Create a model for use with VNRecognizedObjectObservation

Hi,

I can't find any documentation on how to train a model for use with VNRecognizedObjectObservation.
If you look at ObjectDetector.mlmodel, it's type is: Pipeline
Where standard Inception models types are: Image Classifier
For Object detection, I'm guessing I need to create a Pipeline, but can't find any documentation for it....

Are there any other constructors I can use other than:

let datasource: CreateML.MLImageClassifier.DataSource = .labeledDirectories(at: root)
let classifier = try MLImageClassifier(trainingData: datasource)
try classifier.write(to: modelURL)

for example:

let datasource: CreateML.MLPipelineClassifier.DataSource = .labeledDirectories(at: root, withOther: data)
let classifier = try MLPipelineClassifier(trainingData: datasource)
try classifier.write(to: modelURL)

Kind Regards
Chris