MPSCNNDropout(3) MetalPerformanceShaders.framework MPSCNNDropout(3)
NAME
MPSCNNDropout
SYNOPSIS
#import <MPSCNNDropout.h>
Inherits MPSCNNKernel.
Instance Methods
(nonnull instancetype) - initWithDevice:
(nullable instancetype) - initWithCoder:device:
(nonnull instancetype) - initWithDevice:keepProbability:seed:maskStrideInPixels:
(MPSCNNDropoutGradientState *__nullable) - temporaryResultStateForCommandBuffer:sourceImage:sourceStates:destinationImage:
(MPSCNNDropoutGradientStateBatch *__nullable) - temporaryResultStateBatchForCommandBuffer:
Properties
float keepProbability
NSUInteger seed
MTLSize maskStrideInPixels
Additional Inherited Members
Detailed Description
This depends on Metal.framework Dropout is a regularization technique used to prevent neural networks from overfitting during training.
With probability keepProbability, this filter outputs the input element scaled by 1 / keepProbability. Otherwise, it outputs 0. Each input
element is kept or dropped independently. The scaling is performed to keep the energy of the output unchanged.
Method Documentation
- (nullable instancetype) initWithCoder: (NSCoder *__nonnull) aDecoder(nonnull id< MTLDevice >) device
<NSSecureCoding> support
Reimplemented from MPSCNNKernel.
- (nonnull instancetype) initWithDevice: (nonnull id< MTLDevice >) device
Standard init with default properties per filter type
Parameters:
device The device that the filter will be used on. May not be NULL.
Returns:
A pointer to the newly initialized object. This will fail, returning nil if the device is not supported. Devices must be
MTLFeatureSet_iOS_GPUFamily2_v1 or later.
Reimplemented from MPSCNNKernel.
- (nonnull instancetype) initWithDevice: (nonnull id< MTLDevice >) device(float) keepProbability(NSUInteger) seed(MTLSize) maskStrideInPixels
Standard init with default properties per filter type.
Parameters:
device The device that the filter will be used on.
keepProbability The probability that each element in the input is kept. The valid range is (0.0f, 1.0f).
seed The seed used to generate random numbers.
maskStrideInPixels The mask stride in the x, y, and z dimensions, which allows for the broadcasting of mask data. The only valid values
are 0 and 1 for each dimension. For no broadcasting, set the values for each dimension to 1. For broadcasting, set desired values to 0.
Returns:
A valid MPSCNNDropout object or nil, if failure.
- (MPSCNNDropoutGradientStateBatch * __nullable) temporaryResultStateBatchForCommandBuffer: (__nonnull id< MTLCommandBuffer >) commandBuffer
- (MPSCNNDropoutGradientState * __nullable) temporaryResultStateForCommandBuffer: (__nonnull id< MTLCommandBuffer >) commandBuffer(MPSImage
*__nonnull) sourceImage(NSArray< MPSState * > *__nullable) sourceStates(MPSImage *__nonnull) dest
Property Documentation
- keepProbability [read], [nonatomic], [assign]
The probability that each element in the input is kept. The valid range is (0.0f, 1.0f).
- maskStrideInPixels [read], [nonatomic], [assign]
The mask stride in the x, y, and x dimensions, which allows for the broadcasting the mask data. The only valid values are 0 and 1 for each
dimension. For no broadcasting, set the values for each dimension to 1. For broadcasting, set desired values to 0.
- seed [read], [nonatomic], [assign]
The seed used to generate random numbers.
Author
Generated automatically by Doxygen for MetalPerformanceShaders.framework from the source code.
Version MetalPerformanceShaders-100 Thu Feb 8 2018 MPSCNNDropout(3)