Skip to content

Latest commit

 

History

History
44 lines (30 loc) · 3.12 KB

intro.rst

File metadata and controls

44 lines (30 loc) · 3.12 KB

Introduction to Amazon SageMaker

You have several options for how you can use Amazon SageMaker.

  1. IDE: SageMaker Studio
  2. Console: SageMaker Notebook Instances
  3. Command line & SDK: AWS CLI, boto3, & SageMaker Python SDK
  4. 3rd party integrations: Kubeflow & Kubernetes operators

If you're new to SageMaker we recommend starting with more feature-rich SageMaker Studio. It uses the familiar JupyterLab interface and has seamless integration with a variety of deep learning and data science environments and scalable compute resources for training, inference, and other ML operations. Studio offers teams and companies easy on-boarding for their team members, freeing them up from complex systems admin and security processes. Administrators control data access and resource provisioning for their users.

Notebook Instances are another option. They have the familiar Jupyter and JuypterLab interfaces that work well for single users, or small teams where users are also administrators.

Advanced users also use SageMaker solely with the AWS CLI and Python scripts using boto3 and/or the SageMaker Python SDK. Kubeflow users can setup workflows to use SageMaker directly as well. And Kubernetes users can setup their environment to use SageMaker operators for training and inference.

This website offers examples for all types of customers and levels of expertise.

But if you're just getting started, we recommend that you check out the videos for Studio and Notebook Instances to decide which is best for you.

After you watch the video and complete on-boarding, you will have cloned this examples repo and be able to run the notebooks from this website from within SageMaker.

Get started with SageMaker Studio

Get started with SageMaker Notebook Instances