Skip to main content
Version: Next

Deploy a model to H2O MLOps through the H2O Hydrogen Torch UI

Overview​

After building an H2O Hydrogen Torch model, you can deploy it to H2O MLOps utilizing the H2O Hydrogen Torch UI.

note

You can also deploy a built model to H2O MLOps by utilizing the model's H2O MLOps pipeline. To learn more, see Download an experiment's H2O MLOps pipeline

Instructions​

Deploy a model to H2O MLOps​

To deploy a built model (experiment) to H2O MLOps utilizing the H2O Hydrogen Torch UI, consider the following instructions:

  1. In the H2O Hydrogen Torch navigation menu, click View experiments.
  2. Click the name of the experiment (model) you want to depoy to H2O MLOps.
  3. Click the Deployments tab.
  4. In the Project box, select or enter an H2O MLOps project.
    note
    • If a project does not exist in H2O MLOps, H2O Hydrogen Torch creates a new project named HT, which can be utilized.
    • You can enter a non-existing project name that creates a new project to deploy the model.
  5. In the Environment list, select an environment to deploy the built model.
  6. (Optional) Click the GPU deployment toggle.
  7. In the Runtime list, select a runtime to deploy the built model.
  8. Click Deploy.
note

After the deployed model achieves a HEALTHY status (the deployment is alive and healthy), you can view certain deployment pieces of information (for example, Endpoint URL). To view the deployed model's details, consider the following instructions:

  1. In the List of deployments table, select the deployed model you want to view its deployment details.
note
  • To learn about the different deployment states a model can be in, see States.
  • To learn about the available deployment details of a deployed model, see Deployment details.

H2O MLOps deployment settings​

H2O Hydrogen Torch offers several H2O MLOps deployment settings to deploy an H2O Hydrogen Torch model to H2O MLOps quickly. Available settings are described in the following subsections.

note

Project​

This setting defines the H2O MLOps project to deploy (store) the built H2O Hydrogen Torch model. To learn more about H2O MLOps projects, see Understand projects in H2O MLOps.

note
  • To deploy the model, you can enter a non-existent project name, which will create a project in H2O MLOps under that name to deploy the model.
  • Not chaning the default value for the Project setting, will create an H2O MLOPs project named HT where the model will be deployed.
  • To deploy the model, you can enter or select an existing H2O MLOps project name where the model will be deployed.

GPU deployment​

This setting determines whether to deploy the built model on GPUs.

note

CPU deployments are often sufficient and less demanding on resources.

Environment​

This setting defines the target deployment environment H2O MLOps utilizes to deploy the built H2O Hydrogen Torch model. To learn more about H2O MLOps deployments, see Understand deployments in H2O MLOps.

Runtime​

This setting defines the runtime H2O MLOps utilizes to deploy the built H2O Hydrogen Torch model.

Requests/Limits​

This setting defines the amount of resources H2O MLOps utilizes for the deployment.

  • Requests: Resource requests determine the amount of resources the deployment asks Kubernetes to provide. For example, if a replica requests 256Mi of memory, Kubernetes schedules the deployment on any node with enough memory to satisfy the stated requirement.
  • Limits: Resource limits determine the maximum available resources for deployment. If a deployment exceeds a limit on the amount of memory allocated, the deployment is restarted. Note that the deployment does not restart if it exceeds a CPU limit.
note
  • To add additional fields for any existing custom resources that an admin has set up in your Kubernetes cluster, consider the following instructions:
    1. Click Add row.
  • The resource requests and limits fields must be defined using the quantity suffixes used in Kubernetes. The default values for Memory and CPU requests are 256Mi and 100m, respectively. For more information, see Resource units in Kubernetes.
  • By default, resources are not limited.
  • When specifying custom resources, the pod cannot be scheduled if Kubernetes cannot satisfy the stated custom value(s).
  • For more information on resource requests and limits in Kubernetes, see Resource Management for Pods and Containers.

Access all H2O MLOps deployment settings​

By default, H2O Hydrogen Torch only displays specific H2O MLOps deployment settings (for example, Project, Environment, GPU deployment, etc.).

You can access other H2O MLOps deployment settings (for example, deployment limits). To access other (non-default) H2O MLOps deployment settings, consider the following instructions:

  1. In H2O Hydrogen Torch navigation menu, click Settings
  2. Click the Allow custom MLOps limits toggle.
  3. Click Save settings persistently.
Feedback