diff options
author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2018-06-28 16:29:29 +0100 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:54:54 +0000 |
commit | 215b4ea6c9dee480a22070d5873b0b8cb52531a0 (patch) | |
tree | 398e552c4d01c0b84d03a873098a9183ba8f82e4 /src/runtime/CL/functions/CLFlattenLayer.cpp | |
parent | ad486e21e5870f41774f30825c270762e08ae71e (diff) | |
download | ComputeLibrary-215b4ea6c9dee480a22070d5873b0b8cb52531a0.tar.gz |
COMPMID-1277 - Optimizing CLIm2ColKernel for NHWC.
This patch includes:
- Im2Col optimizations for NHWC using a new data layout
- Refactoring of CLIm2ColKernel adding validation method and auto-init
- Removed im2col_reduced from CLIm2ColKernel and created a new kernel CLFlattenLayerKernel
Change-Id: I1620640b6796baa268324b33ae92cdd8de53e27c
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/141241
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Diffstat (limited to 'src/runtime/CL/functions/CLFlattenLayer.cpp')
-rw-r--r-- | src/runtime/CL/functions/CLFlattenLayer.cpp | 12 |
1 files changed, 8 insertions, 4 deletions
diff --git a/src/runtime/CL/functions/CLFlattenLayer.cpp b/src/runtime/CL/functions/CLFlattenLayer.cpp index f5809a218a..b372c35dd9 100644 --- a/src/runtime/CL/functions/CLFlattenLayer.cpp +++ b/src/runtime/CL/functions/CLFlattenLayer.cpp @@ -23,8 +23,7 @@ */ #include "arm_compute/runtime/CL/functions/CLFlattenLayer.h" -#include "arm_compute/core/CL/kernels/CLIm2ColKernel.h" -#include "arm_compute/core/Size2D.h" +#include "arm_compute/core/CL/kernels/CLFlattenLayerKernel.h" #include "arm_compute/runtime/CL/CLScheduler.h" #include "support/ToolchainSupport.h" @@ -32,8 +31,13 @@ using namespace arm_compute; void CLFlattenLayer::configure(const ICLTensor *input, ICLTensor *output) { - auto k = arm_compute::support::cpp14::make_unique<CLIm2ColKernel>(); - k->configure(input, output, Size2D(1, 1), PadStrideInfo(1, 1, 0, 0), false); + auto k = arm_compute::support::cpp14::make_unique<CLFlattenLayerKernel>(); + k->configure(input, output); _kernel = std::move(k); CLScheduler::get().tune_kernel_static(*_kernel); } + +Status CLFlattenLayer::validate(const ITensorInfo *input, const ITensorInfo *output) +{ + return CLFlattenLayerKernel::validate(input, output); +}
\ No newline at end of file |