21.02
|
Interface for the Neon kernel to perform Winograd weights transform. More...
#include <NEWinogradConvolutionLayerKernel.h>
Public Member Functions | |
INEWinogradLayerTransformWeightsKernel (const INEWinogradLayerTransformWeightsKernel &)=default | |
Prevent instances of this class from being copied (As this class contains pointers) More... | |
INEWinogradLayerTransformWeightsKernel & | operator= (const INEWinogradLayerTransformWeightsKernel &)=default |
Prevent instances of this class from being copied (As this class contains pointers) More... | |
INEWinogradLayerTransformWeightsKernel (INEWinogradLayerTransformWeightsKernel &&)=default | |
Allow instances of this class to be moved. More... | |
INEWinogradLayerTransformWeightsKernel & | operator= (INEWinogradLayerTransformWeightsKernel &&)=default |
Allow instances of this class to be moved. More... | |
INEWinogradLayerTransformWeightsKernel () | |
virtual | ~INEWinogradLayerTransformWeightsKernel () |
virtual unsigned int | get_weight_storage_size (int num_output_channels, int num_input_channels) const =0 |
Determine how much memory (in units of T) to allocate for the transformed weights. More... | |
virtual int | get_matrix_stride (int num_output_channels, int num_input_channels) const =0 |
Gets the stride between matrices in the kernel worspace. More... | |
virtual void | configure (const ITensor *weights_hwio, ITensor *output, const int matrix_stride, const int num_output_channels, const int num_input_channels)=0 |
Configure the weights transform kernel. More... | |
Public Member Functions inherited from ICPPKernel | |
virtual | ~ICPPKernel ()=default |
Default destructor. More... | |
virtual void | run (const Window &window, const ThreadInfo &info) |
Execute the kernel on the passed window. More... | |
virtual void | run_nd (const Window &window, const ThreadInfo &info, const Window &thread_locator) |
legacy compatibility layer for implemantions which do not support thread_locator In these cases we simply narrow the interface down the legacy version More... | |
virtual void | run_op (ITensorPack &tensors, const Window &window, const ThreadInfo &info) |
Execute the kernel on the passed window. More... | |
virtual const char * | name () const =0 |
Name of the kernel. More... | |
Public Member Functions inherited from IKernel | |
IKernel () | |
Constructor. More... | |
virtual | ~IKernel ()=default |
Destructor. More... | |
virtual bool | is_parallelisable () const |
Indicates whether or not the kernel is parallelisable. More... | |
virtual BorderSize | border_size () const |
The size of the border for that kernel. More... | |
const Window & | window () const |
The maximum window the kernel can be executed on. More... | |
Static Public Member Functions | |
static Status | validate (const ITensorInfo *input, const ITensorInfo *weights) |
Static function to check if given info will lead to a valid configuration of NEWinogradLayerTransformWeightsKernel. More... | |
Interface for the Neon kernel to perform Winograd weights transform.
Definition at line 441 of file NEWinogradConvolutionLayerKernel.h.
|
default |
Prevent instances of this class from being copied (As this class contains pointers)
Allow instances of this class to be moved.
|
inline |
Definition at line 453 of file NEWinogradConvolutionLayerKernel.h.
|
inlinevirtual |
Definition at line 456 of file NEWinogradConvolutionLayerKernel.h.
References INEWinogradLayerTransformInputKernel::configure(), INEWinogradLayerTransformInputKernel::get_matrix_stride(), arm_compute::test::validation::input, and arm_compute::validate().
|
pure virtual |
Configure the weights transform kernel.
[in] | weights_hwio | Pointer to the weights tensor |
[out] | output | Pointer to working space for the output tensor in the Winograd domain. |
[in] | matrix_stride | Stride across matrices in the output workspace. |
[in] | num_output_channels | Number of filters. |
[in] | num_input_channels | Number of channels in each filter. |
Referenced by NEWinogradLayerTransformWeightsKernel< T, OutputTileRows, OutputTileCols, KernelRows, KernelCols >::get_matrix_stride().
|
pure virtual |
Gets the stride between matrices in the kernel worspace.
[in] | num_output_channels | Number of output feature maps. |
[in] | num_input_channels | Number of input feature maps. |
Implemented in NEWinogradLayerTransformWeightsKernel< T, OutputTileRows, OutputTileCols, KernelRows, KernelCols >.
|
pure virtual |
Determine how much memory (in units of T) to allocate for the transformed weights.
[in] | num_output_channels | Number of output feature maps. |
[in] | num_input_channels | Number of input feature maps. |
Implemented in NEWinogradLayerTransformWeightsKernel< T, OutputTileRows, OutputTileCols, KernelRows, KernelCols >.
|
default |
Prevent instances of this class from being copied (As this class contains pointers)
|
default |
Allow instances of this class to be moved.
|
static |
Static function to check if given info will lead to a valid configuration of NEWinogradLayerTransformWeightsKernel.
[in] | input | First tensor input info. Data types supported: F16/F32. |
[in] | weights | Weights tensor info. Data types supported: same as input . |
Definition at line 180 of file NEWinogradConvolutionLayerKernel.cpp.
References ARM_COMPUTE_RETURN_ERROR_ON, ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES, ARM_COMPUTE_RETURN_ERROR_ON_MSG, arm_compute::test::validation::data_layout, ITensorInfo::data_layout(), ITensorInfo::data_type(), ITensorInfo::dimension(), arm_compute::F16, arm_compute::F32, arm_compute::get_data_layout_dimension_index(), arm_compute::HEIGHT, ITensorInfo::num_dimensions(), and arm_compute::WIDTH.