C++ API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2
Provides methods to run implementations of the of the backward 2D convolution layer This class is associated with the daal::algorithms::neural_networks::layers::convolution2d::backward::Batch class and supports the method of backward 2D convolution layer computation in the batch processing mode. More...
algorithmFPType | Data type to use in intermediate computations of backward 2D convolution layer, double or float |
method | Computation method of the layer, daal::algorithms::neural_networks::layers::convolution2d::Method |
cpu | Version of the cpu-specific implementation of the layer, daal::CpuType |
BatchContainer | ( | daal::services::Environment::env * | daalEnv | ) |
Constructs a container for the backward 2D convolution layer with a specified environment in the batch processing mode
[in] | daalEnv | Environment object |
~BatchContainer | ( | ) |
Default destructor
|
virtual |
Computes the result of the backward 2D convolution layer in the batch processing mode
Implements AlgorithmContainer< mode >.
|
virtual |
Resets internal datastructures for compute function.
Reimplemented from AlgorithmContainerImpl< mode >.
|
virtual |
Setups internal datastructures for compute function.
Reimplemented from AlgorithmContainerImpl< mode >.
For more complete information about compiler optimizations, see our Optimization Notice.