coreml

type ErrorHandler ¶

type ErrorHandler = func(error)

ErrorHandler handles The closures the task calls during the update process. The error can be type-asserted to *foundation.NSError for Domain, Code, and UserInfo.

Used by:

• MLModelAsset.FunctionNamesWithCompletionHandler
• [MLUpdateTask.UpdateTaskForModelAtURLTrainingDataConfigurationProgressHandlersError]
• [MLUpdateTask.UpdateTaskForModelAtURLTrainingDataProgressHandlersError]

type IMLArrayBatchProvider ¶

type IMLArrayBatchProvider interface {
	objectivec.IObject
	MLBatchProvider

	// Creates the batch provider based on the array of feature providers.
	InitWithFeatureProviderArray(array []objectivec.IObject) MLArrayBatchProvider
	// Creates a batch provider based on feature names and their associated arrays of data.
	InitWithDictionaryError(dictionary foundation.INSDictionary) (MLArrayBatchProvider, error)

	// The array of feature providers.
	Array() []objectivec.IObject
}

An interface definition for the MLArrayBatchProvider class.

Creating a batch provider ¶

• [IMLArrayBatchProvider.InitWithFeatureProviderArray]: Creates the batch provider based on the array of feature providers.
• [IMLArrayBatchProvider.InitWithDictionaryError]: Creates a batch provider based on feature names and their associated arrays of data.

Accessing the feature providers ¶

• [IMLArrayBatchProvider.Array]: The array of feature providers.

See: https://developer.apple.com/documentation/CoreML/MLArrayBatchProvider

type IMLCPUComputeDevice ¶

type IMLCPUComputeDevice interface {
	objectivec.IObject
	MLComputeDeviceProtocol
}

An interface definition for the MLCPUComputeDevice class.

See: https://developer.apple.com/documentation/CoreML/MLCPUComputeDevice

type IMLComputePlan ¶

type IMLComputePlan interface {
	objectivec.IObject

	// The model structure.
	ModelStructure() IMLModelStructure

	// Returns The anticipated compute devices that would be used for executing an ML Program operation.
	ComputeDeviceUsageForMLProgramOperation(operation IMLModelStructureProgramOperation) IMLComputePlanDeviceUsage
	// Returns the anticipated compute devices that would be used for executing a NeuralNetwork layer.
	ComputeDeviceUsageForNeuralNetworkLayer(layer IMLModelStructureNeuralNetworkLayer) IMLComputePlanDeviceUsage

	// Returns the estimated cost of executing an ML Program operation.
	EstimatedCostOfMLProgramOperation(operation IMLModelStructureProgramOperation) IMLComputePlanCost
}

An interface definition for the MLComputePlan class.

Getting the model structure ¶

• [IMLComputePlan.ModelStructure]: The model structure.

Getting the device usage ¶

• [IMLComputePlan.ComputeDeviceUsageForMLProgramOperation]: Returns The anticipated compute devices that would be used for executing an ML Program operation.
• [IMLComputePlan.ComputeDeviceUsageForNeuralNetworkLayer]: Returns the anticipated compute devices that would be used for executing a NeuralNetwork layer.

Getting the estimated cost ¶

• [IMLComputePlan.EstimatedCostOfMLProgramOperation]: Returns the estimated cost of executing an ML Program operation.

See: https://developer.apple.com/documentation/CoreML/MLComputePlan-85vdw

type IMLComputePlanCost ¶

type IMLComputePlanCost interface {
	objectivec.IObject

	// The estimated workload of executing the operation over the total model execution. The value is between [0.0, 1.0].
	Weight() float64
}

An interface definition for the MLComputePlanCost class.

Accessing the weight ¶

• [IMLComputePlanCost.Weight]: The estimated workload of executing the operation over the total model execution. The value is between [0.0, 1.0].

See: https://developer.apple.com/documentation/CoreML/MLComputePlanCost

type IMLComputePlanDeviceUsage ¶

type IMLComputePlanDeviceUsage interface {
	objectivec.IObject

	// The compute device that the framework prefers to execute the layer/operation.
	PreferredComputeDevice() MLComputeDeviceProtocol
	// The compute devices that can execute the layer/operation.
	SupportedComputeDevices() []objectivec.IObject
}

An interface definition for the MLComputePlanDeviceUsage class.

Getting the compute device ¶

• [IMLComputePlanDeviceUsage.PreferredComputeDevice]: The compute device that the framework prefers to execute the layer/operation.
• [IMLComputePlanDeviceUsage.SupportedComputeDevices]: The compute devices that can execute the layer/operation.

See: https://developer.apple.com/documentation/CoreML/MLComputePlanDeviceUsage

type IMLDictionaryConstraint ¶

type IMLDictionaryConstraint interface {
	objectivec.IObject

	// The key type for the dictionary.
	KeyType() MLFeatureType

	// The constraint for a dictionary feature.
	DictionaryConstraint() IMLDictionaryConstraint
	SetDictionaryConstraint(value IMLDictionaryConstraint)
	// The size and format constraints for an image feature.
	ImageConstraint() IMLImageConstraint
	SetImageConstraint(value IMLImageConstraint)
	// The constraints on a multidimensional array feature.
	MultiArrayConstraint() IMLMultiArrayConstraint
	SetMultiArrayConstraint(value IMLMultiArrayConstraint)
	// The constraints for a sequence feature.
	SequenceConstraint() IMLSequenceConstraint
	SetSequenceConstraint(value IMLSequenceConstraint)
	// The state feature value constraint.
	StateConstraint() IMLStateConstraint
	SetStateConstraint(value IMLStateConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLDictionaryConstraint class.

Accessing the constraint ¶

• [IMLDictionaryConstraint.KeyType]: The key type for the dictionary.

See: https://developer.apple.com/documentation/CoreML/MLDictionaryConstraint

type IMLDictionaryFeatureProvider ¶

type IMLDictionaryFeatureProvider interface {
	objectivec.IObject
	MLFeatureProvider

	// Creates the feature provider based on a dictionary.
	InitWithDictionaryError(dictionary foundation.INSDictionary) (MLDictionaryFeatureProvider, error)

	// Subscript interface for the feature provider to pass through to the dictionary.
	ObjectForKeyedSubscript(featureName string) IMLFeatureValue
	// The backing dictionary.
	Dictionary() foundation.INSDictionary

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLDictionaryFeatureProvider class.

Creating the provider ¶

• [IMLDictionaryFeatureProvider.InitWithDictionaryError]: Creates the feature provider based on a dictionary.

Accessing the features ¶

• [IMLDictionaryFeatureProvider.ObjectForKeyedSubscript]: Subscript interface for the feature provider to pass through to the dictionary.
• [IMLDictionaryFeatureProvider.Dictionary]: The backing dictionary.

See: https://developer.apple.com/documentation/CoreML/MLDictionaryFeatureProvider

type IMLFeatureDescription ¶

type IMLFeatureDescription interface {
	objectivec.IObject

	// The name of this feature.
	Name() string
	// The type of this feature.
	Type() MLFeatureType
	// A Boolean value that indicates whether this feature is optional.
	Optional() bool

	// Checks whether the model will accept an input feature value.
	IsAllowedValue(value IMLFeatureValue) bool

	// The state feature value constraint.
	StateConstraint() IMLStateConstraint
	// The size and format constraints for an image feature.
	ImageConstraint() IMLImageConstraint
	// The constraint for a dictionary feature.
	DictionaryConstraint() IMLDictionaryConstraint
	// The constraints on a multidimensional array feature.
	MultiArrayConstraint() IMLMultiArrayConstraint
	// The constraints for a sequence feature.
	SequenceConstraint() IMLSequenceConstraint

	// A dictionary of input feature descriptions, which the model keys by the input’s name.
	InputDescriptionsByName() IMLFeatureDescription
	SetInputDescriptionsByName(value IMLFeatureDescription)
	// A dictionary of output feature descriptions, which the model keys by the output’s name.
	OutputDescriptionsByName() IMLFeatureDescription
	SetOutputDescriptionsByName(value IMLFeatureDescription)
	// Description of the state features.
	StateDescriptionsByName() IMLFeatureDescription
	SetStateDescriptionsByName(value IMLFeatureDescription)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLFeatureDescription class.

Inspecting a feature ¶

• [IMLFeatureDescription.Name]: The name of this feature.
• [IMLFeatureDescription.Type]: The type of this feature.
• [IMLFeatureDescription.Optional]: A Boolean value that indicates whether this feature is optional.

Checking for validity ¶

• [IMLFeatureDescription.IsAllowedValue]: Checks whether the model will accept an input feature value.

Accessing feature constraints ¶

• [IMLFeatureDescription.StateConstraint]: The state feature value constraint.
• [IMLFeatureDescription.ImageConstraint]: The size and format constraints for an image feature.
• [IMLFeatureDescription.DictionaryConstraint]: The constraint for a dictionary feature.
• [IMLFeatureDescription.MultiArrayConstraint]: The constraints on a multidimensional array feature.
• [IMLFeatureDescription.SequenceConstraint]: The constraints for a sequence feature.

See: https://developer.apple.com/documentation/CoreML/MLFeatureDescription

type IMLFeatureValue ¶

type IMLFeatureValue interface {
	objectivec.IObject

	// The type of the feature value.
	Type() MLFeatureType

	// A Boolean value that indicates whether the feature value is undefined or missing.
	Undefined() bool
	// The underlying integer of the feature value.
	Int64Value() int64
	// The underlying double of the feature value.
	DoubleValue() float64
	// The underlying string of the feature value.
	StringValue() string
	// The underlying image of the feature value as a pixel buffer.
	ImageBufferValue() corevideo.CVImageBufferRef
	// The underlying multiarray of the feature value.
	MultiArrayValue() IMLMultiArray
	// The underlying sequence of the feature value.
	SequenceValue() IMLSequence
	// The underlying dictionary of the feature value.
	DictionaryValue() foundation.INSDictionary

	// Returns a Boolean value that indicates whether a feature value is equal to another.
	IsEqualToFeatureValue(value IMLFeatureValue) bool

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLFeatureValue class.

Accessing the feature’s type ¶

• [IMLFeatureValue.Type]: The type of the feature value.

Accessing the feature’s value ¶

• [IMLFeatureValue.Undefined]: A Boolean value that indicates whether the feature value is undefined or missing.
• [IMLFeatureValue.Int64Value]: The underlying integer of the feature value.
• [IMLFeatureValue.DoubleValue]: The underlying double of the feature value.
• [IMLFeatureValue.StringValue]: The underlying string of the feature value.
• [IMLFeatureValue.ImageBufferValue]: The underlying image of the feature value as a pixel buffer.
• [IMLFeatureValue.MultiArrayValue]: The underlying multiarray of the feature value.
• [IMLFeatureValue.SequenceValue]: The underlying sequence of the feature value.
• [IMLFeatureValue.DictionaryValue]: The underlying dictionary of the feature value.

Comparing feature values ¶

• [IMLFeatureValue.IsEqualToFeatureValue]: Returns a Boolean value that indicates whether a feature value is equal to another.

See: https://developer.apple.com/documentation/CoreML/MLFeatureValue

type IMLGPUComputeDevice ¶

type IMLGPUComputeDevice interface {
	objectivec.IObject
	MLComputeDeviceProtocol

	// The device that represents the underlying metal device.
	MetalDevice() objectivec.IObject
}

An interface definition for the MLGPUComputeDevice class.

Getting the metal device ¶

• [IMLGPUComputeDevice.MetalDevice]: The device that represents the underlying metal device.

See: https://developer.apple.com/documentation/CoreML/MLGPUComputeDevice

type IMLImageConstraint ¶

type IMLImageConstraint interface {
	objectivec.IObject

	// The model’s default width for an image feature.
	PixelsWide() int
	// The model’s default height for an image feature.
	PixelsHigh() int
	// The model’s pixel format for an image feature.
	PixelFormatType() uint32

	// Additional sizes this image feature supports.
	SizeConstraint() IMLImageSizeConstraint

	// The size and format constraints for an image feature.
	ImageConstraint() IMLImageConstraint
	SetImageConstraint(value IMLImageConstraint)
	// The type of this feature.
	Type() MLFeatureType
	SetType(value MLFeatureType)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLImageConstraint class.

Accessing the constraints ¶

• [IMLImageConstraint.PixelsWide]: The model’s default width for an image feature.
• [IMLImageConstraint.PixelsHigh]: The model’s default height for an image feature.
• [IMLImageConstraint.PixelFormatType]: The model’s pixel format for an image feature.

Inspecting acceptable sizes ¶

• [IMLImageConstraint.SizeConstraint]: Additional sizes this image feature supports.

See: https://developer.apple.com/documentation/CoreML/MLImageConstraint

type IMLImageSize ¶

type IMLImageSize interface {
	objectivec.IObject

	// The height of an image feature in pixels.
	PixelsHigh() int
	// The width of an image feature in pixels.
	PixelsWide() int

	// An array of image sizes a model’s image feature accepts as input or produces as output.
	EnumeratedImageSizes() IMLImageSize
	SetEnumeratedImageSizes(value IMLImageSize)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLImageSize class.

Accessing an image size ¶

• [IMLImageSize.PixelsHigh]: The height of an image feature in pixels.
• [IMLImageSize.PixelsWide]: The width of an image feature in pixels.

See: https://developer.apple.com/documentation/CoreML/MLImageSize

type IMLImageSizeConstraint ¶

type IMLImageSizeConstraint interface {
	objectivec.IObject

	// Indicator of which properties to inspect for this image size constraint.
	Type() MLImageSizeConstraintType

	// The range of widths a model’s image feature accepts as input or produces as output.
	PixelsWideRange() foundation.NSRange
	// The range of heights a model’s image feature accepts as input or produces as output.
	PixelsHighRange() foundation.NSRange

	// An array of image sizes a model’s image feature accepts as input or produces as output.
	EnumeratedImageSizes() []MLImageSize

	// The model’s default height for an image feature.
	PixelsHigh() int
	SetPixelsHigh(value int)
	// The model’s default width for an image feature.
	PixelsWide() int
	SetPixelsWide(value int)
	// Additional sizes this image feature supports.
	SizeConstraint() IMLImageSizeConstraint
	SetSizeConstraint(value IMLImageSizeConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLImageSizeConstraint class.

Determining relevant constraints ¶

• [IMLImageSizeConstraint.Type]: Indicator of which properties to inspect for this image size constraint.

Accessing the image size ranges ¶

• [IMLImageSizeConstraint.PixelsWideRange]: The range of widths a model’s image feature accepts as input or produces as output.
• [IMLImageSizeConstraint.PixelsHighRange]: The range of heights a model’s image feature accepts as input or produces as output.

Accessing the enumerated image sizes ¶

• [IMLImageSizeConstraint.EnumeratedImageSizes]: An array of image sizes a model’s image feature accepts as input or produces as output.

See: https://developer.apple.com/documentation/CoreML/MLImageSizeConstraint

type IMLKey ¶

type IMLKey interface {
	objectivec.IObject

	// The name of the machine learning key.
	Name() string
	// The applicable scope of the machine learning key.
	Scope() string

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLKey class.

Retrieving a key’s information ¶

• [IMLKey.Name]: The name of the machine learning key.
• [IMLKey.Scope]: The applicable scope of the machine learning key.

See: https://developer.apple.com/documentation/CoreML/MLKey

type IMLMetricKey ¶

type IMLMetricKey interface {
	IMLKey

	// The training metrics of the model for the update task, contained in a dictionary.
	Metrics() IMLMetricKey
	SetMetrics(value IMLMetricKey)
}

An interface definition for the MLMetricKey class.

See: https://developer.apple.com/documentation/CoreML/MLMetricKey

type IMLModel ¶

type IMLModel interface {
	objectivec.IObject

	PredictionFromFeaturesError(input objectivec.IObject) (objectivec.IObject, error)
	PredictionFromFeaturesOptionsError(input objectivec.IObject, options IMLPredictionOptions) (objectivec.IObject, error)
	// Generates predictions for each input feature provider within the batch provider.
	PredictionsFromBatchError(inputBatch MLBatchProvider) (MLBatchProvider, error)
	// Generates a prediction for each input feature provider within the batch provider using the prediction options.
	PredictionsFromBatchOptionsError(inputBatch MLBatchProvider, options IMLPredictionOptions) (MLBatchProvider, error)
	PredictionFromFeaturesUsingStateError(inputFeatures objectivec.IObject, state IMLState) (objectivec.IObject, error)
	PredictionFromFeaturesUsingStateOptionsError(inputFeatures objectivec.IObject, state IMLState, options IMLPredictionOptions) (objectivec.IObject, error)

	// The list of available compute devices that the model’s prediction methods use.
	AvailableComputeDevices() objectivec.IObject
	SetAvailableComputeDevices(value objectivec.IObject)
	// The configuration of the model set during initialization.
	Configuration() IMLModelConfiguration
	// Model information you use at runtime during development, which Xcode also displays in its Core ML model editor view.
	ModelDescription() IMLModelDescription
	// Returns a model parameter value for a key.
	ParameterValueForKeyError(key IMLParameterKey) (objectivec.IObject, error)

	// A dictionary of the model’s creation information, such as its description, author, version, and license.
	Metadata() MLModelMetadataKey
	SetMetadata(value MLModelMetadataKey)
}

An interface definition for the MLModel class.

Making predictions ¶

• [IMLModel.PredictionFromFeaturesError]
• [IMLModel.PredictionFromFeaturesOptionsError]
• [IMLModel.PredictionsFromBatchError]: Generates predictions for each input feature provider within the batch provider.
• [IMLModel.PredictionsFromBatchOptionsError]: Generates a prediction for each input feature provider within the batch provider using the prediction options.
• [IMLModel.PredictionFromFeaturesUsingStateError]
• [IMLModel.PredictionFromFeaturesUsingStateOptionsError]

Inspecting a model ¶

• [IMLModel.AvailableComputeDevices]: The list of available compute devices that the model’s prediction methods use.
• [IMLModel.SetAvailableComputeDevices]
• [IMLModel.Configuration]: The configuration of the model set during initialization.
• [IMLModel.ModelDescription]: Model information you use at runtime during development, which Xcode also displays in its Core ML model editor view.
• [IMLModel.ParameterValueForKeyError]: Returns a model parameter value for a key.

See: https://developer.apple.com/documentation/CoreML/MLModel

type IMLModelAsset ¶

type IMLModelAsset interface {
	objectivec.IObject

	// The list of function names in the model asset.
	FunctionNamesWithCompletionHandler(handler ErrorHandler)

	// The default model descripton.
	ModelDescriptionWithCompletionHandler(handler MLModelDescriptionErrorHandler)
	// The model descripton for a specified function.
	ModelDescriptionOfFunctionNamedCompletionHandler(functionName string, handler MLModelDescriptionErrorHandler)
}

An interface definition for the MLModelAsset class.

Getting function names ¶

• [IMLModelAsset.FunctionNamesWithCompletionHandler]: The list of function names in the model asset.

Getting the model description ¶

• [IMLModelAsset.ModelDescriptionWithCompletionHandler]: The default model descripton.
• [IMLModelAsset.ModelDescriptionOfFunctionNamedCompletionHandler]: The model descripton for a specified function.

See: https://developer.apple.com/documentation/CoreML/MLModelAsset

type IMLModelConfiguration ¶

type IMLModelConfiguration interface {
	objectivec.IObject

	// Function name that [MLModel] will use.
	FunctionName() string
	SetFunctionName(value string)
	// A human readable name of a model for display purposes.
	ModelDisplayName() string
	SetModelDisplayName(value string)
	// A dictionary of configuration settings your app can override when loading a model.
	Parameters() foundation.INSDictionary
	SetParameters(value foundation.INSDictionary)

	// The metal device you prefer this model use to make predictions (inference) and update the model.
	PreferredMetalDevice() objectivec.IObject
	SetPreferredMetalDevice(value objectivec.IObject)
	// A Boolean value that determines whether to allow low-precision accumulation on a GPU.
	AllowLowPrecisionAccumulationOnGPU() bool
	SetAllowLowPrecisionAccumulationOnGPU(value bool)

	// The processing unit or units the model uses to make predictions.
	ComputeUnits() MLComputeUnits
	SetComputeUnits(value MLComputeUnits)

	// A group of hints for CoreML to optimize
	OptimizationHints() IMLOptimizationHints
	SetOptimizationHints(value IMLOptimizationHints)

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLModelConfiguration class.

Configuring model parameters ¶

• [IMLModelConfiguration.FunctionName]: Function name that MLModel will use.
• [IMLModelConfiguration.SetFunctionName]
• [IMLModelConfiguration.ModelDisplayName]: A human readable name of a model for display purposes.
• [IMLModelConfiguration.SetModelDisplayName]
• [IMLModelConfiguration.Parameters]: A dictionary of configuration settings your app can override when loading a model.
• [IMLModelConfiguration.SetParameters]

Configuring GPU usage ¶

• [IMLModelConfiguration.PreferredMetalDevice]: The metal device you prefer this model use to make predictions (inference) and update the model.
• [IMLModelConfiguration.SetPreferredMetalDevice]
• [IMLModelConfiguration.AllowLowPrecisionAccumulationOnGPU]: A Boolean value that determines whether to allow low-precision accumulation on a GPU.
• [IMLModelConfiguration.SetAllowLowPrecisionAccumulationOnGPU]

Allowing access to processing units ¶

• [IMLModelConfiguration.ComputeUnits]: The processing unit or units the model uses to make predictions.
• [IMLModelConfiguration.SetComputeUnits]

Getting optimization hints ¶

• [IMLModelConfiguration.OptimizationHints]: A group of hints for CoreML to optimize
• [IMLModelConfiguration.SetOptimizationHints]

See: https://developer.apple.com/documentation/CoreML/MLModelConfiguration

type IMLModelDescription ¶

type IMLModelDescription interface {
	objectivec.IObject

	// Description of the state features.
	StateDescriptionsByName() foundation.INSDictionary
	// A dictionary of input feature descriptions, which the model keys by the input’s name.
	InputDescriptionsByName() foundation.INSDictionary
	// A dictionary of output feature descriptions, which the model keys by the output’s name.
	OutputDescriptionsByName() foundation.INSDictionary

	// An array of labels, which can be either strings or a numbers, for classifier models.
	ClassLabels() []objectivec.IObject
	// A dictionary of the model’s creation information, such as its description, author, version, and license.
	Metadata() foundation.INSDictionary

	// The name of the primary prediction feature output description.
	PredictedFeatureName() string
	// The name of the feature output description for all probabilities of a prediction.
	PredictedProbabilitiesName() string

	// A Boolean value that indicates whether you can update the model with additional training.
	IsUpdatable() bool
	// A dictionary of the training input feature descriptions, which the model keys by the input’s name.
	TrainingInputDescriptionsByName() foundation.INSDictionary
	// A dictionary of the descriptions for the model’s parameters.
	ParameterDescriptionsByKey() foundation.INSDictionary

	// The list of available compute devices that the model’s prediction methods use.
	AvailableComputeDevices() objectivec.IObject
	SetAvailableComputeDevices(value objectivec.IObject)
	// The configuration of the model set during initialization.
	Configuration() IMLModelConfiguration
	SetConfiguration(value IMLModelConfiguration)
	// Model information you use at runtime during development, which Xcode also displays in its Core ML model editor view.
	ModelDescription() IMLModelDescription
	SetModelDescription(value IMLModelDescription)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLModelDescription class.

Accessing feature descriptions ¶

• [IMLModelDescription.StateDescriptionsByName]: Description of the state features.
• [IMLModelDescription.InputDescriptionsByName]: A dictionary of input feature descriptions, which the model keys by the input’s name.
• [IMLModelDescription.OutputDescriptionsByName]: A dictionary of output feature descriptions, which the model keys by the output’s name.

Accessing metadata ¶

• [IMLModelDescription.ClassLabels]: An array of labels, which can be either strings or a numbers, for classifier models.
• [IMLModelDescription.Metadata]: A dictionary of the model’s creation information, such as its description, author, version, and license.

Accessing prediction names ¶

• [IMLModelDescription.PredictedFeatureName]: The name of the primary prediction feature output description.
• [IMLModelDescription.PredictedProbabilitiesName]: The name of the feature output description for all probabilities of a prediction.

Accessing update descriptions ¶

• [IMLModelDescription.IsUpdatable]: A Boolean value that indicates whether you can update the model with additional training.
• [IMLModelDescription.TrainingInputDescriptionsByName]: A dictionary of the training input feature descriptions, which the model keys by the input’s name.
• [IMLModelDescription.ParameterDescriptionsByKey]: A dictionary of the descriptions for the model’s parameters.

See: https://developer.apple.com/documentation/CoreML/MLModelDescription

type IMLModelStructure ¶

type IMLModelStructure interface {
	objectivec.IObject

	// If the model is of NeuralNetwork type then it is the structure of the NeuralNetwork otherwise `nil`.
	NeuralNetwork() IMLModelStructureNeuralNetwork
	// If the model is of Pipeline type then it is the structure of the Pipeline otherwise `nil`.
	Pipeline() IMLModelStructurePipeline
	// If the model is of ML Program type then it is the structure of the ML Program otherwise `nil`.
	Program() IMLModelStructureProgram
}

An interface definition for the MLModelStructure class.

Model structures ¶

• [IMLModelStructure.NeuralNetwork]: If the model is of NeuralNetwork type then it is the structure of the NeuralNetwork otherwise `nil`.
• [IMLModelStructure.Pipeline]: If the model is of Pipeline type then it is the structure of the Pipeline otherwise `nil`.
• [IMLModelStructure.Program]: If the model is of ML Program type then it is the structure of the ML Program otherwise `nil`.

See: https://developer.apple.com/documentation/CoreML/MLModelStructure-c.class

type IMLModelStructureNeuralNetwork ¶

type IMLModelStructureNeuralNetwork interface {
	objectivec.IObject

	// The topologically sorted layers in the NeuralNetwork.
	Layers() []MLModelStructureNeuralNetworkLayer
}

An interface definition for the MLModelStructureNeuralNetwork class.

Accessing the layers ¶

• [IMLModelStructureNeuralNetwork.Layers]: The topologically sorted layers in the NeuralNetwork.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureNeuralNetwork

type IMLModelStructureNeuralNetworkLayer ¶

type IMLModelStructureNeuralNetworkLayer interface {
	objectivec.IObject

	// The input names.
	InputNames() []string
	// The layer name.
	Name() string
	// The output names.
	OutputNames() []string
	// The type of the layer, e,g, “elementwise”, “pooling”, etc.
	Type() string
}

An interface definition for the MLModelStructureNeuralNetworkLayer class.

Accessing the network layer properties ¶

• [IMLModelStructureNeuralNetworkLayer.InputNames]: The input names.
• [IMLModelStructureNeuralNetworkLayer.Name]: The layer name.
• [IMLModelStructureNeuralNetworkLayer.OutputNames]: The output names.
• [IMLModelStructureNeuralNetworkLayer.Type]: The type of the layer, e,g, “elementwise”, “pooling”, etc.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureNeuralNetworkLayer

type IMLModelStructurePipeline ¶

type IMLModelStructurePipeline interface {
	objectivec.IObject

	// The names of the sub models in the pipeline.
	SubModelNames() []string
	// The structure of the sub models in the pipeline.
	SubModels() []MLModelStructure
}

An interface definition for the MLModelStructurePipeline class.

Accessing the models and model names ¶

• [IMLModelStructurePipeline.SubModelNames]: The names of the sub models in the pipeline.
• [IMLModelStructurePipeline.SubModels]: The structure of the sub models in the pipeline.

See: https://developer.apple.com/documentation/CoreML/MLModelStructurePipeline

type IMLModelStructureProgram ¶

type IMLModelStructureProgram interface {
	objectivec.IObject

	// The functions in the program.
	Functions() foundation.INSDictionary
}

An interface definition for the MLModelStructureProgram class.

Accessing the program functions ¶

• [IMLModelStructureProgram.Functions]: The functions in the program.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgram

type IMLModelStructureProgramArgument ¶

type IMLModelStructureProgramArgument interface {
	objectivec.IObject

	// The array of bindings.
	Bindings() []MLModelStructureProgramBinding
}

An interface definition for the MLModelStructureProgramArgument class.

Accessing the bindings ¶

• [IMLModelStructureProgramArgument.Bindings]: The array of bindings.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramArgument

type IMLModelStructureProgramBinding ¶

type IMLModelStructureProgramBinding interface {
	objectivec.IObject

	// The name of the variable in the Program.
	Name() string
	// The compile time constant value in the Program.
	Value() IMLModelStructureProgramValue
}

An interface definition for the MLModelStructureProgramBinding class.

Accessing the program binding properties ¶

• [IMLModelStructureProgramBinding.Name]: The name of the variable in the Program.
• [IMLModelStructureProgramBinding.Value]: The compile time constant value in the Program.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramBinding

type IMLModelStructureProgramBlock ¶

type IMLModelStructureProgramBlock interface {
	objectivec.IObject

	// The named inputs to the block.
	Inputs() []MLModelStructureProgramNamedValueType
	// The list of topologically sorted operations in the block.
	Operations() []MLModelStructureProgramOperation
	// The output names.
	OutputNames() []string
}

An interface definition for the MLModelStructureProgramBlock class.

Accessing the program block properties ¶

• [IMLModelStructureProgramBlock.Inputs]: The named inputs to the block.
• [IMLModelStructureProgramBlock.Operations]: The list of topologically sorted operations in the block.
• [IMLModelStructureProgramBlock.OutputNames]: The output names.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramBlock

type IMLModelStructureProgramFunction ¶

type IMLModelStructureProgramFunction interface {
	objectivec.IObject

	// The active block in the function.
	Block() IMLModelStructureProgramBlock
	// The named inputs to the function.
	Inputs() []MLModelStructureProgramNamedValueType
}

An interface definition for the MLModelStructureProgramFunction class.

Accessing the program function properties ¶

• [IMLModelStructureProgramFunction.Block]: The active block in the function.
• [IMLModelStructureProgramFunction.Inputs]: The named inputs to the function.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramFunction

type IMLModelStructureProgramNamedValueType ¶

type IMLModelStructureProgramNamedValueType interface {
	objectivec.IObject

	// The name of the parameter.
	Name() string
	// The type of the parameter.
	Type() IMLModelStructureProgramValueType
}

An interface definition for the MLModelStructureProgramNamedValueType class.

Accessing the value type properties ¶

• [IMLModelStructureProgramNamedValueType.Name]: The name of the parameter.
• [IMLModelStructureProgramNamedValueType.Type]: The type of the parameter.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramNamedValueType

type IMLModelStructureProgramOperation ¶

type IMLModelStructureProgramOperation interface {
	objectivec.IObject

	// Nested blocks for loops and conditionals, e.g., a conditional block will have two entries here.
	Blocks() []MLModelStructureProgramBlock
	// The arguments to the Operation.
	Inputs() foundation.INSDictionary
	// The name of the operator, e.g., “conv”, “pool”, “softmax”, etc.
	OperatorName() string
	// The outputs of the Operation.
	Outputs() []MLModelStructureProgramNamedValueType
}

An interface definition for the MLModelStructureProgramOperation class.

Accessing the program operation properties ¶

• [IMLModelStructureProgramOperation.Blocks]: Nested blocks for loops and conditionals, e.g., a conditional block will have two entries here.
• [IMLModelStructureProgramOperation.Inputs]: The arguments to the Operation.
• [IMLModelStructureProgramOperation.OperatorName]: The name of the operator, e.g., “conv”, “pool”, “softmax”, etc.
• [IMLModelStructureProgramOperation.Outputs]: The outputs of the Operation.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramOperation

type IMLModelStructureProgramValue ¶

type IMLModelStructureProgramValue interface {
	objectivec.IObject
}

An interface definition for the MLModelStructureProgramValue class.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramValue

type IMLModelStructureProgramValueType ¶

type IMLModelStructureProgramValueType interface {
	objectivec.IObject
}

An interface definition for the MLModelStructureProgramValueType class.

See: https://developer.apple.com/documentation/CoreML/MLModelStructureProgramValueType

type IMLMultiArray ¶

type IMLMultiArray interface {
	objectivec.IObject

	// Creates a multidimensional array with a shape and type.
	InitWithShapeDataTypeError(shape []foundation.NSNumber, dataType MLMultiArrayDataType) (MLMultiArray, error)
	// Creates a multiarray from a data pointer.
	InitWithDataPointerShapeDataTypeStridesDeallocatorError(dataPointer unsafe.Pointer, shape []foundation.NSNumber, dataType MLMultiArrayDataType, strides []foundation.NSNumber, deallocator func(unsafe.Pointer)) (MLMultiArray, error)
	// Creates a multiarray sharing the surface of a pixel buffer.
	InitWithPixelBufferShape(pixelBuffer corevideo.CVImageBufferRef, shape []foundation.NSNumber) MLMultiArray

	// The total number of elements in the multiarray.
	Count() int
	// The underlying type of the multiarray.
	DataType() MLMultiArrayDataType
	// The multiarray’s multidimensional shape as a number array in which each element’s value is the size of the corresponding dimension.
	Shape() []foundation.NSNumber
	// A number array in which each element is the number of memory locations that span the length of the corresponding dimension.
	Strides() []foundation.NSNumber

	// Transfer the contents to the destination multi-array.
	TransferToMultiArray(destinationMultiArray IMLMultiArray)

	// A reference to the multiarray’s underlying pixel buffer.
	PixelBuffer() corevideo.CVImageBufferRef

	// A dictionary of input feature descriptions, which the model keys by the input’s name.
	InputDescriptionsByName() IMLFeatureDescription
	SetInputDescriptionsByName(value IMLFeatureDescription)
	// Model information you use at runtime during development, which Xcode also displays in its Core ML model editor view.
	ModelDescription() IMLModelDescription
	SetModelDescription(value IMLModelDescription)
	// The constraints on a multidimensional array feature.
	MultiArrayConstraint() IMLMultiArrayConstraint
	SetMultiArrayConstraint(value IMLMultiArrayConstraint)
	// A dictionary of output feature descriptions, which the model keys by the output’s name.
	OutputDescriptionsByName() IMLFeatureDescription
	SetOutputDescriptionsByName(value IMLFeatureDescription)
	// The constraint on the shape of the multiarray.
	ShapeConstraint() IMLMultiArrayShapeConstraint
	SetShapeConstraint(value IMLMultiArrayShapeConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLMultiArray class.

Creating a multiarray ¶

• [IMLMultiArray.InitWithShapeDataTypeError]: Creates a multidimensional array with a shape and type.
• [IMLMultiArray.InitWithDataPointerShapeDataTypeStridesDeallocatorError]: Creates a multiarray from a data pointer.
• [IMLMultiArray.InitWithPixelBufferShape]: Creates a multiarray sharing the surface of a pixel buffer.

Inspecting a multiarray ¶

• [IMLMultiArray.Count]: The total number of elements in the multiarray.
• [IMLMultiArray.DataType]: The underlying type of the multiarray.
• [IMLMultiArray.Shape]: The multiarray’s multidimensional shape as a number array in which each element’s value is the size of the corresponding dimension.
• [IMLMultiArray.Strides]: A number array in which each element is the number of memory locations that span the length of the corresponding dimension.

Transfering the contents ¶

• [IMLMultiArray.TransferToMultiArray]: Transfer the contents to the destination multi-array.

Accessing a multiarray’s elements ¶

• [IMLMultiArray.PixelBuffer]: A reference to the multiarray’s underlying pixel buffer.

See: https://developer.apple.com/documentation/CoreML/MLMultiArray

type IMLMultiArrayConstraint ¶

type IMLMultiArrayConstraint interface {
	objectivec.IObject

	// The shape of the multi array.
	Shape() []foundation.NSNumber
	// The type for the multi array.
	DataType() MLMultiArrayDataType
	// The constraint on the shape of the multiarray.
	ShapeConstraint() IMLMultiArrayShapeConstraint

	// The constraint for a dictionary feature.
	DictionaryConstraint() IMLDictionaryConstraint
	SetDictionaryConstraint(value IMLDictionaryConstraint)
	// The size and format constraints for an image feature.
	ImageConstraint() IMLImageConstraint
	SetImageConstraint(value IMLImageConstraint)
	// The constraints on a multidimensional array feature.
	MultiArrayConstraint() IMLMultiArrayConstraint
	SetMultiArrayConstraint(value IMLMultiArrayConstraint)
	// The constraints for a sequence feature.
	SequenceConstraint() IMLSequenceConstraint
	SetSequenceConstraint(value IMLSequenceConstraint)
	// The state feature value constraint.
	StateConstraint() IMLStateConstraint
	SetStateConstraint(value IMLStateConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLMultiArrayConstraint class.

Accessing the Constraints ¶

• [IMLMultiArrayConstraint.Shape]: The shape of the multi array.
• [IMLMultiArrayConstraint.DataType]: The type for the multi array.
• [IMLMultiArrayConstraint.ShapeConstraint]: The constraint on the shape of the multiarray.

See: https://developer.apple.com/documentation/CoreML/MLMultiArrayConstraint

type IMLMultiArrayShapeConstraint ¶

type IMLMultiArrayShapeConstraint interface {
	objectivec.IObject

	// Array of allowed shapes for a multiarray feature.
	EnumeratedShapes() []foundation.NSArray
	// The allowable range for a dimention of the multiarray.
	SizeRangeForDimension() []foundation.NSValue
	// The type of the shape constraint.
	Type() MLMultiArrayShapeConstraintType

	// The type for the multi array.
	DataType() MLMultiArrayDataType
	SetDataType(value MLMultiArrayDataType)
	// The shape of the multi array.
	Shape() foundation.NSNumber
	SetShape(value foundation.NSNumber)
	// The constraint on the shape of the multiarray.
	ShapeConstraint() IMLMultiArrayShapeConstraint
	SetShapeConstraint(value IMLMultiArrayShapeConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLMultiArrayShapeConstraint class.

Accessing the Constraints ¶

• [IMLMultiArrayShapeConstraint.EnumeratedShapes]: Array of allowed shapes for a multiarray feature.
• [IMLMultiArrayShapeConstraint.SizeRangeForDimension]: The allowable range for a dimention of the multiarray.
• [IMLMultiArrayShapeConstraint.Type]: The type of the shape constraint.

See: https://developer.apple.com/documentation/CoreML/MLMultiArrayShapeConstraint

type IMLNeuralEngineComputeDevice ¶

type IMLNeuralEngineComputeDevice interface {
	objectivec.IObject
	MLComputeDeviceProtocol

	// The total number of cores in the Neural Engine.
	TotalCoreCount() int
}

An interface definition for the MLNeuralEngineComputeDevice class.

Getting the Total Core Count ¶

• [IMLNeuralEngineComputeDevice.TotalCoreCount]: The total number of cores in the Neural Engine.

See: https://developer.apple.com/documentation/CoreML/MLNeuralEngineComputeDevice

type IMLNumericConstraint ¶

type IMLNumericConstraint interface {
	objectivec.IObject

	// The smallest numerical value allowed by this constraint.
	MinNumber() foundation.NSNumber
	// The largest numerical value allowed by this constraint.
	MaxNumber() foundation.NSNumber
	// A set of the numbers allowed in this constraint.
	EnumeratedNumbers() foundation.INSSet

	// The constraints of this paramter description value, if and only if the value is numerical.
	NumericConstraint() IMLNumericConstraint
	SetNumericConstraint(value IMLNumericConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLNumericConstraint class.

Numeric Constraints ¶

• [IMLNumericConstraint.MinNumber]: The smallest numerical value allowed by this constraint.
• [IMLNumericConstraint.MaxNumber]: The largest numerical value allowed by this constraint.
• [IMLNumericConstraint.EnumeratedNumbers]: A set of the numbers allowed in this constraint.

See: https://developer.apple.com/documentation/CoreML/MLNumericConstraint

type IMLOptimizationHints ¶

type IMLOptimizationHints interface {
	objectivec.IObject

	// The anticipated reshape frequency
	ReshapeFrequency() MLReshapeFrequencyHint
	SetReshapeFrequency(value MLReshapeFrequencyHint)

	// Optimization strategy for the model specialization.
	SpecializationStrategy() MLSpecializationStrategy
	SetSpecializationStrategy(value MLSpecializationStrategy)

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLOptimizationHints class.

Getting the reshape frequency ¶

• [IMLOptimizationHints.ReshapeFrequency]: The anticipated reshape frequency
• [IMLOptimizationHints.SetReshapeFrequency]

Getting the specialization strategy ¶

• [IMLOptimizationHints.SpecializationStrategy]: Optimization strategy for the model specialization.
• [IMLOptimizationHints.SetSpecializationStrategy]

See: https://developer.apple.com/documentation/CoreML/MLOptimizationHints-c.class

type IMLParameterDescription ¶

type IMLParameterDescription interface {
	objectivec.IObject

	// The default value for the parameter.
	DefaultValue() objectivec.IObject
	// The key for this parameter description value.
	Key() IMLParameterKey

	// The constraints of this paramter description value, if and only if the value is numerical.
	NumericConstraint() IMLNumericConstraint

	// A Boolean value that indicates whether you can update the model with additional training.
	IsUpdatable() bool
	SetIsUpdatable(value bool)
	// A dictionary of the descriptions for the model’s parameters.
	ParameterDescriptionsByKey() IMLParameterDescription
	SetParameterDescriptionsByKey(value IMLParameterDescription)
	// A dictionary of the training input feature descriptions, which the model keys by the input’s name.
	TrainingInputDescriptionsByName() IMLFeatureDescription
	SetTrainingInputDescriptionsByName(value IMLFeatureDescription)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLParameterDescription class.

Describing the model parameter ¶

• [IMLParameterDescription.DefaultValue]: The default value for the parameter.
• [IMLParameterDescription.Key]: The key for this parameter description value.

Constraining numeric values ¶

• [IMLParameterDescription.NumericConstraint]: The constraints of this paramter description value, if and only if the value is numerical.

See: https://developer.apple.com/documentation/CoreML/MLParameterDescription

type IMLParameterKey ¶

type IMLParameterKey interface {
	IMLKey

	// Creates a copy of a parameter key and adds the scope to it.
	ScopedTo(scope string) IMLParameterKey

	// The list of available compute devices that the model’s prediction methods use.
	AvailableComputeDevices() objectivec.IObject
	SetAvailableComputeDevices(value objectivec.IObject)
	// The configuration of the model set during initialization.
	Configuration() IMLModelConfiguration
	SetConfiguration(value IMLModelConfiguration)
	// Model information you use at runtime during development, which Xcode also displays in its Core ML model editor view.
	ModelDescription() IMLModelDescription
	SetModelDescription(value IMLModelDescription)
	// A dictionary of configuration settings your app can override when loading a model.
	Parameters() IMLParameterKey
	SetParameters(value IMLParameterKey)
}

An interface definition for the MLParameterKey class.

Scoping parameter keys ¶

• [IMLParameterKey.ScopedTo]: Creates a copy of a parameter key and adds the scope to it.

See: https://developer.apple.com/documentation/CoreML/MLParameterKey

type IMLPredictionOptions ¶

type IMLPredictionOptions interface {
	objectivec.IObject

	// A dictionary of feature names and client-allocated buffers.
	OutputBackings() foundation.INSDictionary
	SetOutputBackings(value foundation.INSDictionary)
}

An interface definition for the MLPredictionOptions class.

Getting features ¶

• [IMLPredictionOptions.OutputBackings]: A dictionary of feature names and client-allocated buffers.
• [IMLPredictionOptions.SetOutputBackings]

See: https://developer.apple.com/documentation/CoreML/MLPredictionOptions

type IMLSequence ¶

type IMLSequence interface {
	objectivec.IObject

	// The underlying type of the sequence’s elements.
	Type() MLFeatureType

	// An array of strings in the sequence.
	StringValues() []string
	// An array of 64-bit integers in the sequence.
	Int64Values() []foundation.NSNumber

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLSequence class.

Identifying the sequence’s element type ¶

• [IMLSequence.Type]: The underlying type of the sequence’s elements.

Retrieving the Sequence’s Values ¶

• [IMLSequence.StringValues]: An array of strings in the sequence.
• [IMLSequence.Int64Values]: An array of 64-bit integers in the sequence.

See: https://developer.apple.com/documentation/CoreML/MLSequence

type IMLSequenceConstraint ¶

type IMLSequenceConstraint interface {
	objectivec.IObject

	// The description that all sequence elements must match.
	ValueDescription() IMLFeatureDescription
	// The range of values allowed for the sequence’s length.
	CountRange() foundation.NSRange

	// The constraint for a dictionary feature.
	DictionaryConstraint() IMLDictionaryConstraint
	SetDictionaryConstraint(value IMLDictionaryConstraint)
	// The size and format constraints for an image feature.
	ImageConstraint() IMLImageConstraint
	SetImageConstraint(value IMLImageConstraint)
	// The constraints on a multidimensional array feature.
	MultiArrayConstraint() IMLMultiArrayConstraint
	SetMultiArrayConstraint(value IMLMultiArrayConstraint)
	// The constraints for a sequence feature.
	SequenceConstraint() IMLSequenceConstraint
	SetSequenceConstraint(value IMLSequenceConstraint)
	// The state feature value constraint.
	StateConstraint() IMLStateConstraint
	SetStateConstraint(value IMLStateConstraint)
	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLSequenceConstraint class.

Accessing the constraints ¶

• [IMLSequenceConstraint.ValueDescription]: The description that all sequence elements must match.
• [IMLSequenceConstraint.CountRange]: The range of values allowed for the sequence’s length.

See: https://developer.apple.com/documentation/CoreML/MLSequenceConstraint

type IMLState ¶

type IMLState interface {
	objectivec.IObject
}

An interface definition for the MLState class.

See: https://developer.apple.com/documentation/CoreML/MLState

type IMLStateConstraint ¶

type IMLStateConstraint interface {
	objectivec.IObject

	// The shape of the state buffer.
	BufferShape() int
	SetBufferShape(value int)
	// The data type of scalars in the state buffer.
	DataType() MLMultiArrayDataType

	EncodeWithCoder(coder foundation.INSCoder)
}

An interface definition for the MLStateConstraint class.

Inspecting a state constraint ¶

• [IMLStateConstraint.BufferShape]: The shape of the state buffer.
• [IMLStateConstraint.SetBufferShape]
• [IMLStateConstraint.DataType]: The data type of scalars in the state buffer.

See: https://developer.apple.com/documentation/CoreML/MLStateConstraint

type IMLTask ¶

type IMLTask interface {
	objectivec.IObject

	// A unique name of the task to distinguish it from all other tasks at runtime.
	TaskIdentifier() string

	// Begins or resumes a machine learning task.
	Resume()
	// Cancels a machine learning task before it completes.
	Cancel()

	// The current state of the machine learning task.
	State() MLTaskState
	// The underlying error if the task is in a failed state.
	Error() foundation.INSError
}

An interface definition for the MLTask class.

Identifying a task ¶

• [IMLTask.TaskIdentifier]: A unique name of the task to distinguish it from all other tasks at runtime.

Starting and stopping a task ¶

• [IMLTask.Resume]: Begins or resumes a machine learning task.
• [IMLTask.Cancel]: Cancels a machine learning task before it completes.

Checking the state of a task ¶

• [IMLTask.State]: The current state of the machine learning task.
• [IMLTask.Error]: The underlying error if the task is in a failed state.

See: https://developer.apple.com/documentation/CoreML/MLTask

type IMLUpdateContext ¶

type IMLUpdateContext interface {
	objectivec.IObject

	// The event type that triggered an update task to notify your app’s completion and update progress handlers.
	Event() MLUpdateProgressEvent
	// The update task that generated the update context.
	Task() IMLUpdateTask
	// The parameters for the update task.
	Parameters() foundation.INSDictionary

	// The training metrics of the model for the update task, contained in a dictionary.
	Metrics() foundation.INSDictionary

	// The underlying Core ML model stored in memory.
	Model() IMLModel
}

An interface definition for the MLUpdateContext class.

Getting the update context ¶

• [IMLUpdateContext.Event]: The event type that triggered an update task to notify your app’s completion and update progress handlers.
• [IMLUpdateContext.Task]: The update task that generated the update context.
• [IMLUpdateContext.Parameters]: The parameters for the update task.

Evaluating the update ¶

• [IMLUpdateContext.Metrics]: The training metrics of the model for the update task, contained in a dictionary.

Saving an updated model ¶

• [IMLUpdateContext.Model]: The underlying Core ML model stored in memory.

See: https://developer.apple.com/documentation/CoreML/MLUpdateContext

type IMLUpdateProgressHandlers ¶

type IMLUpdateProgressHandlers interface {
	objectivec.IObject

	// Creates the collection of closures an update task uses to notify your app of its progress.
	InitForEventsProgressHandlerCompletionHandler(interestedEvents MLUpdateProgressEvent, progressHandler MLUpdateContextHandler, completionHandler MLUpdateContextHandler) MLUpdateProgressHandlers
}

An interface definition for the MLUpdateProgressHandlers class.

Creating progress handlers ¶

• [IMLUpdateProgressHandlers.InitForEventsProgressHandlerCompletionHandler]: Creates the collection of closures an update task uses to notify your app of its progress.

See: https://developer.apple.com/documentation/CoreML/MLUpdateProgressHandlers

type IMLUpdateTask ¶

type IMLUpdateTask interface {
	IMLTask

	// Resumes a model update with updated parameter values.
	ResumeWithParameters(updateParameters foundation.INSDictionary)
}

An interface definition for the MLUpdateTask class.

Starting and Resuming an Update ¶

• [IMLUpdateTask.ResumeWithParameters]: Resumes a model update with updated parameter values.

See: https://developer.apple.com/documentation/CoreML/MLUpdateTask

type MLArrayBatchProvider ¶

type MLArrayBatchProvider struct {
	objectivec.Object
}

A convenience wrapper for batches of feature providers.

Overview ¶

This batch provider supports an array of feature providers or a dictionary of arrays of feature values.

Creating a batch provider ¶

• MLArrayBatchProvider.InitWithFeatureProviderArray: Creates the batch provider based on the array of feature providers.
• MLArrayBatchProvider.InitWithDictionaryError: Creates a batch provider based on feature names and their associated arrays of data.

Accessing the feature providers ¶

• MLArrayBatchProvider.Array: The array of feature providers.

See: https://developer.apple.com/documentation/CoreML/MLArrayBatchProvider

func MLArrayBatchProviderFromID(id objc.ID) MLArrayBatchProvider

func NewArrayBatchProviderWithDictionaryError(dictionary foundation.INSDictionary) (MLArrayBatchProvider, error)

func NewArrayBatchProviderWithFeatureProviderArray(array []objectivec.IObject) MLArrayBatchProvider

func NewMLArrayBatchProvider() MLArrayBatchProvider

func (a MLArrayBatchProvider) Array() []objectivec.IObject

func (a MLArrayBatchProvider) Autorelease() MLArrayBatchProvider

func (a MLArrayBatchProvider) Count() int

func (a MLArrayBatchProvider) FeaturesAtIndex(index int) MLFeatureProvider

func (a MLArrayBatchProvider) Init() MLArrayBatchProvider

func (a MLArrayBatchProvider) InitWithDictionaryError(dictionary foundation.INSDictionary) (MLArrayBatchProvider, error)

func (a MLArrayBatchProvider) InitWithFeatureProviderArray(array []objectivec.IObject) MLArrayBatchProvider

type MLArrayBatchProviderClass ¶

type MLArrayBatchProviderClass struct {
	// contains filtered or unexported fields
}

func GetMLArrayBatchProviderClass() MLArrayBatchProviderClass

func (mc MLArrayBatchProviderClass) Alloc() MLArrayBatchProvider

type MLBatchProvider ¶

type MLBatchProvider interface {
	objectivec.IObject

	// Returns the feature provider at the given index.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLBatchProvider/features(at:)
	FeaturesAtIndex(index int) MLFeatureProvider

	// The number of feature providers in this batch.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLBatchProvider/count
	Count() int
}

An interface that represents a collection of feature providers.

See: https://developer.apple.com/documentation/CoreML/MLBatchProvider

type MLBatchProviderObject ¶

type MLBatchProviderObject struct {
	objectivec.Object
}

MLBatchProviderObject wraps an existing Objective-C object that conforms to the MLBatchProvider protocol.

func MLBatchProviderObjectFromID(id objc.ID) MLBatchProviderObject

func (o MLBatchProviderObject) BaseObject() objectivec.Object

func (o MLBatchProviderObject) Count() int

func (o MLBatchProviderObject) FeaturesAtIndex(index int) MLFeatureProvider

type MLCPUComputeDevice ¶

type MLCPUComputeDevice struct {
	objectivec.Object
}

An object that represents a CPU compute device.

See: https://developer.apple.com/documentation/CoreML/MLCPUComputeDevice

func MLCPUComputeDeviceFromID(id objc.ID) MLCPUComputeDevice

func NewMLCPUComputeDevice() MLCPUComputeDevice

func (c MLCPUComputeDevice) Autorelease() MLCPUComputeDevice

func (c MLCPUComputeDevice) Init() MLCPUComputeDevice

type MLCPUComputeDeviceClass ¶

type MLCPUComputeDeviceClass struct {
	// contains filtered or unexported fields
}

func GetMLCPUComputeDeviceClass() MLCPUComputeDeviceClass

func (mc MLCPUComputeDeviceClass) Alloc() MLCPUComputeDevice

type MLComputeDeviceProtocol ¶

type MLComputeDeviceProtocol interface {
	objectivec.IObject
}

An interface that represents a compute device type.

See: https://developer.apple.com/documentation/CoreML/MLComputeDeviceProtocol

type MLComputeDeviceProtocolObject ¶

type MLComputeDeviceProtocolObject struct {
	objectivec.Object
}

MLComputeDeviceProtocolObject wraps an existing Objective-C object that conforms to the MLComputeDeviceProtocol protocol.

func MLAllComputeDevices() []MLComputeDeviceProtocolObject

func MLComputeDeviceProtocolObjectFromID(id objc.ID) MLComputeDeviceProtocolObject

func (o MLComputeDeviceProtocolObject) BaseObject() objectivec.Object

type MLComputePlan ¶

type MLComputePlan struct {
	objectivec.Object
}

A class describing the plan for executing a model.

Overview ¶

The application can use the plan to estimate the necessary cost and resources of the model before running the predictions.

Getting the model structure ¶

• MLComputePlan.ModelStructure: The model structure.

Getting the device usage ¶

• MLComputePlan.ComputeDeviceUsageForMLProgramOperation: Returns The anticipated compute devices that would be used for executing an ML Program operation.
• MLComputePlan.ComputeDeviceUsageForNeuralNetworkLayer: Returns the anticipated compute devices that would be used for executing a NeuralNetwork layer.

Getting the estimated cost ¶

• MLComputePlan.EstimatedCostOfMLProgramOperation: Returns the estimated cost of executing an ML Program operation.

See: https://developer.apple.com/documentation/CoreML/MLComputePlan-85vdw

func MLComputePlanFromID(id objc.ID) MLComputePlan

func NewMLComputePlan() MLComputePlan

func (c MLComputePlan) Autorelease() MLComputePlan

func (c MLComputePlan) ComputeDeviceUsageForMLProgramOperation(operation IMLModelStructureProgramOperation) IMLComputePlanDeviceUsage

func (c MLComputePlan) ComputeDeviceUsageForNeuralNetworkLayer(layer IMLModelStructureNeuralNetworkLayer) IMLComputePlanDeviceUsage

func (c MLComputePlan) EstimatedCostOfMLProgramOperation(operation IMLModelStructureProgramOperation) IMLComputePlanCost

func (c MLComputePlan) Init() MLComputePlan

func (c MLComputePlan) ModelStructure() IMLModelStructure

type MLComputePlanClass ¶

type MLComputePlanClass struct {
	// contains filtered or unexported fields
}

func GetMLComputePlanClass() MLComputePlanClass

func (mc MLComputePlanClass) Alloc() MLComputePlan

func (cc MLComputePlanClass) LoadContentsOfURLConfiguration(ctx context.Context, url foundation.INSURL, configuration IMLModelConfiguration) (*MLComputePlan, error)

func (_MLComputePlanClass MLComputePlanClass) LoadContentsOfURLConfigurationCompletionHandler(url foundation.INSURL, configuration IMLModelConfiguration, handler MLComputePlanErrorHandler)

func (cc MLComputePlanClass) LoadModelAssetConfiguration(ctx context.Context, asset IMLModelAsset, configuration IMLModelConfiguration) (*MLComputePlan, error)

func (_MLComputePlanClass MLComputePlanClass) LoadModelAssetConfigurationCompletionHandler(asset IMLModelAsset, configuration IMLModelConfiguration, handler MLComputePlanErrorHandler)

type MLComputePlanCost ¶

type MLComputePlanCost struct {
	objectivec.Object
}

A class that represents the estimated cost of executing a layer or operation.

Accessing the weight ¶

• MLComputePlanCost.Weight: The estimated workload of executing the operation over the total model execution. The value is between [0.0, 1.0].

See: https://developer.apple.com/documentation/CoreML/MLComputePlanCost

func MLComputePlanCostFromID(id objc.ID) MLComputePlanCost

func NewMLComputePlanCost() MLComputePlanCost

func (c MLComputePlanCost) Autorelease() MLComputePlanCost

func (c MLComputePlanCost) Init() MLComputePlanCost

func (c MLComputePlanCost) Weight() float64

type MLComputePlanCostClass ¶

type MLComputePlanCostClass struct {
	// contains filtered or unexported fields
}

func GetMLComputePlanCostClass() MLComputePlanCostClass

func (mc MLComputePlanCostClass) Alloc() MLComputePlanCost

type MLComputePlanDeviceUsage ¶

type MLComputePlanDeviceUsage struct {
	objectivec.Object
}

The anticipated compute devices to use for executing a layer or operation.

Getting the compute device ¶

• MLComputePlanDeviceUsage.PreferredComputeDevice: The compute device that the framework prefers to execute the layer/operation.
• MLComputePlanDeviceUsage.SupportedComputeDevices: The compute devices that can execute the layer/operation.

See: https://developer.apple.com/documentation/CoreML/MLComputePlanDeviceUsage

func MLComputePlanDeviceUsageFromID(id objc.ID) MLComputePlanDeviceUsage

func NewMLComputePlanDeviceUsage() MLComputePlanDeviceUsage

func (c MLComputePlanDeviceUsage) Autorelease() MLComputePlanDeviceUsage

func (c MLComputePlanDeviceUsage) Init() MLComputePlanDeviceUsage

func (c MLComputePlanDeviceUsage) PreferredComputeDevice() MLComputeDeviceProtocol

func (c MLComputePlanDeviceUsage) SupportedComputeDevices() []objectivec.IObject

type MLComputePlanDeviceUsageClass ¶

type MLComputePlanDeviceUsageClass struct {
	// contains filtered or unexported fields
}

func GetMLComputePlanDeviceUsageClass() MLComputePlanDeviceUsageClass

func (mc MLComputePlanDeviceUsageClass) Alloc() MLComputePlanDeviceUsage

type MLComputePlanErrorHandler ¶

type MLComputePlanErrorHandler = func(*MLComputePlan, error)

MLComputePlanErrorHandler handles When the compute plan is constructed successfully or unsuccessfully, the completion handler is invoked with a valid MLComputePlan instance or NSError object. The error can be type-asserted to *foundation.NSError for Domain, Code, and UserInfo.

Used by:

• [MLComputePlan.LoadContentsOfURLConfigurationCompletionHandler]
• [MLComputePlan.LoadModelAssetConfigurationCompletionHandler]

type MLComputeUnits ¶

type MLComputeUnits int

See: https://developer.apple.com/documentation/CoreML/MLComputeUnits

const (
	// MLComputeUnitsAll: The option you choose to allow the model to use all compute units available, including the neural engine.
	MLComputeUnitsAll MLComputeUnits = 2
	// MLComputeUnitsCPUAndGPU: The option you choose to allow the model to use both the CPU and GPU, but not the neural engine.
	MLComputeUnitsCPUAndGPU MLComputeUnits = 1
	// MLComputeUnitsCPUAndNeuralEngine: The option you choose to allow the model to use both the CPU and neural engine, but not the GPU.
	MLComputeUnitsCPUAndNeuralEngine MLComputeUnits = 3
	// MLComputeUnitsCPUOnly: The option you choose to limit the model to only use the CPU.
	MLComputeUnitsCPUOnly MLComputeUnits = 0
)

func (e MLComputeUnits) String() string

type MLCustomLayer ¶

type MLCustomLayer interface {
	objectivec.IObject

	// Assigns the weights for the connections within the layer.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLCustomLayer/setWeightData(_:)
	SetWeightDataError(weights []foundation.NSData) (bool, error)

	// Calculates the shapes of the output of this layer for the given input shapes.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLCustomLayer/outputShapes(forInputShapes:)
	OutputShapesForInputShapesError(inputShapes []foundation.NSArray) ([]foundation.NSArray, error)

	// Evaluates the custom layer with the given inputs.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLCustomLayer/evaluate(inputs:outputs:)
	EvaluateOnCPUWithInputsOutputsError(inputs []MLMultiArray, outputs []MLMultiArray) (bool, error)
}

An interface that defines the behavior of a custom layer in your neural network model.

See: https://developer.apple.com/documentation/CoreML/MLCustomLayer

type MLCustomLayerObject ¶

type MLCustomLayerObject struct {
	objectivec.Object
}

MLCustomLayerObject wraps an existing Objective-C object that conforms to the MLCustomLayer protocol.

func MLCustomLayerObjectFromID(id objc.ID) MLCustomLayerObject

func (o MLCustomLayerObject) BaseObject() objectivec.Object

func (o MLCustomLayerObject) EncodeToCommandBufferInputsOutputsError(commandBuffer metal.MTLCommandBuffer, inputs []objectivec.IObject, outputs []objectivec.IObject) (bool, error)

func (o MLCustomLayerObject) EvaluateOnCPUWithInputsOutputsError(inputs []MLMultiArray, outputs []MLMultiArray) (bool, error)

func (o MLCustomLayerObject) OutputShapesForInputShapesError(inputShapes []foundation.NSArray) ([]foundation.NSArray, error)

func (o MLCustomLayerObject) SetWeightDataError(weights []foundation.NSData) (bool, error)

type MLCustomModel ¶

type MLCustomModel interface {
	objectivec.IObject

	// Predicts output values from the given input features.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLCustomModel/prediction(from:options:)
	PredictionFromFeaturesOptionsError(input MLFeatureProvider, options IMLPredictionOptions) (MLFeatureProvider, error)
}

An interface that defines the behavior of a custom model.

See: https://developer.apple.com/documentation/CoreML/MLCustomModel

type MLCustomModelObject ¶

type MLCustomModelObject struct {
	objectivec.Object
}

MLCustomModelObject wraps an existing Objective-C object that conforms to the MLCustomModel protocol.

func MLCustomModelObjectFromID(id objc.ID) MLCustomModelObject

func (o MLCustomModelObject) BaseObject() objectivec.Object

func (o MLCustomModelObject) PredictionFromFeaturesOptionsError(input MLFeatureProvider, options IMLPredictionOptions) (MLFeatureProvider, error)

func (o MLCustomModelObject) PredictionsFromBatchOptionsError(inputBatch MLBatchProvider, options IMLPredictionOptions) (MLBatchProvider, error)

type MLDictionaryConstraint ¶

type MLDictionaryConstraint struct {
	objectivec.Object
}

The constraint on the keys for a dictionary feature.

Accessing the constraint ¶

• MLDictionaryConstraint.KeyType: The key type for the dictionary.

See: https://developer.apple.com/documentation/CoreML/MLDictionaryConstraint

func MLDictionaryConstraintFromID(id objc.ID) MLDictionaryConstraint

func NewMLDictionaryConstraint() MLDictionaryConstraint

func (d MLDictionaryConstraint) Autorelease() MLDictionaryConstraint

func (d MLDictionaryConstraint) DictionaryConstraint() IMLDictionaryConstraint

func (d MLDictionaryConstraint) EncodeWithCoder(coder foundation.INSCoder)

func (d MLDictionaryConstraint) ImageConstraint() IMLImageConstraint

func (d MLDictionaryConstraint) Init() MLDictionaryConstraint

func (d MLDictionaryConstraint) KeyType() MLFeatureType

func (d MLDictionaryConstraint) MultiArrayConstraint() IMLMultiArrayConstraint

func (d MLDictionaryConstraint) SequenceConstraint() IMLSequenceConstraint

func (d MLDictionaryConstraint) SetDictionaryConstraint(value IMLDictionaryConstraint)

func (d MLDictionaryConstraint) SetImageConstraint(value IMLImageConstraint)

func (d MLDictionaryConstraint) SetMultiArrayConstraint(value IMLMultiArrayConstraint)

func (d MLDictionaryConstraint) SetSequenceConstraint(value IMLSequenceConstraint)

func (d MLDictionaryConstraint) SetStateConstraint(value IMLStateConstraint)

func (d MLDictionaryConstraint) StateConstraint() IMLStateConstraint

type MLDictionaryConstraintClass ¶

type MLDictionaryConstraintClass struct {
	// contains filtered or unexported fields
}

func GetMLDictionaryConstraintClass() MLDictionaryConstraintClass

func (mc MLDictionaryConstraintClass) Alloc() MLDictionaryConstraint

type MLDictionaryFeatureProvider ¶

type MLDictionaryFeatureProvider struct {
	objectivec.Object
}

A convenience wrapper for the given dictionary of data.

Overview ¶

If your input data is stored in a dictionary, consider this type of MLFeatureProvider that is backed by a dictionary. It is a convenience interface, saving you the trouble of iterating through the dictionary to assign all of its values.

Creating the provider ¶

• MLDictionaryFeatureProvider.InitWithDictionaryError: Creates the feature provider based on a dictionary.

Accessing the features ¶

• MLDictionaryFeatureProvider.ObjectForKeyedSubscript: Subscript interface for the feature provider to pass through to the dictionary.
• MLDictionaryFeatureProvider.Dictionary: The backing dictionary.

See: https://developer.apple.com/documentation/CoreML/MLDictionaryFeatureProvider

func MLDictionaryFeatureProviderFromID(id objc.ID) MLDictionaryFeatureProvider

func NewDictionaryFeatureProviderWithDictionaryError(dictionary foundation.INSDictionary) (MLDictionaryFeatureProvider, error)

func NewMLDictionaryFeatureProvider() MLDictionaryFeatureProvider

func (d MLDictionaryFeatureProvider) Autorelease() MLDictionaryFeatureProvider

func (d MLDictionaryFeatureProvider) Dictionary() foundation.INSDictionary

func (d MLDictionaryFeatureProvider) EncodeWithCoder(coder foundation.INSCoder)

func (d MLDictionaryFeatureProvider) FeatureNames() foundation.INSSet

func (d MLDictionaryFeatureProvider) FeatureValueForName(featureName string) IMLFeatureValue

func (d MLDictionaryFeatureProvider) Init() MLDictionaryFeatureProvider

func (d MLDictionaryFeatureProvider) InitWithDictionaryError(dictionary foundation.INSDictionary) (MLDictionaryFeatureProvider, error)

func (d MLDictionaryFeatureProvider) ObjectForKeyedSubscript(featureName string) IMLFeatureValue

type MLDictionaryFeatureProviderClass ¶

type MLDictionaryFeatureProviderClass struct {
	// contains filtered or unexported fields
}

func GetMLDictionaryFeatureProviderClass() MLDictionaryFeatureProviderClass

func (mc MLDictionaryFeatureProviderClass) Alloc() MLDictionaryFeatureProvider

type MLFeatureDescription ¶

type MLFeatureDescription struct {
	objectivec.Object
}

The name, type, and constraints of an input or output feature.

Overview ¶

In Core ML, a is a single input or output of a model. A model can have any number of or . Each feature has a name and a value type, which are defined in the feature’s MLFeatureDescription. Model authors use feature descriptions to help developers integrate their model properly. Each MLFeatureDescription instance has read-only properties that indicate the feature’s name, its type, and whether it’s optional.

For examples of features, see Integrating a Core ML Model into Your App. Note the three input features named `solarPanels`, `greenhouses`, and `size`, and the output feature is named `price`. All four features are of type [Double].

An MLFeatureDescription may also include constraints, which specify the limitations of the model’s input and output features. For each input feature, the constraints describe what values the model expects from your app. For each output feature, the constraints describe what values your app should expect from the model. You can also write code to inspect these descriptions before using the model in your app.

Inspecting a feature ¶

• MLFeatureDescription.Name: The name of this feature.
• MLFeatureDescription.Type: The type of this feature.
• MLFeatureDescription.Optional: A Boolean value that indicates whether this feature is optional.

Checking for validity ¶

• MLFeatureDescription.IsAllowedValue: Checks whether the model will accept an input feature value.

Accessing feature constraints ¶

• MLFeatureDescription.StateConstraint: The state feature value constraint.
• MLFeatureDescription.ImageConstraint: The size and format constraints for an image feature.
• MLFeatureDescription.DictionaryConstraint: The constraint for a dictionary feature.
• MLFeatureDescription.MultiArrayConstraint: The constraints on a multidimensional array feature.
• MLFeatureDescription.SequenceConstraint: The constraints for a sequence feature.

See: https://developer.apple.com/documentation/CoreML/MLFeatureDescription

func MLFeatureDescriptionFromID(id objc.ID) MLFeatureDescription

func NewMLFeatureDescription() MLFeatureDescription

func (f MLFeatureDescription) Autorelease() MLFeatureDescription

func (f MLFeatureDescription) DictionaryConstraint() IMLDictionaryConstraint

func (f MLFeatureDescription) EncodeWithCoder(coder foundation.INSCoder)

func (f MLFeatureDescription) ImageConstraint() IMLImageConstraint

func (f MLFeatureDescription) Init() MLFeatureDescription

func (f MLFeatureDescription) InputDescriptionsByName() IMLFeatureDescription

func (f MLFeatureDescription) IsAllowedValue(value IMLFeatureValue) bool

func (f MLFeatureDescription) MultiArrayConstraint() IMLMultiArrayConstraint

func (f MLFeatureDescription) Name() string

func (f MLFeatureDescription) Optional() bool

func (f MLFeatureDescription) OutputDescriptionsByName() IMLFeatureDescription

func (f MLFeatureDescription) SequenceConstraint() IMLSequenceConstraint

func (f MLFeatureDescription) SetInputDescriptionsByName(value IMLFeatureDescription)

func (f MLFeatureDescription) SetOutputDescriptionsByName(value IMLFeatureDescription)

func (f MLFeatureDescription) SetStateDescriptionsByName(value IMLFeatureDescription)

func (f MLFeatureDescription) StateConstraint() IMLStateConstraint

func (f MLFeatureDescription) StateDescriptionsByName() IMLFeatureDescription

func (f MLFeatureDescription) Type() MLFeatureType

type MLFeatureDescriptionClass ¶

type MLFeatureDescriptionClass struct {
	// contains filtered or unexported fields
}

func GetMLFeatureDescriptionClass() MLFeatureDescriptionClass

func (mc MLFeatureDescriptionClass) Alloc() MLFeatureDescription

type MLFeatureProvider ¶

type MLFeatureProvider interface {
	objectivec.IObject

	// Accesses the feature value given the feature’s name.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLFeatureProvider/featureValue(for:)
	FeatureValueForName(featureName string) IMLFeatureValue

	// The set of valid feature names.
	//
	// See: https://developer.apple.com/documentation/CoreML/MLFeatureProvider/featureNames
	FeatureNames() foundation.INSSet
}

An interface that represents a collection of values for either a model’s input or its output.

See: https://developer.apple.com/documentation/CoreML/MLFeatureProvider

type MLFeatureProviderObject ¶

type MLFeatureProviderObject struct {
	objectivec.Object
}

MLFeatureProviderObject wraps an existing Objective-C object that conforms to the MLFeatureProvider protocol.

func MLFeatureProviderObjectFromID(id objc.ID) MLFeatureProviderObject

func (o MLFeatureProviderObject) BaseObject() objectivec.Object

func (o MLFeatureProviderObject) FeatureNames() foundation.INSSet

func (o MLFeatureProviderObject) FeatureValueForName(featureName string) IMLFeatureValue

type MLFeatureType ¶

type MLFeatureType int

See: https://developer.apple.com/documentation/CoreML/MLFeatureType

const (
	// MLFeatureTypeDictionary: The type for dictionary features and feature values.
	MLFeatureTypeDictionary MLFeatureType = 6
	// MLFeatureTypeDouble: The type for double features and feature values.
	MLFeatureTypeDouble MLFeatureType = 2
	// MLFeatureTypeImage: The type for image features and feature values.
	MLFeatureTypeImage MLFeatureType = 4
	// MLFeatureTypeInt64: The type for integer features and feature values.
	MLFeatureTypeInt64 MLFeatureType = 1
	// MLFeatureTypeInvalid: The type for invalid feature values.
	MLFeatureTypeInvalid MLFeatureType = 0
	// MLFeatureTypeMultiArray: The type for multidimensional array features and feature values.
	MLFeatureTypeMultiArray MLFeatureType = 5
	// MLFeatureTypeSequence: The type for sequence features and feature values.
	MLFeatureTypeSequence MLFeatureType = 7
	// MLFeatureTypeState: MLState.
	MLFeatureTypeState MLFeatureType = 8
	// MLFeatureTypeString: The type for string features and feature values.
	MLFeatureTypeString MLFeatureType = 3
)

func (e MLFeatureType) String() string

type MLFeatureValue ¶

type MLFeatureValue struct {
	objectivec.Object
}

A generic wrapper around an underlying value and the value’s type.

Overview ¶

A Core ML wraps an underlying value and bundles it with that value’s type, which is one of the types that MLFeatureType defines. Apps typically access feature values indirectly by using the methods in the wrapper class Xcode automatically generates for Core ML model files.

If your app accesses an MLModel directly, it must create and consume MLFeatureProvider instances. For each prediction, Core ML accepts a feature provider for its inputs, and generates a separate feature provider for its outputs. The input feature provider contains one MLFeatureValue instance per input, and the output feature provider contains one per output. See MLFeatureDescription for more information about the model input and output features.

Accessing the feature’s type ¶

• MLFeatureValue.Type: The type of the feature value.

Accessing the feature’s value ¶

• MLFeatureValue.Undefined: A Boolean value that indicates whether the feature value is undefined or missing.
• MLFeatureValue.Int64Value: The underlying integer of the feature value.
• MLFeatureValue.DoubleValue: The underlying double of the feature value.
• MLFeatureValue.StringValue: The underlying string of the feature value.
• MLFeatureValue.ImageBufferValue: The underlying image of the feature value as a pixel buffer.
• MLFeatureValue.MultiArrayValue: The underlying multiarray of the feature value.
• MLFeatureValue.SequenceValue: The underlying sequence of the feature value.
• MLFeatureValue.DictionaryValue: The underlying dictionary of the feature value.

Comparing feature values ¶

• MLFeatureValue.IsEqualToFeatureValue: Returns a Boolean value that indicates whether a feature value is equal to another.

See: https://developer.apple.com/documentation/CoreML/MLFeatureValue

func MLFeatureValueFromID(id objc.ID) MLFeatureValue

func NewFeatureValueUndefinedFeatureValueWithType(type_ MLFeatureType) MLFeatureValue

func NewFeatureValueWithCGImageConstraintOptionsError(cgImage coregraphics.CGImageRef, constraint IMLImageConstraint, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithCGImageOrientationConstraintOptionsError(cgImage coregraphics.CGImageRef, orientation objectivec.IObject, constraint IMLImageConstraint, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithCGImageOrientationPixelsWidePixelsHighPixelFormatTypeOptionsError(cgImage coregraphics.CGImageRef, orientation objectivec.IObject, pixelsWide int, pixelsHigh int, pixelFormatType uint32, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithCGImagePixelsWidePixelsHighPixelFormatTypeOptionsError(cgImage coregraphics.CGImageRef, pixelsWide int, pixelsHigh int, pixelFormatType uint32, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithDictionaryError(value foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithDouble(value float64) MLFeatureValue

func NewFeatureValueWithImageAtURLConstraintOptionsError(url foundation.INSURL, constraint IMLImageConstraint, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithImageAtURLOrientationConstraintOptionsError(url foundation.INSURL, orientation objectivec.IObject, constraint IMLImageConstraint, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithImageAtURLOrientationPixelsWidePixelsHighPixelFormatTypeOptionsError(url foundation.INSURL, orientation objectivec.IObject, pixelsWide int, pixelsHigh int, pixelFormatType uint32, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithImageAtURLPixelsWidePixelsHighPixelFormatTypeOptionsError(url foundation.INSURL, pixelsWide int, pixelsHigh int, pixelFormatType uint32, options foundation.INSDictionary) (MLFeatureValue, error)

func NewFeatureValueWithInt64(value int64) MLFeatureValue

func NewFeatureValueWithMultiArray(value IMLMultiArray) MLFeatureValue

func NewFeatureValueWithPixelBuffer(value corevideo.CVImageBufferRef) MLFeatureValue

func NewFeatureValueWithSequence(sequence IMLSequence) MLFeatureValue

func NewFeatureValueWithString(value string) MLFeatureValue

func NewMLFeatureValue() MLFeatureValue

func (f MLFeatureValue) Autorelease() MLFeatureValue

func (f MLFeatureValue) DictionaryValue() foundation.INSDictionary

func (f MLFeatureValue) DoubleValue() float64

func (f MLFeatureValue) EncodeWithCoder(coder foundation.INSCoder)

func (f MLFeatureValue) ImageBufferValue() corevideo.CVImageBufferRef

func (f MLFeatureValue) Init() MLFeatureValue

func (f MLFeatureValue) Int64Value() int64

func (f MLFeatureValue) IsEqualToFeatureValue(value IMLFeatureValue) bool

func (f MLFeatureValue) MultiArrayValue() IMLMultiArray

func (f MLFeatureValue) SequenceValue() IMLSequence

func (f MLFeatureValue) StringValue() string

func (f MLFeatureValue) Type() MLFeatureType

func (f MLFeatureValue) Undefined() bool

type MLFeatureValueClass ¶

type MLFeatureValueClass struct {
	// contains filtered or unexported fields
}

func GetMLFeatureValueClass() MLFeatureValueClass

func (mc MLFeatureValueClass) Alloc() MLFeatureValue

type MLFeatureValueImageOption ¶

type MLFeatureValueImageOption = string

MLFeatureValueImageOption is the initializer options you use to crop and scale an image when creating an image feature value.

See: https://developer.apple.com/documentation/CoreML/MLFeatureValue/ImageOption

type MLGPUComputeDevice ¶

type MLGPUComputeDevice struct {
	objectivec.Object
}

An object that represents a GPU compute device.

Getting the metal device ¶

• MLGPUComputeDevice.MetalDevice: The device that represents the underlying metal device.

See: https://developer.apple.com/documentation/CoreML/MLGPUComputeDevice

func MLGPUComputeDeviceFromID(id objc.ID) MLGPUComputeDevice

func NewMLGPUComputeDevice() MLGPUComputeDevice

func (g MLGPUComputeDevice) Autorelease() MLGPUComputeDevice

func (g MLGPUComputeDevice) Init() MLGPUComputeDevice

func (g MLGPUComputeDevice) MetalDevice() objectivec.IObject

type MLGPUComputeDeviceClass ¶

type MLGPUComputeDeviceClass struct {
	// contains filtered or unexported fields
}

func GetMLGPUComputeDeviceClass() MLGPUComputeDeviceClass

func (mc MLGPUComputeDeviceClass) Alloc() MLGPUComputeDevice

type MLImageConstraint ¶

type MLImageConstraint struct {
	objectivec.Object
}

The width, height, and pixel format constraints of an image feature.

Overview ¶

In CoreML, an is a collection of pixels represented by CVPixelBuffer (Swift) or CVPixelBuffer (Objective-C). An is a model input or output that accepts or produces, respectively, an image bundled in an MLFeatureValue. MLImageConstraint defines the image feature’s limitations for the images within an MLFeatureValue.

If a model has an image feature for an input or output, the model author uses an by creating an MLFeatureDescription. The feature description for an image input or output has:

- Its MLImageConstraint.Type property set to MLFeatureType.image - Its [ImageConstraint] property set to an MLImageConstraint instance configured to the image feature’s size and format

Image features that support additional image sizes provide a range of sizes, or a list of discrete sizes, in their image constraint’s MLImageConstraint.SizeConstraint property.

Accessing the constraints ¶

• MLImageConstraint.PixelsWide: The model’s default width for an image feature.
• MLImageConstraint.PixelsHigh: The model’s default height for an image feature.
• MLImageConstraint.PixelFormatType: The model’s pixel format for an image feature.

Inspecting acceptable sizes ¶

• MLImageConstraint.SizeConstraint: Additional sizes this image feature supports.

See: https://developer.apple.com/documentation/CoreML/MLImageConstraint

func MLImageConstraintFromID(id objc.ID) MLImageConstraint

func NewMLImageConstraint() MLImageConstraint

func (i MLImageConstraint) Autorelease() MLImageConstraint

func (i MLImageConstraint) EncodeWithCoder(coder foundation.INSCoder)

func (i MLImageConstraint) ImageConstraint() IMLImageConstraint

func (i MLImageConstraint) Init() MLImageConstraint

func (i MLImageConstraint) PixelFormatType() uint32

func (i MLImageConstraint) PixelsHigh() int

func (i MLImageConstraint) PixelsWide() int

func (i MLImageConstraint) SetImageConstraint(value IMLImageConstraint)

func (i MLImageConstraint) SetType(value MLFeatureType)

func (i MLImageConstraint) SizeConstraint() IMLImageSizeConstraint

func (i MLImageConstraint) Type() MLFeatureType

type MLImageConstraintClass ¶

type MLImageConstraintClass struct {
	// contains filtered or unexported fields
}

func GetMLImageConstraintClass() MLImageConstraintClass

func (mc MLImageConstraintClass) Alloc() MLImageConstraint

type MLImageSize ¶

type MLImageSize struct {
	objectivec.Object
}

The width and height of an image feature size.

Accessing an image size ¶

• MLImageSize.PixelsHigh: The height of an image feature in pixels.
• MLImageSize.PixelsWide: The width of an image feature in pixels.

See: https://developer.apple.com/documentation/CoreML/MLImageSize

func MLImageSizeFromID(id objc.ID) MLImageSize

func NewMLImageSize() MLImageSize

func (i MLImageSize) Autorelease() MLImageSize

func (i MLImageSize) EncodeWithCoder(coder foundation.INSCoder)

func (i MLImageSize) EnumeratedImageSizes() IMLImageSize

func (i MLImageSize) Init() MLImageSize

func (i MLImageSize) PixelsHigh() int

func (i MLImageSize) PixelsWide() int

func (i MLImageSize) SetEnumeratedImageSizes(value IMLImageSize)

type MLImageSizeClass ¶

type MLImageSizeClass struct {
	// contains filtered or unexported fields
}

func GetMLImageSizeClass() MLImageSizeClass

func (mc MLImageSizeClass) Alloc() MLImageSize

type MLImageSizeConstraint ¶

type MLImageSizeConstraint struct {
	objectivec.Object
}

A list or range of sizes that augment an image constraint’s default size.

Overview ¶

You use an MLImageSizeConstraint to express what image sizes of an image feature a model will accept as input or produce as output.

Use MLImageSizeConstraint.Type to determine which properties describe what image sizes the model’s image feature expects as input or produces as output.

If `type` is:

- MLImageSizeConstraintType.range, the image feature accepts any image that has a width in MLImageSizeConstraint.PixelsWideRange and a height in MLImageSizeConstraint.PixelsHighRange. - MLImageSizeConstraintType.enumerated, the image feature accepts any image size listed in MLImageSizeConstraint.EnumeratedImageSizes. - MLImageSizeConstraintType.unspecified, the MLImageSizeConstraint instance is not configured and should be ignored. Instead, use the image feature’s default image size constraint, defined by MLImageSizeConstraint.PixelsWide and MLImageSizeConstraint.PixelsHigh.

[media-3027121]

Determining relevant constraints ¶

• MLImageSizeConstraint.Type: Indicator of which properties to inspect for this image size constraint.

Accessing the image size ranges ¶

• MLImageSizeConstraint.PixelsWideRange: The range of widths a model’s image feature accepts as input or produces as output.
• MLImageSizeConstraint.PixelsHighRange: The range of heights a model’s image feature accepts as input or produces as output.

Accessing the enumerated image sizes ¶

• MLImageSizeConstraint.EnumeratedImageSizes: An array of image sizes a model’s image feature accepts as input or produces as output.

See: https://developer.apple.com/documentation/CoreML/MLImageSizeConstraint

func MLImageSizeConstraintFromID(id objc.ID) MLImageSizeConstraint

func NewMLImageSizeConstraint() MLImageSizeConstraint

func (i MLImageSizeConstraint) Autorelease() MLImageSizeConstraint

func (i MLImageSizeConstraint) EncodeWithCoder(coder foundation.INSCoder)

func (i MLImageSizeConstraint) EnumeratedImageSizes() []MLImageSize

func (i MLImageSizeConstraint) Init() MLImageSizeConstraint

func (i MLImageSizeConstraint) PixelsHigh() int

func (i MLImageSizeConstraint) PixelsHighRange() foundation.NSRange

func (i MLImageSizeConstraint) PixelsWide() int

func (i MLImageSizeConstraint) PixelsWideRange() foundation.NSRange

func (i MLImageSizeConstraint) SetPixelsHigh(value int)

func (i MLImageSizeConstraint) SetPixelsWide(value int)

func (i MLImageSizeConstraint) SetSizeConstraint(value IMLImageSizeConstraint)

func (i MLImageSizeConstraint) SizeConstraint() IMLImageSizeConstraint

func (i MLImageSizeConstraint) Type() MLImageSizeConstraintType

type MLImageSizeConstraintClass ¶

type MLImageSizeConstraintClass struct {
	// contains filtered or unexported fields
}