Making a cloud deployment with AWS SageMaker
In the last few years, services such as AWS SageMaker have been gaining ground as an engine to run ML workloads. MLflow provides integrations and easy-to-use commands to deploy your model into the SageMaker infrastructure. The execution of this section will take several minutes (5 to 10 minutes depending on your connection) due to the need to build large Docker images and push the images to the Docker Registry.
The following is a list of some critical prerequisites for you to follow along:
- The AWS CLI configured locally with a default profile (for more details, you can look at https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-configure.html).
- AWS access in the account to SageMaker and its dependencies.
- AWS access in the account to push to Amazon Elastic Container Registry (ECR) service.
- Your MLflow server needs to be running as mentioned in the first Starting up a local model registry section.
To deploy...