Cloudera Docs

Configuring ML Runtime image for Agent Studio

Configure the ML Runtime image for Agent Studio by adding a Runtime repository entry or a private Docker registry to a Cloudera AI Workbench.

In an air-gapped environment, the Agent Studio ML Runtime image is available for download from the Cloudera AI Registry and can be hosted in their private Docker registry.

  1. Optional: Host the Agent Studio ML Runtime image in either a Docker private registry or in your own Runtime repository file.
    • Add the Docker credentials of the private registry to the workbench. For instructions, see Adding Docker registry credentials.
    • Create and internally host your own runtime repository file. If you use a custom repository, you must update the image_identifier value within the file. For instructions on adding a custom ML Runtime repository, see Adding ML Runtimes using Runtime Repo files.
  2. Optional: Add the docker credentials of the private registry in the workbench, for more information see, Adding Docker registry credentials.
  3. Optional: Create and internally host your own runtime repository file. For instructions on adding a custom ML Runtime repository, see Adding ML Runtimes using Runtime Repo files. You must update the image_identifier value within your self-hosted ML Runtime repository file.
  4. In the Cloudera console, click the Cloudera AI tile. The Cloudera AI Workbenches page is displayed.
  5. Click on the name of the workbench. The workbench Home page is displayed.
  6. Click Site Administration in the left navigation pane.
  7. As a site administrator go to Runtimes > Add Runtime Repo and add the ML Runtime repository file entry to obtain the most recent Agent Studio Runtime image.
  8. Manage Runtime updates.
    • Receive automatic updates for the Runtime catalog by selecting the Enable Runtime Updates checkbox.
    • Trigger an update immediately by clicking the Update Runtimes now button.
    Figure 1. Runtime updates section