Contents

init(validation:batchSize:maxIterations:gridSize:algorithm:)

Creates a model parameters instance for an object-detector training session.

Declaration

init(validation: MLObjectDetector.ModelParameters.ValidationData = .split(strategy: .automatic), batchSize: Int? = nil, maxIterations: Int? = nil, gridSize: CGSize = CGSize(width: 13, height: 13), algorithm: MLObjectDetector.ModelParameters.ModelAlgorithmType = .darknetYolo)

Parameters

  • validation:

    An Validationdata instance that contains your validation dataset.

  • batchSize:

    The number of images the object detector uses for each training iteration. If you don’t have a preference, set this parameter to nil to tell Create ML to use an appropriate value when it trains the model.

  • maxIterations:

    The largest number of training iterations the object detector can use. If you don’t have a preference, set this parameter to nil to tell Create ML to use an appropriate value when it trains the model.

  • gridSize:

    The number of rectangles, vertically and horizontally, the training algorithm uses to analyze each input image.

  • algorithm:

    The algorithm the training session uses to train the object detector.

See Also

Creating object detector parameters