EFS volume

  1. From the Amazon SageMaker Studio Control Panel, under Studio Summary, find the Studio ID. The ID will be in the following format: d-xxxxxxxx.

  2. Pass the Studio ID, which is the same as DomainId, to the DescribeDomain API.

    In the response from DescribeDomain, note the value in the HomeEfsFileSystemId field.

  3. Open the Amazon EFS console.

  4. Under File systems, scroll through the list to find the HomeEfsFileSystemId value, which is the same as the File system ID.

Provide Feedback on Amazon SageMaker Studio

To provide feedback

  1. At the upper-right of SageMaker Studio, choose Feedback.

  2. Choose a smiley emoji to let us know how satisfied you are with SageMaker Studio and add any feedback you'd care to share with us.

  3. Decide whether to share your identity with us, then choose Submit.

Update Amazon SageMaker Studio

When you update Amazon SageMaker Studio to the latest release, Amazon SageMaker shuts down and restarts the JupyterServer App. Any unsaved notebook information is lost in the process. The user data in the Amazon EFS volume isn't touched.

After the JupyterServer App is restarted, you must reopen Studio through the Amazon SageMaker Studio Control Panel.

To update Studio

  1. (Optional) Choose Amazon SageMaker Studio on the top-left of Studio to open the landing page. The Studio version is shown on the bottom-left of the page.

  2. On the top menu, choose File then Shut Down.

  3. Choose one of the following options:

    • Shutdown Server – Shuts down the JupyterServer App. Terminal sessions, kernel sessions, SageMaker images, and instances aren't shut down. These resources continue to accrue charges.

    • Shutdown All – Shuts down all Apps, terminal sessions, kernel sessions, SageMaker images, and instances. These resources no longer accrue charges.

  4. Reopen Studio.