Crate objc2_metal_performance_shaders 

Bindings to the MetalPerformanceShaders framework

See Apple’s docs and the general docs on framework crates for more information.

Structs

MPSAccelerationStructureDeprecatedMPSAccelerationStructure and MPSCore and MPSKernel

A data structure built over geometry used to accelerate ray tracing

MPSAccelerationStructureGroupDeprecatedMPSAccelerationStructureGroup

A group of acceleration structures which may be used together in an instance acceleration structure.

MPSAccelerationStructureStatusDeprecatedMPSAccelerationStructure

Possible values of the acceleration structure status property

MPSAccelerationStructureUsageDeprecatedMPSAccelerationStructure

Options describing how an acceleration structure will be used

MPSAliasingStrategyMPSCoreTypes

Apple’s documentation

MPSAlphaTypeMPSImageTypes

Apple’s documentation

MPSBinaryImageKernelMPSCore and MPSImageKernel and MPSKernel

Dependencies: This depends on Metal.framework

MPSBoundingBoxIntersectionTestTypeDeprecatedMPSRayIntersector

Options for the MPSRayIntersector bounding box intersection test type property

MPSCNNAddMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNAddGradientMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNArithmeticMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNArithmeticGradientMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNArithmeticGradientStateMPSCNNMath and MPSCore and MPSNNGradientState and MPSState

Dependencies: This depends on Metal.framework.

MPSCNNBatchNormalizationMPSCNNBatchNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBatchNormalizationFlagsMPSNeuralNetworkTypes

Apple’s documentation

MPSCNNBatchNormalizationGradientMPSCNNBatchNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBatchNormalizationGradientNodeMPSNNGraphNodes

A node representing batch normalization gradient for training

MPSCNNBatchNormalizationNodeMPSNNGraphNodes

A node representing batch normalization for inference or training

MPSCNNBatchNormalizationStateMPSCNNBatchNormalization and MPSCore and MPSNNGradientState and MPSState

MPSCNNBatchNormalizationState encapsulates the data necessary to execute batch normalization.

MPSCNNBatchNormalizationStatisticsMPSCNNBatchNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBatchNormalizationStatisticsGradientMPSCNNBatchNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBinaryConvolutionMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBinaryConvolutionFlagsMPSNeuralNetworkTypes

Apple’s documentation

MPSCNNBinaryConvolutionNodeMPSNNGraphNodes

A MPSNNFilterNode representing a MPSCNNBinaryConvolution kernel

MPSCNNBinaryConvolutionTypeMPSNeuralNetworkTypes

Apple’s documentation

MPSCNNBinaryFullyConnectedMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNBinaryFullyConnectedNodeMPSNNGraphNodes

A MPSNNFilterNode representing a MPSCNNBinaryFullyConnected kernel

MPSCNNBinaryKernelMPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNConvolutionMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNConvolutionDescriptorMPSCNNConvolution

Dependencies: This depends on Metal.framework

MPSCNNConvolutionFlagsMPSNeuralNetworkTypes

Apple’s documentation

MPSCNNConvolutionGradientMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNConvolutionGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNConvolutionGradientOptionMPSCNNConvolution

Apple’s documentation

MPSCNNConvolutionGradientStateMPSCNNConvolution and MPSCore and MPSNNGradientState and MPSState

The MPSCNNConvolutionGradientState is returned by resultStateForSourceImage:sourceStates method on MPSCNNConvolution object. Note that resultStateForSourceImage:sourceStates:destinationImage creates the object on autoreleasepool. It will be consumed by MPSCNNConvolutionGradient. This is also used by MPSCNNConvolutionTranspose encode call that returns MPSImage on left hand side to correctly size the destination. Note that state objects are not usable across batches i.e. when batch is done you should nuke the state object and create new one for next batch.

MPSCNNConvolutionGradientStateNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNConvolutionNodeMPSNNGraphNodes

A MPSNNFilterNode representing a MPSCNNConvolution kernel

MPSCNNConvolutionTransposeMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNConvolutionTransposeGradientMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNConvolutionTransposeGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNConvolutionTransposeGradientStateMPSCNNConvolution and MPSCore and MPSNNGradientState and MPSState

The MPSCNNConvolutionTransposeGradientState is returned by resultStateForSourceImage:sourceStates method on MPSCNNConvolutionTranspose object. Note that resultStateForSourceImage:sourceStates:destinationImage creates the object on autoreleasepool. It will be consumed by MPSCNNConvolutionTransposeGradient. It contains reference to MPSCNNConvolutionGradientState object that connects MPSCNNConvolution and its corresponding MPSCNNConvolutionTranspose in forward pass of autoencoder. In an autoencoder forward pass, MPSCNNConvolutionGradientState is produced by MPSCNNConvolution object and is used by corresponding MPSCNNConvolutionTraspose of forward pass that “undo” the corresponding MPSCNNConvolution. It is used to correctly size destination image that is returned on left hand side by encode call MPSCNNConvolutionTranspose as well as automatically set kernelOffsetX/Y on MPSCNNConvolutionTranspose using the offset and other properties of corresponding MPSCNNConvolution object. During training, same MPSCNNConvolutionGradientState object will be consumed by MPSCNNConvolutionGradient object and the MPSCNNConvolutionTransposeGradientState produced by MPSCNNConvolutionTranspose’s resultStateForSourceImage:sourceStates:destinationImage will be consumed by MPSCNNConvolutionTransposeGradient object

MPSCNNConvolutionTransposeGradientStateNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNConvolutionTransposeNodeMPSNNGraphNodes

A MPSNNFilterNode representing a MPSCNNConvolutionTranspose kernel

MPSCNNConvolutionWeightsAndBiasesStateMPSCNNConvolution and MPSCore and MPSState

The MPSCNNConvolutionWeightsAndBiasesState is returned by exportWeightsAndBiasesWithCommandBuffer: method on MPSCNNConvolution object. This is mainly used for GPU side weights/biases update process. During training, application can keep a copy of weights, velocity, momentum MTLBuffers in its data source, update the weights (in-place or out of place) with gradients obtained from MPSCNNConvolutionGradientState and call [MPSCNNConvolution reloadWeightsAndBiasesWithCommandBuffer] with resulting updated MTLBuffer. If application does not want to keep a copy of weights/biases, it can call [MPSCNNConvolution exportWeightsAndBiasesWithCommandBuffer:] to get the current weights from convolution itself, do the updated and call reloadWithCommandBuffer.

MPSCNNConvolutionWeightsLayoutMPSCNNConvolution

Apple’s documentation

MPSCNNCrossChannelNormalizationMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNCrossChannelNormalizationGradientMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNCrossChannelNormalizationGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNCrossChannelNormalizationNodeMPSNNGraphNodes

Node representing MPSCNNCrossChannelNormalization

MPSCNNDepthWiseConvolutionDescriptorMPSCNNConvolution

MPSCNNDepthWiseConvolutionDescriptor can be used to create MPSCNNConvolution object that does depthwise convolution

MPSCNNDilatedPoolingMaxMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNDilatedPoolingMaxGradientMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNDilatedPoolingMaxGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNDilatedPoolingMaxNodeMPSNNGraphNodes

A node for a MPSCNNDilatedPooling kernel

MPSCNNDivideMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNDropoutMPSCNNDropout and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNDropoutGradientMPSCNNDropout and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNDropoutGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNDropoutGradientStateMPSCNNDropout and MPSCore and MPSNNGradientState and MPSState

Dependencies: This depends on Metal.framework.

MPSCNNDropoutNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNFullyConnectedMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNFullyConnectedGradientMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNFullyConnectedGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNFullyConnectedNodeMPSNNGraphNodes

A MPSNNFilterNode representing a MPSCNNFullyConnected kernel

MPSCNNGradientKernelMPSCNNKernel and MPSCore and MPSKernel

Gradient kernels are the backwards pass of a MPSCNNKernel used during training to calculate gradient back propagation. These take as arguments the gradient result from the next filter and the source image for the forward version of the filter. There is also a MPSNNGradientState passed from MPSCNNKernel to MPSCNNGradientKernel that contains information about the MPSCNNKernel parameters at the time it encoded and possibly also additional MTLResources to enable it to do its job.

MPSCNNGroupNormalizationMPSCNNGroupNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNGroupNormalizationGradientMPSCNNGroupNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNGroupNormalizationGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNGroupNormalizationGradientStateMPSCNNGroupNormalization and MPSCore and MPSNNGradientState and MPSState

Dependencies: This depends on Metal.framework

MPSCNNGroupNormalizationNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNInstanceNormalizationMPSCNNInstanceNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNInstanceNormalizationGradientMPSCNNInstanceNormalization and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNInstanceNormalizationGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNInstanceNormalizationGradientStateMPSCNNInstanceNormalization and MPSCore and MPSNNGradientState and MPSState

Dependencies: This depends on Metal.framework

MPSCNNInstanceNormalizationNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNKernelMPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNLocalContrastNormalizationMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNLocalContrastNormalizationGradientMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNLocalContrastNormalizationGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNLocalContrastNormalizationNodeMPSNNGraphNodes

Node representing MPSCNNLocalContrastNormalization

MPSCNNLogSoftMaxMPSCNNKernel and MPSCNNSoftMax and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNLogSoftMaxGradientMPSCNNKernel and MPSCNNSoftMax and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNLogSoftMaxGradientNodeMPSNNGraphNodes

Node representing a MPSCNNLogSoftMaxGradient kernel

MPSCNNLogSoftMaxNodeMPSNNGraphNodes

Node representing a MPSCNNLogSoftMax kernel

MPSCNNLossMPSCNNKernel and MPSCNNLoss and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNLossDataDescriptorMPSCNNLoss

Dependencies: This depends on Metal.framework.

MPSCNNLossDescriptorMPSCNNLoss

Dependencies: This depends on Metal.framework.

MPSCNNLossLabelsMPSCNNLoss and MPSCore and MPSState

Dependencies: This depends on Metal.framework.

MPSCNNLossNodeMPSNNGraphNodes

This node calculates loss information during training typically immediately after the inference portion of network evaluation is performed. The result image of the loss operations is typically the first gradient image to be comsumed by the gradient passes that work their way back up the graph. In addition, the node will update the loss image in the MPSNNLabels with the desired estimate of correctness.

MPSCNNLossTypeMPSCNNTypes

Apple’s documentation

MPSCNNMultiaryKernelMPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNMultiplyMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNMultiplyGradientMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNNeuronMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronAbsoluteMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronAbsoluteNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronAbsolute kernel

MPSCNNNeuronELUMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronELUNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronELU kernel

MPSCNNNeuronExponentialMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNNeuronExponentialNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronExponential kernel

MPSCNNNeuronGeLUNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronGeLU kernel

MPSCNNNeuronGradientMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronGradientNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronGradient

MPSCNNNeuronHardSigmoidMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronHardSigmoidNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronHardSigmoid kernel

MPSCNNNeuronLinearMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronLinearNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronLinear kernel

MPSCNNNeuronLogarithmMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNNeuronLogarithmNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronLogarithm kernel

MPSCNNNeuronNodeMPSNNGraphNodes

virtual base class for MPSCNNNeuron nodes

MPSCNNNeuronPReLUMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronPReLUNodeMPSNNGraphNodes

A ReLU node with parameter a provided independently for each feature channel

MPSCNNNeuronPowerMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNNeuronPowerNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronPower kernel

MPSCNNNeuronReLUMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronReLUNMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronReLUNNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronReLUN kernel

MPSCNNNeuronReLUNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronReLU kernel

MPSCNNNeuronSigmoidMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronSigmoidNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronSigmoid kernel

MPSCNNNeuronSoftPlusMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronSoftPlusNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronSoftPlus kernel

MPSCNNNeuronSoftSignMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronSoftSignNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronSoftSign kernel

MPSCNNNeuronTanHMPSCNNKernel and MPSCNNNeuron and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNNeuronTanHNodeMPSNNGraphNodes

A node representing a MPSCNNNeuronTanH kernel

MPSCNNNeuronTypeMPSCNNNeuronType

Apple’s documentation

MPSCNNNormalizationGammaAndBetaStateMPSCNNNormalizationWeights and MPSCore and MPSState

A state which contains gamma and beta terms used to apply a scale and bias in either an MPSCNNInstanceNormalization or MPSCNNBatchNormalization operation.

MPSCNNNormalizationMeanAndVarianceStateMPSCNNBatchNormalization and MPSCore and MPSState

A state which contains mean and variance terms used to apply a normalization in a MPSCNNBatchNormalization operation.

MPSCNNNormalizationNodeMPSNNGraphNodes

virtual base class for CNN normalization nodes

MPSCNNPoolingMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingAverageMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingAverageGradientMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingAverageGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNPoolingAverageNodeMPSNNGraphNodes

A node representing a MPSCNNPoolingAverage kernel

MPSCNNPoolingGradientMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNPoolingL2NormMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingL2NormGradientMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingL2NormGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNPoolingL2NormNodeMPSNNGraphNodes

A node representing a MPSCNNPoolingL2Norm kernel

MPSCNNPoolingMaxMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingMaxGradientMPSCNNKernel and MPSCNNPooling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNPoolingMaxGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNPoolingMaxNodeMPSNNGraphNodes

A node representing a MPSCNNPoolingMax kernel

MPSCNNPoolingNodeMPSNNGraphNodes

A node for a MPSCNNPooling kernel

MPSCNNReductionTypeMPSCNNTypes

Apple’s documentation

MPSCNNSoftMaxMPSCNNKernel and MPSCNNSoftMax and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNSoftMaxGradientMPSCNNKernel and MPSCNNSoftMax and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNSoftMaxGradientNodeMPSNNGraphNodes

Node representing a MPSCNNSoftMaxGradient kernel

MPSCNNSoftMaxNodeMPSNNGraphNodes

Node representing a MPSCNNSoftMax kernel

MPSCNNSpatialNormalizationMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNSpatialNormalizationGradientMPSCNNKernel and MPSCNNNormalization and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNSpatialNormalizationGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSCNNSpatialNormalizationNodeMPSNNGraphNodes

Node representing MPSCNNSpatialNormalization

MPSCNNSubPixelConvolutionDescriptorMPSCNNConvolution

MPSCNNSubPixelConvolutionDescriptor can be used to create MPSCNNConvolution object that does sub pixel upsamling and reshaping opeartion as described in http://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Shi_Real-Time_Single_Image_CVPR_2016_paper.pdf

MPSCNNSubtractMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNSubtractGradientMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNUpsamplingMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNUpsamplingBilinearMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNUpsamplingBilinearGradientMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNUpsamplingBilinearGradientNodeMPSNNGraphNodes

Node representing a MPSCNNUpsamplingBilinear kernel

MPSCNNUpsamplingBilinearNodeMPSNNGraphNodes

Node representing a MPSCNNUpsamplingBilinear kernel

MPSCNNUpsamplingGradientMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSCNNUpsamplingNearestMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNUpsamplingNearestGradientMPSCNNKernel and MPSCNNUpsampling and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSCNNUpsamplingNearestGradientNodeMPSNNGraphNodes

Node representing a MPSCNNUpsamplingNearest kernel

MPSCNNUpsamplingNearestNodeMPSNNGraphNodes

Node representing a MPSCNNUpsamplingNearest kernel

MPSCNNWeightsQuantizationTypeMPSCNNConvolution

Apple’s documentation

MPSCNNYOLOLossMPSCNNKernel and MPSCNNLoss and MPSCore and MPSKernel

Apple’s documentation

MPSCNNYOLOLossDescriptorMPSCNNLoss

Dependencies: This depends on Metal.framework.

MPSCNNYOLOLossNodeMPSNNGraphNodes

This node calculates loss information during training typically immediately after the inference portion of network evaluation is performed. The result image of the loss operations is typically the first gradient image to be comsumed by the gradient passes that work their way back up the graph. In addition, the node will update the loss image in the MPSNNLabels with the desired estimate of correctness.

MPSCommandBufferMPSCommandBuffer

Dependencies: This depends on Metal.framework

MPSCustomKernelArgumentCountMPSKernelTypes

Apple’s documentation

MPSCustomKernelIndexMPSKernelTypes

Apple’s documentation

MPSDataLayoutMPSImage

Apple’s documentation

MPSDataTypeMPSCoreTypes

Apple’s documentation

MPSDeviceCapsValuesMPSKernelTypes

Apple’s documentation

MPSDeviceOptions

Apple’s documentation

MPSDimensionSliceMPSCoreTypes

Describes a sub-region of an array dimension

MPSFloatDataTypeBitMPSCoreTypes

Apple’s documentation

MPSFloatDataTypeShiftMPSCoreTypes

Apple’s documentation

MPSGRUDescriptorMPSRNNLayer

Dependencies: This depends on Metal.framework

MPSImageMPSImage

Dependencies: This depends on Metal.framework

MPSImageAddMPSCore and MPSImageKernel and MPSImageMath and MPSKernel

Dependencies: This depends on Metal.framework.

MPSImageAreaMaxMPSCore and MPSImageKernel and MPSImageMorphology and MPSKernel

The MPSImageAreaMax kernel finds the maximum pixel value in a rectangular region centered around each pixel in the source image. If there are multiple channels in the source image, each channel is processed independently. The edgeMode property is assumed to always be MPSImageEdgeModeClamp for this filter.

MPSImageAreaMinMPSCore and MPSImageKernel and MPSImageMorphology and MPSKernel

The MPSImageAreaMin finds the minimum pixel value in a rectangular region centered around each pixel in the source image. If there are multiple channels in the source image, each channel is processed independently. It has the same methods as MPSImageAreaMax The edgeMode property is assumed to always be MPSImageEdgeModeClamp for this filter.

MPSImageArithmeticMPSCore and MPSImageKernel and MPSImageMath and MPSKernel

Dependencies: This depends on Metal.framework.

MPSImageBilinearScaleMPSCore and MPSImageKernel and MPSImageResampling and MPSKernel

Resize an image and / or change its aspect ratio

MPSImageBoxMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageBox convolves an image with given filter of odd width and height. The kernel elements all have equal weight, achieving a blur effect. (Each result is the unweighted average of the surrounding pixels.) This allows for much faster algorithms, espcially for larger blur radii. The box height and width must be odd numbers. The box blur is a separable filter. The implementation is aware of this and will act accordingly to give best performance for multi-dimensional blurs.

MPSImageCannyMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageCanny implements the Canny edge detection algorithm. When the color model of the source and destination textures match, the filter is applied to each channel seperately. If the destination is monochrome but source multichannel, the source will be converted to grayscale using the linear gray color transform vector (v). Luminance = v[0] * pixel.x + v[1] * pixel.y + v[2] * pixel.z;

MPSImageConversionMPSCore and MPSImageConversion and MPSImageKernel and MPSKernel

The MPSImageConversion filter performs a conversion from source to destination

MPSImageConvolutionMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageConvolution convolves an image with given filter of odd width and height. The center of the kernel aligns with the MPSImageConvolution.offset. That is, the position of the top left corner of the area covered by the kernel is given by MPSImageConvolution.offset - {kernel_width>>1, kernel_height>>1, 0}

MPSImageCoordinateMPSCoreTypes

A unsigned coordinate with x, y and channel components

MPSImageCopyToMatrixMPSCore and MPSImageCopy and MPSKernel

The MPSImageCopyToMatrix copies image data to a MPSMatrix. The image data is stored in a row of a matrix. The dataLayout specifies the order in which the feature channels in the MPSImage get stored in the matrix. If MPSImage stores a batch of images, the images are copied into multiple rows, one row per image.

MPSImageDescriptorMPSImage

Dependencies: This depends on Metal.framework

MPSImageDilateMPSCore and MPSImageKernel and MPSImageMorphology and MPSKernel

The MPSImageDilate finds the maximum pixel value in a rectangular region centered around each pixel in the source image. It is like the MPSImageAreaMax, except that the intensity at each position is calculated relative to a different value before determining which is the maximum pixel value, allowing for shaped, non-rectangular morphological probes.

MPSImageDivideMPSCore and MPSImageKernel and MPSImageMath and MPSKernel

Dependencies: This depends on Metal.framework.

MPSImageEDLinesMPSCore and MPSImageEDLines and MPSKernel

The MPSImageEDLInes class implements the EDLines line segmenting algorithm using edge-drawing (ED) described here https://ieeexplore.ieee.org/document/6116138

MPSImageEdgeModeMPSCoreTypes

Apple’s documentation

MPSImageErodeMPSCore and MPSImageKernel and MPSImageMorphology and MPSKernel

The MPSImageErode filter finds the minimum pixel value in a rectangular region centered around each pixel in the source image. It is like the MPSImageAreaMin, except that the intensity at each position is calculated relative to a different value before determining which is the maximum pixel value, allowing for shaped, non-rectangular morphological probes.

MPSImageEuclideanDistanceTransformMPSCore and MPSImageDistanceTransform and MPSImageKernel and MPSKernel

Perform a Euclidean Distance Transform

MPSImageFeatureChannelFormatMPSCoreTypes

Apple’s documentation

MPSImageFindKeypointsMPSCore and MPSImageKeypoint and MPSKernel

The MPSImageFindKeypoints kernel is used to find a list of keypoints whose values are >= minimumPixelThresholdValue in MPSImageKeypointRangeInfo. The keypoints are generated for a specified region in the image. The pixel format of the source image must be MTLPixelFormatR8Unorm.

MPSImageGaussianBlurMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageGaussianBlur convolves an image with gaussian of given sigma in both x and y direction.

MPSImageGaussianPyramidMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

A Gaussian image pyramid is constructed as follows: The mipmap level zero is the source of the operation and is left untouched and the subsequent mipmap levels are constructed from it recursively:

MPSImageGuidedFilterMPSCore and MPSImageGuidedFilter and MPSKernel

Perform Guided Filter to produce a coefficients image The filter is broken into two stages:

MPSImageHistogramMPSCore and MPSImageHistogram and MPSKernel

The MPSImageHistogram computes the histogram of an image.

MPSImageHistogramEqualizationMPSCore and MPSImageHistogram and MPSImageKernel and MPSKernel

The MPSImageHistogramEqualization performs equalizes the histogram of an image. The process is divided into three steps.

MPSImageHistogramSpecificationMPSCore and MPSImageHistogram and MPSImageKernel and MPSKernel

The MPSImageHistogramSpecification performs a histogram specification operation on an image. It is a generalized version of histogram equalization operation. The histogram specificaiton filter converts the image so that its histogram matches the desired histogram.

MPSImageIntegralMPSCore and MPSImageIntegral and MPSImageKernel and MPSKernel

The MPSImageIntegral calculates the sum of pixels over a specified region in the image. The value at each position is the sum of all pixels in a source image rectangle, sumRect:

MPSImageIntegralOfSquaresMPSCore and MPSImageIntegral and MPSImageKernel and MPSKernel

The MPSImageIntegralOfSquares calculates the sum of squared pixels over a specified region in the image. The value at each position is the sum of all squared pixels in a source image rectangle, sumRect:

MPSImageKeypointRangeInfoMPSImageKeypoint

Specifies information to find the keypoints in an image.

MPSImageLanczosScaleMPSCore and MPSImageKernel and MPSImageResampling and MPSKernel

Resize an image and / or change its aspect ratio

MPSImageLaplacianMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageLaplacian is an optimized variant of the MPSImageConvolution filter provided primarily for ease of use. This filter uses an optimized convolution filter with a 3 x 3 kernel with the following weights: [ 0 1 0 1 -4 1 0 1 0 ]

MPSImageLaplacianPyramidMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

Laplacian pyramid levels are constructed as difference between the current source level and 2x interpolated version of the half-resolution source level immediately above it.

MPSImageLaplacianPyramidAddMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

Apple’s documentation

MPSImageLaplacianPyramidSubtractMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

Apple’s documentation

MPSImageMedianMPSCore and MPSImageKernel and MPSImageMedian and MPSKernel

The MPSImageMedian applies a median filter to an image. A median filter finds the median color value for each channel within a kernelDiameter x kernelDiameter window surrounding the pixel of interest. It is a common means of noise reduction and also as a smoothing filter with edge preserving qualities.

MPSImageMultiplyMPSCore and MPSImageKernel and MPSImageMath and MPSKernel

Dependencies: This depends on Metal.framework.

MPSImageNormalizedHistogramMPSCore and MPSImageHistogram and MPSKernel

The MPSImageNormalizedHistogram computes the normalized histogram of an image. The minimum and maximum pixel values for a given region of an image are first computed. The max(computed minimum pixel value, MPSImageHistogramInfo.minPixelValue) and the min(computed maximum pixel value, MPSImageHistogramInfo.maxPixelValue) are used to compute the normalized histogram.

MPSImagePyramidMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImagePyramid is a base class for creating different kinds of pyramid images

MPSImageReadWriteParamsMPSImage

these parameters are passed in to allow user to read/write to a particular set of featureChannels in an MPSImage

MPSImageReduceColumnMaxMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceColumnMax performs a reduction operation returning the maximum value for each column of an image

MPSImageReduceColumnMeanMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceColumnMean performs a reduction operation returning the mean value for each column of an image

MPSImageReduceColumnMinMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceColumnMin performs a reduction operation returning the mininmum value for each column of an image

MPSImageReduceColumnSumMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceColumnSum performs a reduction operation returning the sum for each column of an image

MPSImageReduceRowMaxMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceRowMax performs a reduction operation returning the maximum value for each row of an image

MPSImageReduceRowMeanMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceRowMean performs a reduction operation returning the mean value for each row of an image

MPSImageReduceRowMinMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceRowMin performs a reduction operation returning the mininmum value for each row of an image

MPSImageReduceRowSumMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduceRowSum performs a reduction operation returning the sum for each row of an image

MPSImageReduceUnaryMPSCore and MPSImageKernel and MPSImageReduce and MPSKernel

The MPSImageReduce performs a reduction operation The reduction operations supported are:

MPSImageRegionMPSCoreTypes

A rectangular subregion of a MPSImage

MPSImageScaleMPSCore and MPSImageKernel and MPSImageResampling and MPSKernel

Resize an image and / or change its aspect ratio

MPSImageSobelMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The MPSImageSobel implements the Sobel filter. When the color model (e.g. RGB, two-channel, grayscale, etc.) of source and destination textures match, the filter is applied to each channel separately. If the destination is monochrome (single channel) but source multichannel, the pixel values are converted to grayscale before applying Sobel operator using the linear gray color transform vector (v).

MPSImageStatisticsMeanMPSCore and MPSImageKernel and MPSImageStatistics and MPSKernel

The MPSImageStatisticsMean computes the mean for a given region of an image.

MPSImageStatisticsMeanAndVarianceMPSCore and MPSImageKernel and MPSImageStatistics and MPSKernel

The MPSImageStatisticsMeanAndVariance computes the mean and variance for a given region of an image. The mean and variance values are written to the destination image at the following pixel locations:

MPSImageStatisticsMinAndMaxMPSCore and MPSImageKernel and MPSImageStatistics and MPSKernel

The MPSImageStatisticsMinAndMax computes the minimum and maximum pixel values for a given region of an image. The min and max values are written to the destination image at the following pixel locations:

MPSImageSubtractMPSCore and MPSImageKernel and MPSImageMath and MPSKernel

Dependencies: This depends on Metal.framework.

MPSImageTentMPSCore and MPSImageConvolution and MPSImageKernel and MPSKernel

The box filter, while fast, may yield square-ish looking blur effects. However, multiple passes of the box filter tend to smooth out with each additional pass. For example, two 3-wide box blurs produces the same effective convolution as a 5-wide tent blur:

MPSImageThresholdBinaryMPSCore and MPSImageKernel and MPSImageThreshold and MPSKernel

The MPSThreshold filter applies a fixed-level threshold to each pixel in the image. The threshold functions convert a single channel image to a binary image. If the input image is not a single channel image, convert the inputimage to a single channel luminance image using the linearGrayColorTransform and then apply the threshold. The ThresholdBinary function is: destinationPixelValue = sourcePixelValue > thresholdValue ? maximumValue : 0

MPSImageThresholdBinaryInverseMPSCore and MPSImageKernel and MPSImageThreshold and MPSKernel

The MPSImageThresholdBinaryInverse filter applies a fixed-level threshold to each pixel in the image. The threshold functions convert a single channel image to a binary image. If the input image is not a single channel image, convert the inputimage to a single channel luminance image using the linearGrayColorTransform and then apply the threshold. The ThresholdBinaryInverse function is: destinationPixelValue = sourcePixelValue > thresholdValue ? 0 : maximumValue

MPSImageThresholdToZeroMPSCore and MPSImageKernel and MPSImageThreshold and MPSKernel

The MPSImageThresholdToZero filter applies a fixed-level threshold to each pixel in the image. The threshold functions convert a single channel image to a binary image. If the input image is not a single channel image, convert the inputimage to a single channel luminance image using the linearGrayColorTransform and then apply the threshold. The ThresholdToZero function is: destinationPixelValue = sourcePixelValue > thresholdValue ? sourcePixelValue : 0

MPSImageThresholdToZeroInverseMPSCore and MPSImageKernel and MPSImageThreshold and MPSKernel

The MPSImageThresholdToZeroInverse filter applies a fixed-level threshold to each pixel in the image. The threshold functions convert a single channel image to a binary image. If the input image is not a single channel image, convert the inputimage to a single channel luminance image using the linearGrayColorTransform and then apply the threshold. The ThresholdToZeroINverse function is: destinationPixelValue = sourcePixelValue > thresholdValue ? 0 : sourcePixelValue

MPSImageThresholdTruncateMPSCore and MPSImageKernel and MPSImageThreshold and MPSKernel

The MPSImageThresholdTruncate filter applies a fixed-level threshold to each pixel in the image: The threshold functions convert a single channel image to a binary image. If the input image is not a single channel image, convert the inputimage to a single channel luminance image using the linearGrayColorTransform and then apply the threshold. The ThresholdTruncate function is: destinationPixelValue = sourcePixelValue > thresholdValue ? thresholdValue : sourcePixelValue

MPSImageTransposeMPSCore and MPSImageKernel and MPSImageTranspose and MPSKernel

The MPSImageTranspose transposes an image

MPSImageTypeMPSKernelTypes

Apple’s documentation

MPSInstanceAccelerationStructureDeprecatedMPSAccelerationStructure and MPSCore and MPSInstanceAccelerationStructure and MPSKernel

An acceleration structure built over instances of other acceleration structures

MPSIntegerDivisionParamsMPSKernelTypes

Apple’s documentation

MPSIntersectionDataTypeMPSRayIntersector

Intersection data type options

MPSIntersectionDistanceMPSRayIntersectorTypes

Returned intersection result which contains the distance from the ray origin to the intersection point

MPSIntersectionDistancePrimitiveIndexMPSRayIntersectorTypes

Intersection result which contains the distance from the ray origin to the intersection point and the index of the intersected primitive

MPSIntersectionDistancePrimitiveIndexBufferIndexMPSRayIntersectorTypes

Intersection result which contains the distance from the ray origin to the intersection point, the index of the intersected primitive, and the polygon buffer index of the intersected primitive.

MPSIntersectionDistancePrimitiveIndexBufferIndexInstanceIndexMPSRayIntersectorTypes

Intersection result which contains the distance from the ray origin to the intersection point, the index of the intersected primitive, the polygon buffer index of the intersected primitive, and the index of the intersected instance.

MPSIntersectionDistancePrimitiveIndexInstanceIndexMPSRayIntersectorTypes

Intersection result which contains the distance from the ray origin to the intersection point, the index of the intersected primitive, and the index of the intersected instance.

MPSIntersectionTypeMPSRayIntersector

Options for the MPSRayIntersector intersection type property

MPSKernelMPSKernel

Dependencies: This depends on Metal.framework

MPSKernelOptionsMPSCoreTypes

Apple’s documentation

MPSKeyedUnarchiverMPSKeyedUnarchiver

A NSKeyedArchiver that supports the MPSDeviceProvider protocol for MPSKernel decoding

MPSLSTMDescriptorMPSRNNLayer

Dependencies: This depends on Metal.framework

MPSMatrixMPSMatrix

Dependencies: This depends on Metal.framework

MPSMatrixBatchNormalizationMPSCore and MPSKernel and MPSMatrix and MPSMatrixBatchNormalization and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixBatchNormalizationGradientMPSCore and MPSKernel and MPSMatrix and MPSMatrixBatchNormalization and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixBinaryKernelMPSCore and MPSKernel and MPSMatrixTypes

Dependencies: This depends on Metal.framework

MPSMatrixCopyMPSCore and MPSKernel and MPSMatrixCombination

Apple’s documentation

MPSMatrixCopyDescriptorMPSMatrixCombination

A list of copy operations

MPSMatrixCopyOffsetsMPSMatrixCombination

A description of each copy operation

MPSMatrixCopyToImageMPSCore and MPSImageCopy and MPSKernel

The MPSMatrixCopyToImage copies matrix data to a MPSImage. The operation is the reverse of MPSImageCopyToMatrix.

MPSMatrixDecompositionCholeskyMPSCore and MPSKernel and MPSMatrixDecomposition and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixDecompositionLUMPSCore and MPSKernel and MPSMatrixDecomposition and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixDecompositionStatusMPSMatrixDecomposition

Apple’s documentation

MPSMatrixDescriptorMPSMatrix

Dependencies: This depends on Metal.framework

MPSMatrixFindTopKMPSCore and MPSKernel and MPSMatrixFindTopK and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixFullyConnectedMPSCore and MPSKernel and MPSMatrix and MPSMatrixFullyConnected and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixFullyConnectedGradientMPSCore and MPSKernel and MPSMatrix and MPSMatrixFullyConnected and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixLogSoftMaxMPSCore and MPSKernel and MPSMatrixSoftMax and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixLogSoftMaxGradientMPSCore and MPSKernel and MPSMatrixSoftMax and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixMultiplicationMPSCore and MPSKernel and MPSMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSMatrixNeuronMPSCore and MPSKernel and MPSMatrix and MPSMatrixNeuron and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixNeuronGradientMPSCore and MPSKernel and MPSMatrix and MPSMatrixNeuron and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixOffsetMPSKernelTypes

Specifies a row and column offset into an MPSMatrix.

MPSMatrixRandomMPSCore and MPSKernel and MPSMatrixRandom

Kernels that implement random number generation.

MPSMatrixRandomDistributionMPSMatrixRandom

Apple’s documentation

MPSMatrixRandomDistributionDescriptorMPSMatrixRandom

Dependencies: This depends on Metal.framework

MPSMatrixRandomMTGP32MPSCore and MPSKernel and MPSMatrixRandom

Generates random numbers using a Mersenne Twister algorithm suitable for GPU execution. It uses a period of 2**11214. For further details see: Mutsuo Saito. A Variant of Mersenne Twister Suitable for Graphic Processors. arXiv:1005.4973

MPSMatrixRandomPhiloxMPSCore and MPSKernel and MPSMatrixRandom

Generates random numbers using a counter based algorithm. For further details see: John K. Salmon, Mark A. Moraes, Ron O. Dror, and David E. Shaw. Parallel Random Numbers: As Easy as 1, 2, 3.

MPSMatrixSoftMaxMPSCore and MPSKernel and MPSMatrixSoftMax and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixSoftMaxGradientMPSCore and MPSKernel and MPSMatrixSoftMax and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixSolveCholeskyMPSCore and MPSKernel and MPSMatrixSolve and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixSolveLUMPSCore and MPSKernel and MPSMatrixSolve and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixSolveTriangularMPSCore and MPSKernel and MPSMatrixSolve and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSMatrixSumMPSCore and MPSKernel and MPSMatrixSum

Dependencies: This depends on Metal.framework

MPSMatrixUnaryKernelMPSCore and MPSKernel and MPSMatrixTypes

Dependencies: This depends on Metal.framework

MPSMatrixVectorMultiplicationMPSCore and MPSKernel and MPSMatrixMultiplication and MPSMatrixTypes

Dependencies: This depends on Metal.framework.

MPSNDArrayMPSNDArray

A MPSNDArray object is a MTLBuffer based storage container for multi-dimensional data.

MPSNDArrayAffineInt4DequantizeMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayQuantizedMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSNDArrayAffineQuantizationDescriptorMPSNDArrayQuantization

Dependencies: This depends on Metal.framework.

MPSNDArrayBinaryKernelMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayBinaryPrimaryGradientKernelMPSCore and MPSKernel and MPSNDArrayKernel

Dependencies: This depends on Metal.framework.

MPSNDArrayBinarySecondaryGradientKernelMPSCore and MPSKernel and MPSNDArrayKernel

Dependencies: This depends on Metal.framework.

MPSNDArrayDescriptorMPSNDArray

Dependencies: This depends on Metal.framework

MPSNDArrayGatherMPSCore and MPSKernel and MPSNDArrayGather and MPSNDArrayKernel

Dependencies: This depends on Metal.framework.

MPSNDArrayGatherGradientMPSCore and MPSKernel and MPSNDArrayGather and MPSNDArrayKernel

Dependencies: This depends on Metal.framework.

MPSNDArrayGatherGradientStateMPSCore and MPSNDArrayGather and MPSNDArrayGradientState and MPSState

at the time an -encode call was made.

MPSNDArrayGradientStateMPSCore and MPSNDArrayGradientState and MPSState

at the time an -encode call was made. The contents are opaque.

MPSNDArrayIdentityMPSCore and MPSKernel and MPSNDArrayIdentity and MPSNDArrayKernel

Dependencies: This depends on Metal.framework.

MPSNDArrayLUTDequantizeMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayQuantizedMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSNDArrayLUTQuantizationDescriptorMPSNDArrayQuantization

Dependencies: This depends on Metal.framework.

MPSNDArrayMatrixMultiplicationMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSNDArrayMultiaryBaseMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayMultiaryGradientKernelMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayMultiaryKernelMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayOffsetsMPSNDArrayTypes

Apple’s documentation

MPSNDArrayQuantizationDescriptorMPSNDArrayQuantization

Dependencies: This depends on Metal.framework.

MPSNDArrayQuantizationSchemeMPSNDArrayQuantization

Apple’s documentation

MPSNDArrayQuantizedMatrixMultiplicationMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayMatrixMultiplication and MPSNDArrayQuantizedMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSNDArraySizesMPSNDArrayTypes

Apple’s documentation

MPSNDArrayStridedSliceMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayStridedSlice

Dependencies: This depends on Metal.framework.

MPSNDArrayStridedSliceGradientMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayStridedSlice

Dependencies: This depends on Metal.framework.

MPSNDArrayUnaryGradientKernelMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayUnaryKernelMPSCore and MPSKernel and MPSNDArrayKernel

Apple’s documentation

MPSNDArrayVectorLUTDequantizeMPSCore and MPSKernel and MPSNDArrayKernel and MPSNDArrayQuantizedMatrixMultiplication

Dependencies: This depends on Metal.framework.

MPSNNAdditionGradientNodeMPSNNGraphNodes

returns gradient for either primary or secondary source image from the inference pass. Use the isSecondarySourceFilter property to indicate whether this filter is computing the gradient for the primary or secondary source image from the inference pass.

MPSNNAdditionNodeMPSNNGraphNodes

returns elementwise sum of left + right

MPSNNArithmeticGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSNNArithmeticGradientStateNodeMPSNNGraphNodes

Apple’s documentation

MPSNNBilinearScaleNodeMPSNNGraphNodes

A MPSNNScale object that uses bilinear interpolation for resampling

MPSNNBinaryArithmeticNodeMPSNNGraphNodes

virtual base class for basic arithmetic nodes

MPSNNBinaryGradientStateMPSCore and MPSNNGradientState and MPSState

at the time an -encode call was made. The contents are opaque.

MPSNNBinaryGradientStateNodeMPSNNGraphNodes

Apple’s documentation

MPSNNCompareMPSCNNKernel and MPSCNNMath and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSNNComparisonNodeMPSNNGraphNodes

returns elementwise comparison of left and right

MPSNNComparisonTypeMPSCNNMath

Apple’s documentation

MPSNNConcatenationGradientNodeMPSNNGraphNodes

A MPSNNSlice filter that operates as the conjugate computation for concatentation operators during training

MPSNNConcatenationNodeMPSNNGraphNodes

Node representing a the concatenation (in the feature channel dimension) of the results from one or more kernels

MPSNNConvolutionAccumulatorPrecisionOptionMPSNeuralNetworkTypes

Apple’s documentation

MPSNNCropAndResizeBilinearMPSCNNKernel and MPSCore and MPSKernel and MPSNNResize

Dependencies: This depends on Metal.framework

MPSNNDefaultPaddingMPSNeuralNetworkTypes

This class provides some pre-rolled padding policies for common tasks

MPSNNDivisionNodeMPSNNGraphNodes

returns elementwise quotient of left / right

MPSNNFilterNodeMPSNNGraphNodes

A placeholder node denoting a neural network filter stage

MPSNNForwardLossMPSCNNKernel and MPSCNNLoss and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSNNForwardLossNodeMPSNNGraphNodes

Node representing a MPSNNForwardLosskernel

MPSNNGradientFilterNodeMPSNNGraphNodes

For each MPSNNFilterNode, there is a corresponding MPSNNGradientFilterNode used for training that back propagates image gradients to refine the various parameters in each node. Generally, it takes as input a gradient corresponding to the result image from the MPSNNFilterNode and returns a gradient image corresponding to the source image of the MPSNNFilterNode. In addition, there is generally a MPSNNState produced by the MPSNNFilterNode that is consumed by the MPSNNGradientNode and the MPSNNGradientNode generally needs to look at the MPSNNFilterNode source image.

MPSNNGradientStateMPSCore and MPSNNGradientState and MPSState

at the time an -encode call was made. The contents are opaque.

MPSNNGradientStateNodeMPSNNGraphNodes

During training, each MPSNNFilterNode has a corresponding MPSNNGradientFilterNode for the gradient computation for trainable parameter update. The two communicate through a MPSNNGradientStateNode or subclass which carries information about the inference pass settings to the gradient pass. You can avoid managing these – there will be many! – by using -[MPSNNFilterNode gradientFilterWithSources:] to make the MPSNNGradientFilterNodes. That method will append the necessary extra information like MPSNNGradientState nodes and inference filter source image nodes to the object as needed.

MPSNNGramMatrixCalculationMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSNNGramMatrixCalculationGradientMPSCNNConvolution and MPSCNNKernel and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSNNGramMatrixCalculationGradientNodeMPSNNGraphNodes

Node representing a MPSNNGramMatrixCalculationGradientkernel

MPSNNGramMatrixCalculationNodeMPSNNGraphNodes

Node representing a MPSNNGramMatrixCalculationkernel

MPSNNGraphMPSCore and MPSKernel and MPSNNGraph

Optimized representation of a graph of MPSNNImageNodes and MPSNNFilterNodes

MPSNNGridSampleMPSCNNKernel and MPSCore and MPSKernel and MPSNNGridSample

Apple’s documentation

MPSNNImageNodeMPSNNGraphNodes

A placeholder node denoting the position of a MPSImage in a graph

MPSNNInitialGradientMPSCNNKernel and MPSCNNLoss and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework

MPSNNInitialGradientNodeMPSNNGraphNodes

A node for a MPSNNInitialGradient kernel

MPSNNLabelsNodeMPSNNGraphNodes

The labels and weights for each MPSImage are passed in separately to the graph in a MPSNNLabels object. If the batch interface is used then there will be a MPSStateBatch of these of the same size as the MPSImageBatch that holds the images. The MPSNNLabelsNode is a place holder in the graph for these nodes. The MPSNNLabels node is taken as an input to the Loss node

MPSNNLanczosScaleNodeMPSNNGraphNodes

A MPSNNScale object that uses the Lanczos resampling filter

MPSNNLocalCorrelationMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNLocalCorrelation filter computes the correlation between two images locally with a varying offset on x-y plane between the two source images (controlled by the window and stride properties) and the end result is summed over the feature channels. The results are stored in the different feature channels of the destination image, ordered such that the offset in the x direction is the faster running index.

MPSNNLossGradientMPSCNNKernel and MPSCNNLoss and MPSCore and MPSKernel

Dependencies: This depends on Metal.framework.

MPSNNLossGradientNodeMPSNNGraphNodes

Node representing a MPSNNLossGradientkernel

MPSNNMultiaryGradientStateMPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSNNMultiaryGradientStateNodeMPSNNGraphNodes

Apple’s documentation

MPSNNMultiplicationGradientNodeMPSNNGraphNodes

returns gradient for either primary or secondary source image from the inference pass. Use the isSecondarySourceFilter property to indicate whether this filter is computing the gradient for the primary or secondary source image from the inference pass.

MPSNNMultiplicationNodeMPSNNGraphNodes

returns elementwise product of left * right

MPSNNNeuronDescriptorMPSCNNNeuron

Dependencies: This depends on Metal.framework

MPSNNOptimizerMPSCore and MPSKernel and MPSNNOptimizers

The MPSNNOptimizer base class, use one of the child classes, not to be directly used. Optimizers are generally used to update trainable neural network parameters. Users are usually expected to call these MPSKernels from the update methods on their Convolution or BatchNormalization data sources.

MPSNNOptimizerAdamMPSCore and MPSKernel and MPSNNOptimizers

The MPSNNOptimizerAdam performs an Adam Update

MPSNNOptimizerDescriptorMPSNNOptimizers

The MPSNNOptimizerDescriptor base class. Optimizers are generally used to update trainable neural network parameters. Users are usually expected to call these MPSKernels from the update methods on their Convolution or BatchNormalization data sources.

MPSNNOptimizerRMSPropMPSCore and MPSKernel and MPSNNOptimizers

The MPSNNOptimizerRMSProp performs an RMSProp Update RMSProp is also known as root mean square propagation.

MPSNNOptimizerStochasticGradientDescentMPSCore and MPSKernel and MPSNNOptimizers

The MPSNNOptimizerStochasticGradientDescent performs a gradient descent with an optional momentum Update RMSProp is also known as root mean square propagation.

MPSNNPadMPSCNNKernel and MPSCore and MPSKernel and MPSNNReshape

Apple’s documentation

MPSNNPadGradientMPSCNNKernel and MPSCore and MPSKernel and MPSNNReshape

Dependencies: This depends on Metal.framework

MPSNNPadGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSNNPadNodeMPSNNGraphNodes

A node for a MPSNNPad kernel

MPSNNPaddingMethodMPSNeuralNetworkTypes

Apple’s documentation

MPSNNReduceBinaryMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduce performs a reduction operation The reduction operations supported are:

MPSNNReduceColumnMaxMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceColumnMax performs a reduction operation returning the maximum value for each column of an image

MPSNNReduceColumnMeanMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceColumnMean performs a reduction operation returning the mean value for each column of an image

MPSNNReduceColumnMinMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceColumnMin performs a reduction operation returning the mininmum value for each column of an image

MPSNNReduceColumnSumMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceColumnSum performs a reduction operation returning the sum for each column of an image

MPSNNReduceFeatureChannelsAndWeightsMeanMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

Apple’s documentation

MPSNNReduceFeatureChannelsAndWeightsSumMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

Apple’s documentation

MPSNNReduceFeatureChannelsArgumentMaxMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsArgumentMax performs returns the argument index that is the location of the maximum value for feature channels of an image

MPSNNReduceFeatureChannelsArgumentMinMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsArgumentMin returns the argument index that is the location of the minimum value for feature channels of an image

MPSNNReduceFeatureChannelsMaxMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsMax performs a reduction operation returning the maximum value for feature channels of an image

MPSNNReduceFeatureChannelsMeanMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsMean performs a reduction operation returning the mean value for each column of an image

MPSNNReduceFeatureChannelsMinMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsMin performs a reduction operation returning the mininmum value for feature channels of an image

MPSNNReduceFeatureChannelsSumMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceFeatureChannelsSum performs a reduction operation returning the sum for each column of an image

MPSNNReduceRowMaxMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceRowMax performs a reduction operation returning the maximum value for each row of an image

MPSNNReduceRowMeanMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceRowMean performs a reduction operation returning the mean value for each row of an image

MPSNNReduceRowMinMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceRowMin performs a reduction operation returning the mininmum value for each row of an image

MPSNNReduceRowSumMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduceRowSum performs a reduction operation returning the sum for each row of an image

MPSNNReduceUnaryMPSCNNKernel and MPSCore and MPSKernel and MPSNNReduce

The MPSNNReduce performs a reduction operation The reduction operations supported are:

MPSNNReductionColumnMaxNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionColumnMeanNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionColumnMinNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionColumnSumNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsArgumentMaxNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsArgumentMinNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsMaxNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsMeanNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsMinNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionFeatureChannelsSumNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionRowMaxNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionRowMeanNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionRowMinNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionRowSumNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionSpatialMeanGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReductionSpatialMeanNodeMPSNNGraphNodes

Apple’s documentation

MPSNNRegularizationTypeMPSNNOptimizers

Apple’s documentation

MPSNNReshapeMPSCNNKernel and MPSCore and MPSKernel and MPSNNReshape

Apple’s documentation

MPSNNReshapeGradientMPSCNNKernel and MPSCore and MPSKernel and MPSNNReshape

Dependencies: This depends on Metal.framework

MPSNNReshapeGradientNodeMPSNNGraphNodes

Apple’s documentation

MPSNNReshapeNodeMPSNNGraphNodes

A node for a MPSNNReshape kernel

MPSNNResizeBilinearMPSCNNKernel and MPSCore and MPSKernel and MPSNNResize

Dependencies: This depends on Metal.framework

MPSNNScaleNodeMPSNNGraphNodes

Abstract Node representing a image resampling operation

MPSNNSliceMPSCNNKernel and MPSCore and MPSKernel and MPSNNSlice

Apple’s documentation

MPSNNStateNodeMPSNNGraphNodes

A placeholder node denoting the position in the graph of a MPSState object

MPSNNSubtractionGradientNodeMPSNNGraphNodes

returns gradient for either primary or secondary source image from the inference pass. Use the isSecondarySourceFilter property to indicate whether this filter is computing the gradient for the primary or secondary source image from the inference pass.

MPSNNSubtractionNodeMPSNNGraphNodes

returns elementwise difference of left - right

MPSNNTrainingStyleMPSNeuralNetworkTypes

Apple’s documentation

MPSNNUnaryReductionNodeMPSNNGraphNodes

A node for a unary MPSNNReduce node.

MPSOffsetMPSCoreTypes

A signed coordinate with x, y and z components

MPSOriginMPSCoreTypes

A position in an image

MPSPackedFloat3MPSRayIntersectorTypes

Apple’s documentation

MPSPolygonAccelerationStructureDeprecatedMPSAccelerationStructure and MPSCore and MPSKernel and MPSPolygonAccelerationStructure

An acceleration structure built over polygonal shapes

MPSPolygonBufferDeprecatedMPSPolygonBuffer

A vertex buffer and optional index and mask buffer for a set of polygons

MPSPolygonTypeDeprecatedMPSPolygonAccelerationStructure

Apple’s documentation

MPSPredicateMPSCommandBuffer

Dependencies: This depends on Metal.framework

MPSPurgeableStateMPSImage

Apple’s documentation

MPSQuadrilateralAccelerationStructureDeprecatedMPSAccelerationStructure and MPSCore and MPSKernel and MPSPolygonAccelerationStructure and MPSQuadrilateralAccelerationStructure

An acceleration structure built over quadrilaterals

MPSRNNBidirectionalCombineModeMPSRNNLayer

Apple’s documentation

MPSRNNDescriptorMPSRNNLayer

Dependencies: This depends on Metal.framework

MPSRNNImageInferenceLayerMPSCNNKernel and MPSCore and MPSKernel and MPSRNNLayer

Dependencies: This depends on Metal.framework

MPSRNNMatrixIdMPSRNNLayer

Apple’s documentation

MPSRNNMatrixInferenceLayerMPSCore and MPSKernel and MPSRNNLayer

Dependencies: This depends on Metal.framework

MPSRNNMatrixTrainingLayerMPSCore and MPSKernel and MPSRNNLayer

Dependencies: This depends on Metal.framework

MPSRNNMatrixTrainingStateMPSCore and MPSRNNLayer and MPSState

Dependencies: This depends on Metal.framework

MPSRNNRecurrentImageStateMPSCore and MPSRNNLayer and MPSState

Dependencies: This depends on Metal.framework

MPSRNNRecurrentMatrixStateMPSCore and MPSRNNLayer and MPSState

Dependencies: This depends on Metal.framework

MPSRNNSequenceDirectionMPSRNNLayer

Apple’s documentation

MPSRNNSingleGateDescriptorMPSRNNLayer

Dependencies: This depends on Metal.framework

MPSRayDataTypeMPSRayIntersector

Options for the MPSRayIntersector ray data type property

MPSRayIntersectorDeprecatedMPSCore and MPSKernel and MPSRayIntersector

Performs intersection tests between rays and the geometry in an MPSAccelerationStructure

MPSRayMaskOperatorDeprecatedMPSRayIntersector

Options for the MPSRayIntersector ray mask operator property

MPSRayMaskOptionsDeprecatedMPSRayIntersector

Options for the MPSRayIntersector ray mask options property

MPSRayOriginMaskDirectionMaxDistanceMPSRayIntersectorTypes

Represents a 3D ray with an origin, a direction, and a mask to filter out intersections

MPSRayOriginMinDistanceDirectionMaxDistanceMPSRayIntersectorTypes

Represents a 3D ray with an origin, a direction, and an intersection distance range from the origin

MPSRayPackedOriginDirectionMPSRayIntersectorTypes

Represents a 3D ray with an origin and a direction

MPSRegionMPSCoreTypes

A region of an image

MPSSVGFMPSCore and MPSKernel and MPSSVGF

Reduces noise in images rendered with Monte Carlo ray tracing methods

MPSSVGFDefaultTextureAllocatorMPSSVGF

A default implementation of the MPSSVGFTextureAllocator protocol. Maintains a cache of textures which is checked first when a texture is requested. If there is no suitable texture in the cache, allocates a texture directly from the Metal device.

MPSSVGFDenoiserMPSSVGF

A convenience object which uses an MPSSVGF object to manage the denoising process

MPSScaleTransformMPSCoreTypes

Transform matrix for explict control over resampling in MPSImageScale.

MPSSizeMPSCoreTypes

A size of a region in an image

MPSStateMPSState

Dependencies: This depends on Metal Framework

MPSStateResourceListMPSState

Apple’s documentation

MPSStateResourceTypeMPSState

Apple’s documentation

MPSStateTextureInfoMPSState

Apple’s documentation

MPSTemporalAAMPSCore and MPSKernel and MPSTemporalAA

Reduces aliasing in an image by accumulating samples over multiple frames

MPSTemporalWeightingMPSSVGF

Controls how samples are weighted over time

MPSTemporaryImageMPSImage

Dependencies: MPSImage

MPSTemporaryMatrixMPSMatrix

A MPSMatrix allocated on GPU private memory.

MPSTemporaryNDArrayMPSNDArray

A MPSNDArray that uses command buffer specific memory to store the array data

MPSTemporaryVectorMPSMatrix

A MPSVector allocated on GPU private memory.

MPSTransformTypeMPSInstanceAccelerationStructure

Instance transformation type options

MPSTriangleAccelerationStructureDeprecatedMPSAccelerationStructure and MPSCore and MPSKernel and MPSPolygonAccelerationStructure and MPSTriangleAccelerationStructure

An acceleration structure built over triangles

MPSTriangleIntersectionTestTypeDeprecatedMPSRayIntersector

Options for the MPSRayIntersector triangle intersection test type property

MPSUnaryImageKernelMPSCore and MPSImageKernel and MPSKernel

Dependencies: This depends on Metal.framework

MPSVectorMPSMatrix

Dependencies: This depends on Metal.framework

MPSVectorDescriptorMPSMatrix

Dependencies: This depends on Metal.framework

Constants

MPSBatchSizeIndexMPSFunctionConstantIndices

Apple’s documentation

MPSDeviceCapsIndexMPSFunctionConstantIndices

Apple’s documentation

MPSFunctionConstantIndexMPSFunctionConstantIndices

Apple’s documentation

MPSFunctionConstantIndexReservedMPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantIndexMPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantMultiDestDstAddressingIndexMPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantMultiDestIndexMPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantMultiDestIndex0MPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantMultiDestIndex1MPSFunctionConstantIndices

Apple’s documentation

MPSNDArrayConstantMultiDestSrcAddressingIndexMPSFunctionConstantIndices

Apple’s documentation

MPSTextureLinkingConstantIndexMPSFunctionConstantIndices

Apple’s documentation

MPSUserAvailableFunctionConstantStartIndexMPSFunctionConstantIndices

Apple’s documentation

MPSUserConstantIndexMPSFunctionConstantIndices

Apple’s documentation

Statics

MPSFunctionConstantNoneMPSKernelTypes

Apple’s documentation

MPSRectNoClip^⚠MPSCoreTypes

This is a special constant to indicate no clipping is to be done. The entire image will be used. This is the default clipping rectangle or the input extent for MPSKernels.

Traits

MPSCNNBatchNormalizationDataSourceMPSCNNBatchNormalization

The MPSCNNBatchNormalizationDataSource protocol declares the methods that an instance of MPSCNNBatchNormalizationState uses to initialize the scale factors, bias terms, and batch statistics.

MPSCNNConvolutionDataSourceMPSCNNConvolution

Provides convolution filter weights and bias terms

MPSCNNGroupNormalizationDataSourceMPSCNNGroupNormalization

The MPSCNNGroupNormalizationDataSource protocol declares the methods that an group of MPSCNNGroupNormalization uses to initialize the scale factors (gamma) and bias terms (beta).

MPSCNNInstanceNormalizationDataSourceMPSCNNInstanceNormalization

The MPSCNNInstanceNormalizationDataSource protocol declares the methods that an instance of MPSCNNInstanceNormalization uses to initialize the scale factors (gamma) and bias terms (beta).

MPSDeviceProviderMPSCoreTypes

A way of extending a NSCoder to enable the setting of MTLDevice for unarchived objects

MPSHandleMPSNNGraphNodes

MPS resource identification

MPSHeapProviderMPSCommandBuffer

Apple’s documentation

MPSImageAllocatorMPSImage

A class that allocates new MPSImage or MPSTemporaryImage

MPSImageSizeEncodingStateMPSNeuralNetworkTypes

MPSStates conforming to this protocol contain information about a image size elsewhere in the graph

MPSImageTransformProviderMPSNNGraphNodes

Apple’s documentation

MPSNDArrayAllocatorMPSNDArray

Apple’s documentation

MPSNNGramMatrixCallbackMPSNNGraphNodes

MPSNNGramMatrixCallback Defines a callback protocol for MPSNNGramMatrixCalculationNodeto set the ‘alpha’ scaling value dynamically just before encoding the underlying MPSNNGramMatrixCalculation kernel.

MPSNNLossCallbackMPSNNGraphNodes

MPSNNLossCallback Defines a callback protocol for MPSNNForwardLossNodeand MPSNNLossGradientNodeto set the scalar weight value just before encoding the underlying kernels.

MPSNNPaddingMPSNeuralNetworkTypes

A method to describe how MPSCNNKernels should pad images when data outside the image is needed

MPSNNTrainableNodeMPSNNGraphNodes

Apple’s documentation

MPSSVGFTextureAllocatorMPSSVGF

Protocol dictating how texture allocator objects should operate so that they can be used by an MPSSVGFDenoiser object to allocate and reuse intermediate and output textures during the denoising process.

Functions

MPSGetImageType^⚠MPSImage and MPSKernelTypes

MPSGetPreferredDevice^⚠

Identify the preferred device for MPS computation

MPSHintTemporaryMemoryHighWaterMark^⚠

Hint to MPS how much memory your application expects to need for the command buffer

MPSImageBatchIncrementReadCount^⚠DeprecatedMPSImage

MPSImageBatchIterate^⚠DeprecatedMPSImage and block2

MPSImageBatchResourceSize^⚠DeprecatedMPSImage

MPSImageBatchSynchronize^⚠DeprecatedMPSImage

MPSSetHeapCacheDuration^⚠

Set the timeout after which unused cached MTLHeaps are released

MPSStateBatchIncrementReadCount^⚠DeprecatedMPSState

MPSStateBatchResourceSize^⚠DeprecatedMPSState

MPSStateBatchSynchronize^⚠DeprecatedMPSState

MPSSupportsMTLDevice^⚠

MPSSupportsMTLDevice

Type Aliases

MPSAccelerationStructureCompletionHandlerDeprecatedMPSAccelerationStructure and MPSCore and MPSKernel and block2

A block of code invoked when an operation on an MPSAccelerationStructure is completed

MPSCNNArithmeticGradientStateBatchMPSCNNMath and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNConvolutionGradientStateBatchMPSCNNConvolution and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNConvolutionTransposeGradientStateBatchMPSCNNConvolution and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNDropoutGradientStateBatchMPSCNNDropout and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNGroupNormalizationGradientStateBatchMPSCNNGroupNormalization and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNInstanceNormalizationGradientStateBatchMPSCNNInstanceNormalization and MPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSCNNLossLabelsBatchMPSCNNLoss and MPSCore and MPSState

Apple’s documentation

MPSCopyAllocatorMPSCore and MPSImageKernel and MPSKernel and block2

Apple’s documentation

MPSDeviceCapsMPSKernelTypes

Apple’s documentation

MPSFunctionConstantMPSKernelTypes

Apple’s documentation

MPSFunctionConstantInMetalMPSKernelTypes

Apple’s documentation

MPSGradientNodeBlockMPSNNGraphNodes and block2

Block callback for customizing gradient nodes as they are constructed

MPSImageBatchMPSImage and MPSCore

Apple’s documentation

MPSNNBinaryGradientStateBatchMPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSNNGradientStateBatchMPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSNNGraphCompletionHandlerMPSCore and MPSImage and MPSNNGraph and block2

A notification when computeAsyncWithSourceImages:completionHandler: has finished

MPSNNMultiaryGradientStateBatchMPSCore and MPSNNGradientState and MPSState

Apple’s documentation

MPSShapeMPSCoreTypes

An array of NSNumbers where dimension lengths provided by the user goes from slowest moving to fastest moving dimension. This is same order as MLMultiArray in coreML and most frameworks in Python.

MPSStateBatchMPSState

Apple’s documentation