MLS-C01 Question 273
Select 3You are building a machine learning workflow on AWS that requires the use of Amazon SageMaker for training and inference. To optimize the environment setup time for your training jobs, you decide to create a custom Amazon Machine Image (AMI) with all the necessary dependencies pre-installed. Which of the following steps are required to achieve this?
- A
Launch an EC2 instance with a base AMI and install the required dependencies.
- B
Create an Amazon SageMaker notebook instance and install the dependencies directly into the notebook.
- C
Use the
aws ec2 create-imagecommand to create a custom AMI from the EC2 instance. - D
Register the custom AMI with Amazon SageMaker to use it for training jobs.
- E
Assign the AMI to an EC2 Auto Scaling group to enable scaling for SageMaker.
Show answer and explanation
Correct answers: A, C, D
Explanation
To create a custom AMI for optimal environment setup time in Amazon SageMaker, you first launch an EC2 instance with a base AMI and install all necessary dependencies. After that, you use the aws ec2 create-image command to create a custom AMI from the configured EC2 instance. Finally, you register the custom AMI with Amazon SageMaker to use it for training jobs. This process ensures that your ML workflows are consistent and efficient.
- A. Correct.
Correct. Launching an EC2 instance with a base AMI and installing the required dependencies is the first step in creating a custom AMI.
- B. Incorrect.
Incorrect. Installing dependencies directly into a SageMaker notebook instance does not create a reusable AMI. Notebook instances are not AMIs.
- C. Correct.
Correct. The
aws ec2 create-imagecommand is used to create a custom AMI from a running EC2 instance. - D. Correct.
Correct. Once the custom AMI is created, it needs to be registered with Amazon SageMaker as a custom image to be used for training jobs.
- E. Incorrect.
Incorrect. SageMaker does not use EC2 Auto Scaling groups for running training jobs. Instead, it uses the custom AMI directly when configured.