H2O MLOps overview

H2O MLOps is an open, interoperable platform for model deployment, management, governance, monitoring, and alerting that features integration with H2O Driverless AI, H2O-3 open source, and third-party models.

Driverless AI compatibility

To determine whether a specific version of Driverless AI is compatible with MLOps, refer to the following table:

MLOps version	Driverless AI version
0.40.0 - 0.41.1	1.9.0.3 - 1.9.0.6
0.41.2 - 0.52.0	1.9.3 - 1.9.3.1
0.52.1 - 0.54.0	1.9.3 - 1.10.1.3
0.54.1 - 0.55.0	1.9.3 - 1.10.2
0.56.0	1.9.3 - 1.10.2
0.56.1	1.9.3 - 1.10.3
0.57.0 - 0.57.2	1.9.3 - 1.10.3.1
0.57.3	1.9.3 - 1.10.3.1, 1.10.4
0.58.x	1.9.3 - 1.10.4

Index

• Release notes

• Install MLOps

• Get started

• Projects:

  • Understand projects in MLOps

  • Create projects

  • View projects

• Experiments:

  • Understand experiments in MLOps

  • Add experiments

  • View experiments

• Model registry

• Model deployment:

  • Understand deployments in MLOps

  • Deploy a model

  • View deployments

  • Configure node affinity and toleration

  • Shapley values support

• Third-party model management integration

• Key terms

• MLOps Python client:

  • Installing the Python client

  • Using the Python client

  • Python client examples:

    • Driverless AI example

    • H2O-3 example

    • MLflow examples:

      • MLflow PyTorch example

      • MLflow scikit-learn example

      • MLflow custom Python model example

    • Model registry:

      • Registered models:

        • Create registered model

        • Get registered model

        • List registered models

        • Count registered models

      • Registered model versions:

        • Create model version

        • Get model version

        • List model versions for model

        • Count model versions

    • External registry:

      • Search external registry

      • List model versions

      • Add external registry model to MLOps

    • A/B Test deployment

    • Champion/Challenger deployment

    • Model monitoring:

      • List monitored deployments

      • Get monitored deployment

      • Get health metrics

      • Get drift metrics

    • Deploying a model with Kubernetes resources and replicas

    • Getting metadata for experiments

    • Getting deployment statuses for experiments

    • Creating secure endpoints

    • Sharing projects

    • Pump data

    • Authentication methods in Keycloak

Feedback

• Send feedback on H2O MLOps to cloud-feedback@h2o.ai