Runtime configuration parameters specific to execution on Kubeflow.
tfx.v1.orchestration.experimental.KubeflowDagRunnerConfig(
pipeline_operator_funcs: Optional[List[OpFunc]] = None,
tfx_image: Optional[str] = None,
kubeflow_metadata_config: Optional[kubeflow_pb2.KubeflowMetadataConfig] = None,
supported_launcher_classes: Optional[List[Type[base_component_launcher.BaseComponentLauncher]]] = None,
metadata_ui_path: str = '/mlpipeline-ui-metadata.json',
**kwargs
)
Args |
|---|
pipeline_operator_funcs
|
A list of ContainerOp modifying functions that
will be applied to every container step in the pipeline.
|
tfx_image
|
The TFX container image to use in the pipeline.
|
kubeflow_metadata_config
|
Runtime configuration to use to connect to
Kubeflow metadata.
|
supported_launcher_classes
|
A list of component launcher classes that are
supported by the current pipeline. List sequence determines the order in
which launchers are chosen for each component being run.
|
metadata_ui_path
|
File location for metadata-ui-metadata.json file.
|
**kwargs
|
keyword args for PipelineConfig.
|
