AWS SageMaker: Sign In, Set up SageMaker Studio and Use Jupyter Notebook Instance
Sign Up for AWS Account
First, head on over to AWS and create an account. Note the following points:
- You have to give your Credit Card Information
- You might be charged depending on your usage
- There is an option for a free tier, but it keeps changing
Sign up for an AWS Free Tier Account here: Link
Once you sign up for your AWS Account, you will see your Amazon Console Home.
Create S3 Bucket for storage
Search for S3 in the Search Bar
Select create a bucket.
Give a name and select the below options.
Great, you have now successfully created your S3 bucket.
Keep the rest of the settings as default and create your S3 bucket!
Create ECR
Here, we will create a repository.
Select Create a repository and enter a unique name.
You can get your ARN by simply copying the URI below. You will need this to setup Sagemaker in the next step.
Set up SageMaker on AWS
You Home Console will look something like this:
Here, search for Sagemaker in the search bar and select the first option.
Once you are on the Amazon SageMaker page, select the “Getting Started” from the left pane.
Create a Role
Select Create a Role and enter the information for your desired role.
Configure ML Activities
Click Next and select all the options as mentioned below. Don’t select more than 10 roles — AWS does not allow it.
Add the bucket name you created in all of the configuration fields. First, enter S3 bucket name.
Next, enter the S3 bucket again and the URI that you copied. We have to convert the URI that we have into ARN format. Here is mine:
arn:aws:ecr:ap-southeast-2:038365619140:repository/mlopsanish
Next, enter your S3 bucket again and primary for the default workgroup.
Click the next button.
Add Additional policies & tags
Review Role
Validate
Search for IAM, and select Roles.
You will see your role show up here, with the prefix SageMaker.
Create Jupyter Notebook Instance
SageMaker Domain
Next, go back to SageMaker and configure a SageMaker Domain.
Do a quick setup.
It will work on creating a domain.
Once it is set, you will be able to access AWS SageMake Studio
Create Jupyter Notebook Instance on AWS SageMaker
Search for SageMaker Studio in AWS Home Console.
Go to Notebook Instances on the left panel as shown.
Select Create Notebook Instance
Create a notebook instance and hit the orange create
You will see the notebook being created.
Once it is set up, you will see the Create Notebook instance being “InService”
Select your notebook and then select Open JupyterLab
Select “Open Jupyter” and Tada! In a few minutes, you will have Jupyter Notebook hosted on the cloud.
Select conda3_python as the interpreter of choice and run!
