Using AWS SageMaker Studio

Amazon SageMaker Studio is a web-based integrated development environment (IDE) for machine learning (ML). It allows data scientists and developers to perform all ML development steps—data preparation, model building, training, deployment, and management—on a single unified platform.

DeepSense AWS SageMaker Studio

DeepSense has acquired Amazon SageMaker Studio Domains per project to give users access to the Studio IDE.

You can find the Studio domain from the SageMaker service dashboard's left-side panel. These domains provide Resources (instances) that can be used for project work.

SageMaker Studio includes a comprehensive set of tools for every stage of ML development, enabling you to:

• Prepare and process data
• Build, train, and deploy models
• Manage experiments and training jobs
• Switch quickly between workflows (JupyterLab, Code Editor, or RStudio)

Use Studio notebooks in your project domain to prepare data, write training code, deploy ML models, and validate results.

Accessing DeepSense AWS SageMaker Studio

1. Log in to the AWS Access Portal (see Onboarding on AWS for setup instructions).
2. Click on the Applications tab.
3. Select your assigned SageMaker Studio to be redirected to the Studio IDE.

Working on DeepSense AWS SageMaker Studio

• From the Studio UI, create a new JupyterLab space:
  • Click on JupyterLab and select Create a new space (top-right corner).
  • If you have a supervisor, select a Shared space; otherwise, create a Private space.
  • After creation, select the space and review its details.

• Launch a Jupyter application:
  • Select the JupyterServer application.
  • Start the specific instance type you are instructed to use.
  • (Optional) In the Change environment dialog box, select a startup script if required.

• Recommended instance usage:
  • Use ml.t3.medium if editing code without execution.
  • For training or running experiments, switch to the instance type specified by your project lead.

• Notebook management:
  • To create a new notebook: go to File → New → Notebook.
  • Save your work frequently.
  • To stop an instance:

1. Navigate to the Running instances section in the SageMaker console.

2. Click Stop under the Actions column.

• • To restart: click Start for the stopped instance.

Best Practices

• Always stop unused notebook instances to avoid unnecessary charges.
• Use Shared spaces when collaborating with supervisors or teammates.
• Keep experiments organized by naming notebooks and spaces clearly.