diff options
author | Felix Thomasmathibalan <felixjohnny.thomasmathibalan@arm.com> | 2023-09-27 17:46:17 +0100 |
---|---|---|
committer | felixjohnny.thomasmathibalan <felixjohnny.thomasmathibalan@arm.com> | 2023-09-28 12:08:05 +0000 |
commit | afd38f0c617d6f89b2b4532c6c44f116617e2b6f (patch) | |
tree | 03bc7d5a762099989b16a656fa8d397b490ed70e /src/runtime/CL/functions/CLDeconvolutionLayer.cpp | |
parent | bdcb4c148ee2fdeaaddf4cf1e57bbb0de02bb894 (diff) | |
download | ComputeLibrary-afd38f0c617d6f89b2b4532c6c44f116617e2b6f.tar.gz |
Apply clang-format on repository
Code is formatted as per a revised clang format configuration
file(not part of this delivery). Version 14.0.6 is used.
Exclusion List:
- files with .cl extension
- files that are not strictly C/C++ (e.g. Android.bp, Sconscript ...)
And the following directories
- compute_kernel_writer/validation/
- tests/
- include/
- src/core/NEON/kernels/convolution/
- src/core/NEON/kernels/arm_gemm/
- src/core/NEON/kernels/arm_conv/
- data/
There will be a follow up for formatting of .cl files and the
files under tests/ and compute_kernel_writer/validation/.
Signed-off-by: Felix Thomasmathibalan <felixjohnny.thomasmathibalan@arm.com>
Change-Id: Ib7eb1fcf4e7537b9feaefcfc15098a804a3fde0a
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10391
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Diffstat (limited to 'src/runtime/CL/functions/CLDeconvolutionLayer.cpp')
-rw-r--r-- | src/runtime/CL/functions/CLDeconvolutionLayer.cpp | 74 |
1 files changed, 48 insertions, 26 deletions
diff --git a/src/runtime/CL/functions/CLDeconvolutionLayer.cpp b/src/runtime/CL/functions/CLDeconvolutionLayer.cpp index 4421a18f2a..e988ab0ac4 100644 --- a/src/runtime/CL/functions/CLDeconvolutionLayer.cpp +++ b/src/runtime/CL/functions/CLDeconvolutionLayer.cpp @@ -25,16 +25,16 @@ #include "arm_compute/core/Types.h" #include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" #include "arm_compute/core/utils/misc/ShapeCalculator.h" #include "arm_compute/core/utils/quantization/AsymmHelpers.h" +#include "arm_compute/core/Validate.h" #include "arm_compute/runtime/CL/CLScheduler.h" + +#include "src/common/utils/Log.h" #include "src/core/CL/ICLKernel.h" #include "src/gpu/cl/IClOperator.h" #include "src/gpu/cl/operators/ClTransposedConvolution.h" -#include "src/common/utils/Log.h" - #include <cmath> #include <memory> #include <tuple> @@ -44,11 +44,11 @@ using namespace arm_compute::misc::shape_calculator; struct CLDeconvolutionLayer::Impl { - const ICLTensor *src{ nullptr }; - const ICLTensor *weights{ nullptr }; - const ICLTensor *biases{ nullptr }; - ICLTensor *dst{ nullptr }; - std::unique_ptr<opencl::IClOperator> op{ nullptr }; + const ICLTensor *src{nullptr}; + const ICLTensor *weights{nullptr}; + const ICLTensor *biases{nullptr}; + ICLTensor *dst{nullptr}; + std::unique_ptr<opencl::IClOperator> op{nullptr}; }; CLDeconvolutionLayer::~CLDeconvolutionLayer() = default; @@ -58,24 +58,35 @@ CLDeconvolutionLayer::CLDeconvolutionLayer(std::shared_ptr<IMemoryManager> memor { } -void CLDeconvolutionLayer::configure(ICLTensor *input, ICLTensor *weights, const ICLTensor *bias, ICLTensor *output, const PadStrideInfo &deconv_info, - const WeightsInfo &weights_info) +void CLDeconvolutionLayer::configure(ICLTensor *input, + ICLTensor *weights, + const ICLTensor *bias, + ICLTensor *output, + const PadStrideInfo &deconv_info, + const WeightsInfo &weights_info) { configure(CLKernelLibrary::get().get_compile_context(), input, weights, bias, output, deconv_info, weights_info); } -void CLDeconvolutionLayer::configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *weights, const ICLTensor *bias, ICLTensor *output, const PadStrideInfo &deconv_info, - const WeightsInfo &weights_info) +void CLDeconvolutionLayer::configure(const CLCompileContext &compile_context, + ICLTensor *input, + ICLTensor *weights, + const ICLTensor *bias, + ICLTensor *output, + const PadStrideInfo &deconv_info, + const WeightsInfo &weights_info) { ARM_COMPUTE_ERROR_ON_NULLPTR(input, weights, output); ARM_COMPUTE_LOG_PARAMS(input, weights, bias, output, deconv_info, weights_info); - switch(CLDeconvolutionLayer::get_deconvolution_method(input->info(), weights->info(), nullptr, output->info(), deconv_info, weights_info)) + switch (CLDeconvolutionLayer::get_deconvolution_method(input->info(), weights->info(), nullptr, output->info(), + deconv_info, weights_info)) { case DeconvolutionMethod::DIRECT: { auto op = std::make_unique<opencl::ClTransposedConvolution>(); - op->configure(compile_context, input->info(), weights->info(), bias != nullptr ? bias->info() : nullptr, output->info(), deconv_info); + op->configure(compile_context, input->info(), weights->info(), bias != nullptr ? bias->info() : nullptr, + output->info(), deconv_info); _impl->src = input; _impl->weights = weights; @@ -105,22 +116,28 @@ void CLDeconvolutionLayer::configure(const CLCompileContext &compile_context, IC } } -Status CLDeconvolutionLayer::validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *bias, ITensorInfo *output, const PadStrideInfo &deconv_info, - const WeightsInfo &weights_info) +Status CLDeconvolutionLayer::validate(const ITensorInfo *input, + const ITensorInfo *weights, + const ITensorInfo *bias, + ITensorInfo *output, + const PadStrideInfo &deconv_info, + const WeightsInfo &weights_info) { ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, weights, output); - switch(CLDeconvolutionLayer::get_deconvolution_method(input, weights, bias, output, deconv_info, weights_info)) + switch (CLDeconvolutionLayer::get_deconvolution_method(input, weights, bias, output, deconv_info, weights_info)) { case DeconvolutionMethod::DIRECT: { // Validate transposed convolution operator - ARM_COMPUTE_RETURN_ON_ERROR(opencl::ClTransposedConvolution::validate(input, weights, bias, output, deconv_info)); + ARM_COMPUTE_RETURN_ON_ERROR( + opencl::ClTransposedConvolution::validate(input, weights, bias, output, deconv_info)); break; } case DeconvolutionMethod::UPSCALE_CONV2D: { // Validate direct convolution layer - ARM_COMPUTE_RETURN_ON_ERROR(CLDirectDeconvolutionLayer::validate(input, weights, bias, output, deconv_info, weights_info)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLDirectDeconvolutionLayer::validate(input, weights, bias, output, deconv_info, weights_info)); break; } case DeconvolutionMethod::GEMM: @@ -137,12 +154,16 @@ Status CLDeconvolutionLayer::validate(const ITensorInfo *input, const ITensorInf return Status{}; } -DeconvolutionMethod CLDeconvolutionLayer::get_deconvolution_method(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *bias, ITensorInfo *output, const PadStrideInfo &deconv_info, - const WeightsInfo &weights_info) +DeconvolutionMethod CLDeconvolutionLayer::get_deconvolution_method(const ITensorInfo *input, + const ITensorInfo *weights, + const ITensorInfo *bias, + ITensorInfo *output, + const PadStrideInfo &deconv_info, + const WeightsInfo &weights_info) { ARM_COMPUTE_UNUSED(output, bias, weights_info); - if(is_data_type_quantized_per_channel(weights->data_type())) + if (is_data_type_quantized_per_channel(weights->data_type())) { return DeconvolutionMethod::UPSCALE_CONV2D; } @@ -154,11 +175,12 @@ DeconvolutionMethod CLDeconvolutionLayer::get_deconvolution_method(const ITensor const size_t idx_n = get_data_layout_dimension_index(data_layout, DataLayoutDimension::BATCHES); const size_t ofm = weights->tensor_shape()[idx_n]; - if(weights->dimension(idx_w) != deconv_info.stride().first || weights->dimension(idx_h) != deconv_info.stride().second) + if (weights->dimension(idx_w) != deconv_info.stride().first || + weights->dimension(idx_h) != deconv_info.stride().second) { // We observe better performance for FP32 types only when ofm <= 16. // A better heuristic is required for selecting the method for FP16 data types. - if(input->data_layout() == DataLayout::NHWC && !((input->data_type() == DataType::F32) && (ofm > 16))) + if (input->data_layout() == DataLayout::NHWC && !((input->data_type() == DataType::F32) && (ofm > 16))) { return DeconvolutionMethod::DIRECT; } @@ -175,7 +197,7 @@ void CLDeconvolutionLayer::run() { prepare(); - if(_impl->op != nullptr) + if (_impl->op != nullptr) { // Optimized Operator will be used ITensorPack pack; @@ -195,7 +217,7 @@ void CLDeconvolutionLayer::run() void CLDeconvolutionLayer::prepare() { - if(_impl->op == nullptr) + if (_impl->op == nullptr) { _function->prepare(); } |