24 #ifndef __ARM_COMPUTE_NEGEMMCONVOLUTIONLAYER_H__ 25 #define __ARM_COMPUTE_NEGEMMCONVOLUTIONLAYER_H__ 182 const ITensor *_original_weights;
184 Tensor _input_im2col_reshaped;
185 Tensor _input_interleaved_reshaped;
194 bool _is_fully_connected_convolution;
195 bool _are_weights_reshaped;
197 bool _is_interleaved;
198 bool _is_activationlayer_enabled;
Base class for all functions.
Function to reshape and perform 1xW transposition on the weights.
GEMM reshape information class.
Basic function to execute NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint on NEON.
Interface for the im2col reshape kernel.
Store the tensor's metadata.
Activation Layer Information class.
Interface for NEON tensor.
NEON kernel to perform col2im reshaping.
Basic function to simulate a convolution layer.
This file contains all available output stages for GEMMLowp on OpenCL.
NEON kernel to perform reshaping on the weights used by convolution and locally connected layer...
Convolution Layer Weights Information class.
NEON kernel to interleave the elements of a matrix.
static Status validate(const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, bool transpose1xW)
Static function to check if given info will lead to a valid configuration of NEConvolutionLayerReshap...
NEON kernel which transposes the elements of a matrix in chunks of 1xW, where W is equal to (16 / ele...
Basic implementation of the tensor interface.
Padding and stride information class.
Interface for the kernel to perform addition between two tensors.
void run() override
Run the kernels contained in the function.
NEConvolutionLayerReshapeWeights(std::shared_ptr< IMemoryManager > memory_manager=nullptr)
Constructor.
Basic function to run NEActivationLayerKernel.
Class for specifying the size of an image or rectangle.
void configure(const ITensor *weights, const ITensor *biases, ITensor *output, bool transpose1xW)
Set the input and output tensors.
DataLayout
Supported tensor data layouts.
Basic function to execute GEMMLowpMatrixMultiplyCore on NEON.
NEON kernel to multiply two input matrices "A" and "B".