Skip to main content

​
Introduction

Edge Inference Services is not yet available in production. The Inference tab currently shows a Coming Soon placeholder. This article describes the intended functionality.
This article describes the minimal steps required to get a working setup in Edge Inference Services. Follow the steps in order.

​
Prerequisites

​
Quick Start for Edge Inference Services

Edge Inference Services manages the lifecycle of AI models at the edge, including import, benchmarking, evaluation, and deployment to ZEDEDA Edge Kubernetes Service clusters.
  1. In the bottom navigation bar, select Inference to switch to the Inference service area.
  2. Create an organization. The name must be 1 to 49 characters long, use only lowercase letters, numbers, and hyphens, start and end with a letter or number, and be unique. The name system is reserved. See Organizations, Users, and Roles.
  3. Add a provider to connect to your model source. See Providers.
  4. Import a model from your provider into your model library. See Manage Models.
  5. Optionally, run inference benchmarks to evaluate model performance before deploying. See Benchmark Models and Manage Device Pools.
  6. Submit a deployment request to deploy the model to a ZEDEDA Edge Kubernetes Service cluster or cluster group. See Submit Model Deployment Requests.
  7. Have an Organization Admin or MLOps Admin in the organization approve the request. You cannot approve your own request. See Approve and Validate Models.

​
In This Section