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
nilto 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
nilto 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.