diff options
| author | Giorgio Arena <giorgio.arena@arm.com> | 2019-07-12 14:49:49 +0100 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-07-26 13:52:08 +0000 |
| commit | 44f5572f3d6ba8e39c4a18a991049992d590ce39 (patch) | |
| tree | c78abd8f4ddd44d2ff28433fa44997be0972bc2d /arm_compute/runtime | |
| parent | c050e0ce189585599b2b70c20aad089e58f657ff (diff) | |
| download | ComputeLibrary-44f5572f3d6ba8e39c4a18a991049992d590ce39.tar.gz | |
COMPMID-2179 New generic depthwise convolution for NEON F32 NHWC
Change-Id: I2b883785c0500d4bdb6ee4700382ee058be2cd36
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1538
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Diffstat (limited to 'arm_compute/runtime')
| -rw-r--r-- | arm_compute/runtime/NEON/functions/NEDepthwiseConvolutionLayer.h | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/arm_compute/runtime/NEON/functions/NEDepthwiseConvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEDepthwiseConvolutionLayer.h index 715f4f5d1d..5b0d1bafcd 100644 --- a/arm_compute/runtime/NEON/functions/NEDepthwiseConvolutionLayer.h +++ b/arm_compute/runtime/NEON/functions/NEDepthwiseConvolutionLayer.h @@ -25,6 +25,7 @@ #define __ARM_COMPUTE_NEDEPTHWISECONVOLUTION_H__ #include "arm_compute/core/NEON/kernels/NEDepthwiseConvolutionLayer3x3Kernel.h" +#include "arm_compute/core/NEON/kernels/NEDepthwiseConvolutionLayerKernel.h" #include "arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h" #include "arm_compute/core/NEON/kernels/NEDepthwiseVectorToTensorKernel.h" #include "arm_compute/core/NEON/kernels/NEDepthwiseWeightsReshapeKernel.h" @@ -280,6 +281,10 @@ private: /** Basic function to execute a generic depthwise convolution. This function calls the following NEON kernels: * + * If data type is F32 and data layout is NHWC: + * -# @ref NEDepthwiseConvolutionLayerKernel + * + * Otherwise: * -# @ref NEDepthwiseIm2ColKernel * -# @ref NEDepthwiseWeightsReshapeKernel * -# @ref NEGEMMMatrixVectorMultiplyKernel @@ -339,8 +344,10 @@ private: NEDepthwiseIm2ColKernel _im2col_kernel; NEDepthwiseWeightsReshapeKernel _weights_reshape_kernel; NEGEMMMatrixVectorMultiplyKernel _v2mm_kernel; + NEDepthwiseConvolutionLayerKernel _depthwise_conv_kernel; NEDepthwiseVectorToTensorKernel _vector_to_tensor_kernel; NEDirectConvolutionLayerOutputStageKernel _output_stage_kernel; + NEFillBorderKernel _fill_border; NEFillBorderKernel _v2mm_input_fill_border; NEFillBorderKernel _v2mm_weights_fill_border; NEPermute _permute_input; @@ -358,6 +365,7 @@ private: bool _is_quantized; bool _is_nhwc; bool _is_activationlayer_enabled; + bool _is_optimized; const ITensor *_original_weights; }; } // namespace arm_compute |