A Kubeflow run is one execution instance of a Kubeflow pipeline. It applies the pipeline’s task graph and runtime parameters to execute an ML workflow. section 5.1.2
A run connects a defined pipeline to an operational execution:
- Execute tasks according to the pipeline DAG, honoring dependencies between component tasks. section 5.1.2
Income-classifier run
A run executes the income-classifier pipeline with values for MinIO locations, feature-store configuration, model identity, MLflow, and output storage. Its tasks fetch data, retrieve features, run inference, and write results.
