diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2021-04-16 15:08:59 +0100 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2021-06-24 11:16:30 +0000 |
| commit | 561c176598cd14245e2e7918fdf136d1c888d1da (patch) | |
| tree | 82adfff6de30292dabbbcc7ced4ae35cac3d45cf /src/core/CL/CLKernels.h | |
| parent | 31c7c26822270f1c4952c8973aa8bfb38e0a7c68 (diff) | |
| download | ComputeLibrary-561c176598cd14245e2e7918fdf136d1c888d1da.tar.gz | |
Rework OpenCL Depthwise Convolution
- Remove dedicated kernels for NCHW. Now we only use NHWC with permute
- Remove specialized kernels for 3x3 NHWC
- Simplify CLDepthwiseConvolutionLayer.cpp to call just the native
implementation for both floating-point and quantized data types
- Develop two parametric opencl kernels for depthwise convolution layer NHWC
(floating-point and quantized)
- Add support to export the weights to cl_image
- Extend test for depthwise convolution on opencl
Resolves COMPMID-4417
Change-Id: I253dd5d959a70783c82e62b1771a5e9f91621cb0
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5806
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Diffstat (limited to 'src/core/CL/CLKernels.h')
| -rw-r--r-- | src/core/CL/CLKernels.h | 2 |
1 files changed, 0 insertions, 2 deletions
diff --git a/src/core/CL/CLKernels.h b/src/core/CL/CLKernels.h index 5dc95dae27..c380a9cc68 100644 --- a/src/core/CL/CLKernels.h +++ b/src/core/CL/CLKernels.h @@ -36,8 +36,6 @@ #include "src/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h" #include "src/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h" #include "src/core/CL/kernels/CLDepthToSpaceLayerKernel.h" -#include "src/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NCHWKernel.h" -#include "src/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NHWCKernel.h" #include "src/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h" #include "src/core/CL/kernels/CLFFTDigitReverseKernel.h" #include "src/core/CL/kernels/CLFFTRadixStageKernel.h" |