AWS SAM launches machine learning inference templates for AWS Lambda
AWS Serverless Application Model (AWS SAM) framework launches four new templates for building machine learning inference-based applications on AWS Lambda. Now, customers can leverage these templates as a starting point to build, test, and deploy their container-based Serverless ML applications.
AWS Serverless Application Model (AWS SAM) framework launches four new templates for building machine learning inference-based applications on AWS Lambda. Now, customers can leverage these templates as a starting point to build, test, and deploy their container-based Serverless ML applications.
Customers can access the ML templates as quick start templates from AWS SAM CLI init command. Upon selecting “Image” as package type and “python3.8-base” as their base image, customers will be able to choose an ML inference template for the framework of their choice. We are offering templates for the TensorFlow, PyTorch, XGBoost, and Scikit-Learn frameworks. Each template will generate a fully functional Serverless application to classify handwritten digits, and starter models are included. The application leverages an AWS Lambda function to perform inference and an Amazon API Gateway endpoint to expose this functionality as a RESTful service. Customers can use these templates as the starting point for their ML applications and then adapt them by adding their own model or code.
To get started, read the AWS Compute Blog and explore the SAM CLI template repository on GitHub. To learn more about AWS SAM, visit the AWS Serverless Application Model documentation.