21.02
|
Basic function to manage the reshape weights generated from CLConvolutionLayerReshapeWeights. More...
#include <CLGEMMConvolutionLayer.h>
Public Member Functions | |
void | configure (const ICLTensor *input, const ICLTensor *biases, unsigned int num_groups) |
Configures the CLConvolutionLayerReshapeWeights function. More... | |
void | configure (const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *biases, unsigned int num_groups) |
Configures the CLConvolutionLayerReshapeWeights function. More... | |
void | run () override |
Run the transformation function. More... | |
ICLTensor * | get_weights () override |
Get a pointer to the transformed weights. More... | |
void | release () override |
Release transformed weights memory. More... | |
uint32_t | uid () override |
Function that returns a unique id of the reshape function. More... | |
Public Member Functions inherited from ITransformWeights | |
ITransformWeights ()=default | |
Default Constructor. More... | |
virtual | ~ITransformWeights ()=default |
Default Destructor. More... | |
ITransformWeights (const ITransformWeights &)=delete | |
Prevent instances of this class to be copy constructed. More... | |
ITransformWeights & | operator= (const ITransformWeights &)=delete |
Prevent instances of this class to be copied. More... | |
ITransformWeights (ITransformWeights &&other) | |
Allow instances of this class to be move constructed. More... | |
ITransformWeights & | operator= (ITransformWeights &&other) |
Allow instances of this class to be moved. More... | |
void | increase_refcount () |
Increase the object's refcount. More... | |
int32_t | decrease_refcount () |
Decrease the object's refcount and return the updated value. More... | |
bool | is_reshape_run () |
Function that returns a flag on whether the weights are reshaped or not. More... | |
Basic function to manage the reshape weights generated from CLConvolutionLayerReshapeWeights.
Definition at line 107 of file CLGEMMConvolutionLayer.h.
|
inline |
Configures the CLConvolutionLayerReshapeWeights function.
[in] | input | Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/F16/F32. |
[in] | biases | Biases tensor. Data type supported: same as input , S32 if input is quantized. |
[in] | num_groups | Number of groups when performing a grouped convolution. |
Definition at line 116 of file CLGEMMConvolutionLayer.h.
References CLConvolutionLayerReshapeWeights::configure(), and CLKernelLibrary::get().
Referenced by CLGEMMConvolutionLayer::configure().
|
inline |
Configures the CLConvolutionLayerReshapeWeights function.
[in] | compile_context | The compile context to be used. |
[in] | input | Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/F16/F32. |
[in] | biases | Biases tensor. Data type supported: same as input , S32 if input is quantized. |
[in] | num_groups | Number of groups when performing a grouped convolution. |
Definition at line 127 of file CLGEMMConvolutionLayer.h.
References arm_compute::test::validation::num_groups.
|
inlineoverridevirtual |
Get a pointer to the transformed weights.
Implements ITransformWeights.
Definition at line 143 of file CLGEMMConvolutionLayer.h.
|
inlineoverridevirtual |
Release transformed weights memory.
Implements ITransformWeights.
Definition at line 149 of file CLGEMMConvolutionLayer.h.
|
inlineoverridevirtual |
Run the transformation function.
Implements ITransformWeights.
Definition at line 135 of file CLGEMMConvolutionLayer.h.
|
inlineoverridevirtual |
Function that returns a unique id of the reshape function.
Implements ITransformWeights.
Definition at line 155 of file CLGEMMConvolutionLayer.h.