Python* API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1
Abstract class which defines interface for the layer. More...
Public Member Functions | |
def | getLayerResult |
def | getLayerInput |
def | getLayerParameter |
def | clone |
def | allocateResult |
def | allocateInput |
def | initializeInput |
def | addInput |
def | getLayerForPrediction |
![]() | |
def | computeNoThrow |
def | compute |
def | checkComputeParams |
def | checkResult |
def | setupCompute |
def | resetCompute |
def | enableResetOnCompute |
def | hostApp |
def | setHostApp |
![]() | |
def | checkComputeParams |
def | getBaseParameter |
![]() | |
def | enableChecks |
def | isChecksEnabled |
![]() | |
def | checkComputeParams |
def | checkResult |
def | getMethod |
def addInput | ( | self, | |
result, | |||
resultIndex, | |||
inputIndex | |||
) |
Connects two layers in neural network by getting tensor with value from the result of the previous layer and adding it to the input object of this layer algorithm
result | Structure that contains results of the previous layer |
resultIndex | Index of the tensor with value in the structure that contains results of the previous layer |
inputIndex | Index in the input object of this layer algorithm where the tensor with value should be placed |
def allocateInput | ( | self | ) |
Allocates memory buffers needed for the computations
def allocateResult | ( | self | ) |
Allocates memory buffers needed for the computations
def clone | ( | self | ) |
Returns a pointer to the newly allocated forward neural network layer with a copy of input objects and parameters of this layer
def getLayerForPrediction | ( | self | ) |
Returns the layer that corresponds to this layer on the prediction stage
def getLayerInput | ( | self | ) |
Returns the structure that contains input objects of the layer
def getLayerParameter | ( | self | ) |
Returns the structure that contains parameters of the layer
def getLayerResult | ( | self | ) |
Returns the structure that contains results of the layer
def initializeInput | ( | self | ) |
Initializes values of weights and biases if needed
For more complete information about compiler optimizations, see our Optimization Notice.