Amazon SageMaker
Amazon SageMaker helps data scientists and developers to prepare, build, train, and deploy high-quality machine learning (ML) models quickly by bringing together a broad set of capabilities purpose-built for ML. You can deploy SageMaker Domain or Jupyter notebook instances on AWS using this module.
Examples
Getting started
AWS CLI
Follow the official guide to install and configure profiles.
After the installation is complete, you can check the aws cli version:
aws --version
aws-cli/2.5.8 Python/3.9.11 Darwin/21.4.0 exe/x86_64 prompt/off
Terraform
Terraform is an open-source infrastructure as code software tool that enables you to safely and predictably create, change, and improve infrastructure.
Install
This is the official guide for terraform binary installation. Please visit this Install Terraform website and follow the instructions.
Or, you can manually get a specific version of terraform binary from the websiate. Move to the Downloads page and look for the appropriate package for your system. Download the selected zip archive package. Unzip and install terraform by navigating to a directory included in your system's PATH
.
Or, you can use tfenv utility. It is very useful and easy solution to install and switch the multiple versions of terraform-cli.
First, install tfenv using brew.
brew install tfenv
Then, you can use tfenv in your workspace like below.
tfenv install <version>
tfenv use <version>
Also this tool is helpful to upgrade terraform v0.12. It is a major release focused on configuration language improvements and thus includes some changes that you'll need to consider when upgrading. But the version 0.11 and 0.12 are very different. So if some codes are written in older version and others are in 0.12 it would be great for us to have nice tool to support quick switching of version.
tfenv list
tfenv install latest
tfenv use <version>
Setup
module "sagemaker" {
source = "Young-ook/sagemaker/aws"
name = "sagemaker"
tags = { env = "test" }
}
Run terraform:
terraform init
terraform apply
Known Issues
Not supported instance in the availability zone
You may receive a 'Failed' error message when creating a SageMaker notebook instance. This happens for a number of reasons, so you should check the detailed messages about the state of your notebook instance.
Error: error waiting for sagemaker notebook instance (sagemaker-lhdhu-default) to create: unexpected state 'Failed',
wanted target 'InService'. last error: %!s(<nil>)
Open the AWS Management Console and go to the SageMaker service page. Then select the Notebook Instances menu in the left navigation bar. This will show the instances that failed to create. Select the instance and check the details in the pop-up message. If you see a message like the below, you should try creating the notebook instance again by changing the Availability Zone or Instance Type values.
The Notebook Instance type 'ml.t3.large' is not available in the availability zone 'ap-northeast-2b'. We apologize for the inconvenience.
Please try again using subnet in a different availability zone, or try a different instance type.
Run terraform to delete a randomly selected subnet index and notebook instance:
terraform destroy -target module.notebook
Then retry terraform apply
terraform apply