Amazon SageMaker is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly. SageMaker abstracts much of the underlying infrastructure from the user. When you create a SageMaker notebook instance, it indeed launches an EC2 instance in the background, but this instance is not visible in your EC2 dashboard. This is because SageMaker notebook instances are run on EC2 instances that are managed by AWS within their own service accounts, not directly within your AWS account. Therefore, you can't directly access or manage the underlying EC2 instance or its EBS volume like you would with a regular EC2 instance that you launched yourself.

IMPORTANT: SageMaker does not allow burstable instances.

Built-in image classification algorithms

RecordIO protobuf format

DeepAR

Invocation Safety Factor

SageMakerVariantInvocationsPerInstance

Creating a Custom Algorithm Container for Amazon SageMaker

Hyper Parameter Optimization

Endpoint Configuration

Neo

Lifecycle Configuration Events

Ground Truth

Parameters

Operations

Model Monitor

SageMaker Debugger

BlazingText