diff options
| author | Giorgio Arena <giorgio.arena@arm.com> | 2018-04-04 17:44:26 +0100 |
|---|---|---|
| committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:50:48 +0000 |
| commit | 7657224de2b697a8a92cccf26d98e53ccd7c1a03 (patch) | |
| tree | 1dcfa4541dbaf753854a628c93991652158d373e /arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h | |
| parent | e74b201ca1abca040ca9f30837fdf19aa610e7c4 (diff) | |
| download | ComputeLibrary-7657224de2b697a8a92cccf26d98e53ccd7c1a03.tar.gz | |
COMPMID-926 Add depth multiplier support to NEON/CL/GLES depthwise convolution
Change-Id: I03f32c62350e5ea43e77bb15fc5a832d83719e3b
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/126657
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h')
| -rw-r--r-- | arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h | 18 |
1 files changed, 10 insertions, 8 deletions
diff --git a/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h b/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h index ca10bfaab2..9c11cfa425 100644 --- a/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h +++ b/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h @@ -54,15 +54,16 @@ public: NEDepthwiseIm2ColKernel &operator=(NEDepthwiseIm2ColKernel &&) = default; /** Set the input and output of the kernel. * - * @param[in] input The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM], - * while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8, F32 - * @param[out] output The output tensor. First 3 lower dimensions represent a transform of each 3D input, - * while every dimension above 3 represents a batch. Data types supported: Same as @p input - * @param[in] kernel_dims The kernel dimensions (width and height). - * @param[in] conv_info Contains padding and stride information described in @ref PadStrideInfo. - * @param[in] has_bias Boolean that specifies if the depthwise convolution has bias. + * @param[in] input The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM], + * while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8, F32 + * @param[out] output The output tensor. First 3 lower dimensions represent a transform of each 3D input, + * while every dimension above 3 represents a batch. Data types supported: Same as @p input + * @param[in] kernel_dims The kernel dimensions (width and height). + * @param[in] conv_info Contains padding and stride information described in @ref PadStrideInfo. + * @param[in] has_bias Boolean that specifies if the depthwise convolution has bias. + * @param[in] depth_multiplier (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1. */ - void configure(const ITensor *input, ITensor *output, const Size2D &kernel_dims, const PadStrideInfo &conv_info, bool has_bias = false); + void configure(const ITensor *input, ITensor *output, const Size2D &kernel_dims, const PadStrideInfo &conv_info, bool has_bias = false, unsigned int depth_multiplier = 1); // Inherited methods overridden: void run(const Window &window, const ThreadInfo &info) override; @@ -87,6 +88,7 @@ private: Size2D _kernel_dims; PadStrideInfo _conv_info; bool _has_bias; + unsigned int _depth_multiplier; }; } // arm_compute #endif /*__ARM_COMPUTE_NEDEPTHWISEIM2COLKERNEL_H__ */ |