Register a model candidate
Make a completed Core ML training result available as a project model candidate.
Before you start
A completed Create ML run with an available saved model. Evaluate it first.
Quick start
- Inspect the completed run result and model identity.
- Choose Register model. Verify registration before selecting it in an inference workspace.
Functions and choices
Register a candidate
Registration records a candidate; it does not prove quality or automatically replace the current model.
Result
A model candidate with its training lineage.
Next step
Select the candidate in Capture Lab, Annotator or Failure Bank and test useful images.