Skip to main content
Version: v1.1.x

Concepts

This page describes concepts that are found in the H2O eScorer documentation.

Snowflake​

Snowflake serves as the data foundation for H2O eScorer. It is a cloud-based enterprise data warehouse that provides secure storage and management capabilities for large datasets. H2O eScorer utilizes Snowflake to store and manage the data required for scoring.

Snowflake SQL queries​

H2O eScorer leverages Snowflake's SQL query functionality to interact with your data. These queries serve as structured instructions for retrieving specific information from your data stored in Snowflake. Essentially, they enable H2O eScorer to efficiently prepare data for model scoring.

Snowpark Container Services (SPCS)​

For advanced users and developers, Snowpark within Snowflake offers the ability to securely run non-SQL code like Python. This capability is particularly valuable for complex data manipulation tasks or model development beyond the capabilities of basic SQL queries. SPCS further enhance this functionality by allowing the execution of containerized applications – pre-packaged software environments – that contain your custom code for model development or processing. These container services essentially provide secure, isolated workspaces within Snowflake, equipped with the necessary tools for building and running your models directly.

Model stats dashboard​

This dashboard is a built-in visualization tool within H2O eScorer that also serves as a live dashboard for model stats across models. It serves as a central hub for displaying key metrics that evaluate the performance of your deployed models. These metrics can encompass various aspects like accuracy, precision, recall, or other relevant statistics depending on the specific model type. The model stats dashboard helps you to gain critical insights into model effectiveness and identify areas for potential improvement.

Amazon SageMaker​

Amazon SageMaker is a fully managed machine learning service from AWS that enables you to build, train, and deploy models at scale. H2O eScorer Standalone can be packaged as a Docker container and deployed to SageMaker, allowing you to create real-time inference endpoints for your H2O MOJO models. This integration combines H2O's model scoring capabilities with SageMaker's managed infrastructure for scalable, production-ready deployments.

Amazon Elastic Container Registry (AWS ECR)​

Amazon ECR is a fully managed container registry that stores, manages, and deploys Docker container images. When deploying H2O eScorer to Amazon SageMaker, you push the eScorer container image to ECR, which SageMaker then uses to create inference endpoints.

H2O eScorer Standalone​

H2O eScorer Standalone is a self-contained JAR file that packages the H2O eScorer scoring service. It can run independently in any Java environment or be containerized for deployment to cloud platforms like Amazon SageMaker. The standalone version is ideal for scenarios where you need to deploy scoring capabilities outside of the H2O AI Cloud managed environment.

Feedback