MPSRNNMatrixInferenceLayer(3)
NAME
MPSRNNMatrixInferenceLayer
SYNOPSIS
#import <MPSRNNLayer.h>
Inherits MPSKernel.
Instance Methods
(nonnull instancetype) - initWithDevice:rnnDescriptor:
(nonnull instancetype) - initWithDevice:rnnDescriptors:
(nonnull instancetype) - initWithDevice:
(void) -
encodeSequenceToCommandBuffer:sourceMatrices:destinationMatrices:recurrentInputState:recurrentOutputStates:
(void) -
encodeBidirectionalSequenceToCommandBuffer:sourceSequence:destinationForwardMatrices:destinationBackwardMatrices:
(nullable instancetype) - initWithCoder:device:
(nonnull instancetype) - copyWithZone:device:
Properties
NSUInteger inputFeatureChannels
NSUInteger outputFeatureChannels
NSUInteger numberOfLayers
BOOL recurrentOutputIsTemporary
BOOL storeAllIntermediateStates
MPSRNNBidirectionalCombineMode bidirectionalCombineMode
Additional Inherited Members
Detailed Description
This depends on Metal.framework The MPSRNNMatrixInferenceLayer
specifies a recurrent neural network layer for inference on
MPSMatrices. Currently two types of recurrent layers are supported:
ones that operate with convolutions on images:
MPSRNNImageInferenceLayer and one that operates on matrices:
MPSRNNMatrixInferenceLayer. The former can be often used to implement
the latter by using 1x1-matrices, but due to image size restrictions
and performance, it is advisable to use MPSRNNMatrixInferenceLayer for
linear recurrent layers. A MPSRNNMatrixInferenceLayer is initialized
using a MPSRNNLayerDescriptor, which further specifies the recurrent
network layer, or an array of MPSRNNLayerDescriptors, which specifies a
stack of recurrent layers, that can operate in parallel a subset of the
inputs in a sequence of inputs and recurrent outputs. Note that
currently stacks with bidirectionally traversing encode functions do
not support starting from a previous set of recurrent states, but this
can be achieved quite easily by defining two separate unidirectional
stacks of layers, and running the same input sequence on them
separately (one forwards and one backwards) and ultimately combining
the two result sequences as desired with auxiliary functions. The input
and output vectors in encode calls are stored as rows of the input and
output matrices and currently MPSRNNMatrixInferenceLayer supports only
matrices with number of rows equal to one. The mathematical operation
then is strictly speaking y^T = W x^T <=> y = x W^T in the linear
transformations of MPSRNNSingleGateDescriptor, MPSLSTMDescriptor and
MPSGRUDescriptor.
Method Documentation
- (nonnull instancetype) copyWithZone: (nullable NSZone *) zone(nullable
id< MTLDevice >) device
Make a copy of this kernel for a new device -
See also:
MPSKernel
Parameters:
zone The NSZone in which to allocate the object
device The device for the new MPSKernel. If nil, then use
self.device.
Returns:
a pointer to a copy of this MPSKernel. This will fail, returning
nil if the device is not supported. Devices must be
MTLFeatureSet_iOS_GPUFamily2_v1 or later.
Reimplemented from MPSKernel.
- (void) encodeBidirectionalSequenceToCommandBuffer: (nonnull id<
MTLCommandBuffer >) commandBuffer(NSArray< MPSMatrix * > *__nonnull)
sourceSequence(NSArray< MPSMatrix * > *__nonnull)
destinationForwardMatrices(NSArray< MPSMatrix * > *__nullable)
destinationBackwardMatrices
Encode an MPSRNNMatrixInferenceLayer kernel stack for an input matrix
sequences into a command buffer bidirectionally. The operation proceeds
as follows: The first source matrix x0 is passed through all forward
traversing layers in the stack, ie. those that were initialized with
MPSRNNSequenceDirectionForward, recurrent input is assumed zero. This
produces forward output yf0 and recurrent states hf00, hf01, hf02, ...
hf0n, one for each forward layer in the stack. Then x1 is passed to
forward layers together with recurrent state hf00, hf01, ..., hf0n,
which produces yf1, and hf10,... This procedure is iterated until the
last matrix in the input sequence x_(N-1), which produces forward
output yf(N-1). The backwards layers iterate the same sequence
backwards, starting from input x_(N-1) (recurrent state zero), that
produces yb(N-1) and recurrent output hb(N-1)0, hf(N-1)1, ... hb(N-1)m,
one for each backwards traversing layer. Then the backwards layers
handle input x_(N-2) using recurrent state hb(N-1)0, ..., et cetera,
until the first matrix of the sequence is computed, producing output
yb0. The result of the operation is either pair of sequences ({yf0,
yf1, ... , yf(N-1)}, {yb0, yb1, ... , yb(N-1)}) or a combined sequence,
{(yf0 + yb0), ... , (yf(N-1) + yb(N-1)) }, where '+' stands either for
sum, or concatenation along feature channels, as specified by
bidirectionalCombineMode.
Parameters:
commandBuffer A valid MTLCommandBuffer to receive the encoded
filter
sourceSequence An array of valid MPSMatrix objects containing the
source matrix sequence (x0, x1, ... x_n-1).
destinationForwardMatrices An array of valid MPSMatrices to be
overwritten by result from forward input matrices. If
bidirectionalCombineMode is either
MPSRNNBidirectionalCombineModeAdd or
MPSRNNBidirectionalCombineModeConcatenate, then will contain the
combined results. destinationForwardMatrix may not alias with any
of the source matrices.
destinationBackwardMatrices If bidirectionalCombineMode is
MPSRNNBidirectionalCombineModeNone, then must be an array of valid
MPSMatrices that will be overwritten by result from backward input
matrices. Otherwise this parameter is ignored and can be nil.
destinationBackwardMatrices may not alias to any of the source
matrices.
- (void) encodeSequenceToCommandBuffer: (nonnull id< MTLCommandBuffer >)
commandBuffer(NSArray< MPSMatrix * > *__nonnull)
sourceMatrices(NSArray< MPSMatrix * > *__nonnull)
destinationMatrices(MPSRNNRecurrentMatrixState *__nullable)
recurrentInputState(NSMutableArray< MPSRNNRecurrentMatrixState * >
*__nullable) recurrentOutputStates
Encode an MPSRNNMatrixInferenceLayer kernel (stack) for a sequence of
inputs into a command buffer. Note that when encoding using this
function the
See also:
layerSequenceDirection is ignored and the layer stack operates as
if all layers were forward feeding layers. In order to run
bidirectional sequences use
encodeBidirectionalSequenceToCommandBuffer:sourceSequence: or
alternatively run two layer stacks and combine results at the end
using utility functions.
Parameters:
commandBuffer A valid MTLCommandBuffer to receive the encoded
filter
sourceMatrices An array of valid MPSMatrix objects containing the
sequence of source matrices.
destinationMatrices An array valid MPSMatrices to be overwritten by
result matrix sequence. destinationMatrices may not alias
sourceMatrices.
recurrentInputState An optional state containing the output
matrices and memory cells (for LSTMs) of the layer obtained from
the previous input matrices in a sequence of inputs. Has to be the
output of a previous call to this function or nil (assumed zero).
Note: can be one of the states returned in
intermediateRecurrentStates.
recurrentOutputStates An optional array that will contain the
recurrent output states. If nil then the recurrent output state is
discarded. If storeAllIntermediateStates is YES, then all
intermediate states of the sequence are returned in the array, the
first one corresponding to the first input in the sequence,
otherwise only the last recurrent output state is returned. If
recurrentOutputIsTemporary is YES and then all returned recurrent
states will be temporary.
See also:
MPSState:isTemporary. Example: In order to get a new state one can
do the following:
MPSRNNRecurrentMatrixState* recurrent0 = nil;
[filter encodeToCommandBuffer: cmdBuf
sourceMatrix: source0
destinationMatrix: destination0
recurrentInputState: nil
recurrentOutputState: &recurrent0];
Then use it for the next input in sequence:
[filter encodeToCommandBuffer: cmdBuf
sourceMatrix: source1
destinationMatrix: destination1
recurrentInputState: recurrent0
recurrentOutputState: &recurrent0];
And discard recurrent output of the third input:
[filter encodeToCommandBuffer: cmdBuf
sourceMatrix: source2
destinationMatrix: destination2
recurrentInputState: recurrent0
recurrentOutputState: nil];
- (nullable instancetype) initWithCoder: (NSCoder *__nonnull)
aDecoder(nonnull id< MTLDevice >) device
NSSecureCoding compatability See MPSKernel::initWithCoder.
Parameters:
aDecoder The NSCoder subclass with your serialized
MPSRNNMatrixInferenceLayer
device The MTLDevice on which to make the
MPSRNNMatrixInferenceLayer
Returns:
A new MPSRNNMatrixInferenceLayer object, or nil if failure.
Reimplemented from MPSKernel.
- (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 MPSKernel.
- (nonnull instancetype) initWithDevice: (nonnull id< MTLDevice >)
device(nonnull const MPSRNNDescriptor *) rnnDescriptor
Initializes a linear (fully connected) RNN kernel
Parameters:
device The MTLDevice on which this MPSRNNMatrixLayer filter will be
used
rnnDescriptor The descriptor that defines the RNN layer
Returns:
A valid MPSRNNMatrixInferenceLayer object or nil, if failure.
- (nonnull instancetype) initWithDevice: (nonnull id< MTLDevice >)
device(NSArray< const MPSRNNDescriptor * > *__nonnull) rnnDescriptors
Initializes a kernel that implements a stack of linear (fully
connected) RNN layers
Parameters:
device The MTLDevice on which this MPSRNNMatrixLayer filter will be
used
rnnDescriptors An array of RNN descriptors that defines a stack of
RNN layers, starting at index zero. The number of layers in stack
is the number of entries in the array. All entries in the array
must be valid MPSRNNDescriptors.
Returns:
A valid MPSRNNMatrixInferenceLayer object or nil, if failure.
Property Documentation
- bidirectionalCombineMode [read], [write], [nonatomic], [assign]
Defines how to combine the output-results, when encoding bidirectional
layers using encodeBidirectionalSequenceToCommandBuffer. Defaults to
MPSRNNBidirectionalCombineModeNone.
- inputFeatureChannels [read], [nonatomic], [assign]
The number of feature channels input vector/matrix.
- numberOfLayers [read], [nonatomic], [assign]
Number of layers in the filter-stack. This will be one when using
initWithDevice:rnnDescriptor to initialize this filter and the number
of entries in the array 'rnnDescriptors' when initializing this filter
with initWithDevice:rnnDescriptors.
- outputFeatureChannels [read], [nonatomic], [assign]
The number of feature channels in the output vector/matrix.
- recurrentOutputIsTemporary [read], [write], [nonatomic], [assign]
How output states from encodeSequenceToCommandBuffer are constructed.
Defaults to NO. For reference
See also:
MPSState.
- storeAllIntermediateStates [read], [write], [nonatomic], [assign]
If YES then calls to encodeSequenceToCommandBuffer return every
recurrent state in the array: recurrentOutputStates. Defaults to NO.
Author
Generated automatically by Doxygen for
MetalPerformanceShaders.framework from the source code.
Version MetalPerformanceShaders-Thu2Jul 13 2017 MPSRNNMatrixInferenceLayer(3)
Mac OS X 10.12.6 - Generated Tue Oct 31 19:54:17 CDT 2017
