Model Serving

Configure endpoint traffic for registered models and inspect artifact and runtime readiness before invoking them.

Open Model Serving

Configure an endpoint

  1. Register a model version with its artifact and valid input/output signatures.
  2. Create a serving endpoint and choose served entities and versions. Traffic percentages must sum to 100.
  3. Inspect endpoint discovery and runtime readiness. Resolve missing artifacts or runtime configuration before invoking it.
  4. Send a small request matching the model signature and inspect the actual prediction response and errors.

Runtime requirements

The executable runtime inspects the selected model format and artifact location. ONNX and TensorFlow Lite have specific execution paths; TensorFlow Lite requires the configured runtime URL and credentials. Registration of a different framework is not proof of an executable backend.

Artifact download hosts and locations are restricted by the runtime configuration. Use the exact artifact recorded on the registered version. A route configuration or serving-mode label alone does not deploy a new cluster.

Discovery and invocation

GET on the invocation route returns endpoint and readiness information. POST requests execute predictions when the runtime is ready. Copy the endpoint from your workspace and use the request shape from the registered model’s signature.

API endpoints

Use your Genorbex deployment’s origin followed by the path below. Replace IDs with resources accessible to your account. Authentication and permissions are checked by each endpoint.

GET /api/orbexdb/ai-ml/serving/{slug}/invocations

Discovers the endpoint configuration and inspected runtime readiness.

Authentication: Genorbex account token and service/resource access; API-key support follows the endpoint’s authentication guard.

POST /api/orbexdb/ai-ml/serving/{slug}/invocations

Runs prediction using the registered version’s expected request signature.

Authentication: Genorbex account token and service/resource access; API-key support follows the endpoint’s authentication guard.

Need help with your setup? Visit the Help Center or contact Genorbex.