Is it possible to create and updatable model in code?

I am creating a model from an MLDataTable as such:


if let dataTable = try? MLDataTable(dictionary: priorityData) {

let (trainingCSVData, testCSVData) = dataTable.randomSplit(by: 0.8, seed: 0)

print( "Training Data = " + trainingCSVData.description)

do {

let deleteProbabilityClassifier = try MLClassifier(trainingData: dataTable, targetColumn: "DeleteProbability")

// evaluate it

let metrics = deleteProbabilityClassifier.evaluation(on: testCSVData)

let modelMetadata = MLModelMetadata(author: "Me", shortDescription: "Model for DeleteProbability Prediction", version: "1.1")

let modelPath = homeDirURL.appendingPathComponent(name + "-deleteProbabilityPredictor.mlmodel", isDirectory: false)

do {

try deleteProbabilityClassifier.write(to: modelPath, metadata: modelMetadata)

print("Success - DeleteProbability Model written" + modelPath.absoluteString )

print(metrics)

return true

}

catch {

print("Failure to write DeleteProbabilityClassifier - " + modelPath.absoluteString)

}

}

catch {

print("ML Classifier failed")

print("ML Classsifier threw error = " + error.localizedDescription)

}

}

else {

print("Failed - MLDataTable ")

print(type(of: priorityData))

}


What do I need to do to update this code to generate an updatable model? Currently I am running this code within a playground. I do not see any where to set the isUpdatable to true here?


Any help would be greatly appreciated.


Thanks,

Rob

hi, did you find any solution or example?

Is it possible to create and updatable model in code?
 
 
Q