Prerequisites for AI Registry standalone API

To set up the AI Registry standalone API, configure the Cloudera AI Inference service and import pretrained Models.

Prerequisites for Cloudera AI Inference service

AI Registry is a prerequisite for Cloudera AI Inference service because the Cloudera AI Inference service needs to deploy the models that are stored in the AI Registry.

  • To use the Cloudera AI Inference service, the latest AI Registry must be present in the same Cloudera Data Platform environment before the Cloudera AI Inference service is created.
  • If there is an older AI Registry in the environment that is created before May 14, 2024, follow the Upgrade AI Registry instructions to upgrade the AI Registry to the latest version before you create the Cloudera AI Inference service.
  • If the AI Registry is re-created, upgraded, or cert-renewed while the Cloudera AI Inference service is present, then follow the steps listed in the Manually updating AI Registry configuration topic to ensure that the configuration of AI Registry and Cloudera AI Inference service are synchronized.

Prerequisites to import pretrained models

You must add the URL details to allow them in the firewall rules.

NVIDIA GPU Cloud (NGC)

Add the following URL details so they can be allowed in the firewall's rules.

  • prod.otel.kaizen.nvidia.com (NVIDIA open telemetry)
  • api.ngc.nvidia.com
  • files.ngc.nvidia.com

Hugging Face

Add the following URL details so they can be allowed in the firewall's rules.
  • huggingface.co
  • cdn-lfs.huggingface.co
  • *.cloudfront.net (CDN)