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>
diff --git a/src/runtime/NEON/functions/NEDeconvolutionLayer.cpp b/src/runtime/NEON/functions/NEDeconvolutionLayer.cpp
index 439aff0..3987370 100644
--- a/src/runtime/NEON/functions/NEDeconvolutionLayer.cpp
+++ b/src/runtime/NEON/functions/NEDeconvolutionLayer.cpp
@@ -25,9 +25,10 @@
#include "arm_compute/core/Helpers.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/Validate.h"
#include "arm_compute/runtime/NEON/NEScheduler.h"
+
#include "src/common/utils/Log.h"
#include "src/core/helpers/AutoConfiguration.h"
@@ -61,7 +62,8 @@
deconv_pad_top += deconv_pad_y / 2;
deconv_pad_bottom += deconv_pad_y / 2;
- return PadStrideInfo(stride_x, stride_y, deconv_pad_left, deconv_pad_right, deconv_pad_top, deconv_pad_bottom, DimensionRoundingType::FLOOR);
+ return PadStrideInfo(stride_x, stride_y, deconv_pad_left, deconv_pad_right, deconv_pad_top, deconv_pad_bottom,
+ DimensionRoundingType::FLOOR);
}
} // namespace
@@ -82,17 +84,24 @@
{
}
-Status NEDeconvolutionLayer::validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *bias, const ITensorInfo *output, const PadStrideInfo &info,
- bool enable_fast_math, const WeightsInfo &weights_info)
+Status NEDeconvolutionLayer::validate(const ITensorInfo *input,
+ const ITensorInfo *weights,
+ const ITensorInfo *bias,
+ const ITensorInfo *output,
+ const PadStrideInfo &info,
+ bool enable_fast_math,
+ const WeightsInfo &weights_info)
{
ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, weights, output);
- ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F32, DataType::F16, DataType::QASYMM8, DataType::QASYMM8_SIGNED);
- const unsigned int width_idx = get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::WIDTH);
- const unsigned int height_idx = get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::HEIGHT);
+ ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F32, DataType::F16, DataType::QASYMM8,
+ DataType::QASYMM8_SIGNED);
+ const unsigned int width_idx = get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::WIDTH);
+ const unsigned int height_idx =
+ get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::HEIGHT);
ARM_COMPUTE_RETURN_ERROR_ON(weights->dimension(width_idx) < 1);
ARM_COMPUTE_RETURN_ERROR_ON(weights->dimension(height_idx) < 1);
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(weights, input);
- if(is_data_type_quantized_per_channel(weights->data_type()) && is_data_type_quantized(input->data_type()))
+ if (is_data_type_quantized_per_channel(weights->data_type()) && is_data_type_quantized(input->data_type()))
{
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(weights, 1, DataType::QSYMM8_PER_CHANNEL);
}
@@ -101,11 +110,13 @@
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, weights);
}
- auto out_dims = deconvolution_output_dimensions(input->dimension(width_idx), input->dimension(height_idx), weights->dimension(width_idx), weights->dimension(height_idx), info);
+ auto out_dims =
+ deconvolution_output_dimensions(input->dimension(width_idx), input->dimension(height_idx),
+ weights->dimension(width_idx), weights->dimension(height_idx), info);
- if(bias != nullptr)
+ if (bias != nullptr)
{
- if(is_data_type_quantized_asymmetric(input->data_type()))
+ if (is_data_type_quantized_asymmetric(input->data_type()))
{
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(bias, 1, DataType::S32);
}
@@ -115,15 +126,18 @@
}
}
- if(output->tensor_shape().total_size() > 0)
+ if (output->tensor_shape().total_size() > 0)
{
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
const TensorShape output_shape = compute_deconvolution_output_shape(out_dims, *input, *weights);
- ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimX) != output_shape.x(), "Output's width is invalid.");
- ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimY) != output_shape.y(), "Output's height is invalid.");
- ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimZ) != output_shape.z(), "Output's depth is invalid.");
+ ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimX) != output_shape.x(),
+ "Output's width is invalid.");
+ ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimY) != output_shape.y(),
+ "Output's height is invalid.");
+ ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->dimension(Window::DimZ) != output_shape.z(),
+ "Output's depth is invalid.");
}
uint32_t deconv_pad_x = 0;
@@ -141,44 +155,61 @@
ARM_COMPUTE_RETURN_ERROR_ON((out_x - weights->dimension(idx_w) + 1) > out_dims.first);
ARM_COMPUTE_RETURN_ERROR_ON((out_y - weights->dimension(idx_h) + 1) > out_dims.second);
- const TensorShape scale_out_shape = compute_deconvolution_upsampled_shape(*input, *weights, stride_x, stride_y, out_dims, deconv_pad_x, deconv_pad_y);
- TensorInfo scale_out_info(input->clone()->set_is_resizable(true).reset_padding().set_tensor_shape(scale_out_shape));
+ const TensorShape scale_out_shape = compute_deconvolution_upsampled_shape(*input, *weights, stride_x, stride_y,
+ out_dims, deconv_pad_x, deconv_pad_y);
+ TensorInfo scale_out_info(input->clone()->set_is_resizable(true).reset_padding().set_tensor_shape(scale_out_shape));
const PadStrideInfo upsample_info = compute_upsample_info(info, deconv_pad_x, deconv_pad_y);
// Do not perform upsampling when the operation uses unit stride in all dimensions
const bool do_upsampling = stride_x != 1 || stride_y != 1;
- const unsigned int batches_idx = get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::BATCHES);
- const unsigned int channel_idx = get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::CHANNEL);
+ const unsigned int batches_idx =
+ get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::BATCHES);
+ const unsigned int channel_idx =
+ get_data_layout_dimension_index(weights->data_layout(), DataLayoutDimension::CHANNEL);
ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(batches_idx) != scale_out_info.dimension(batches_idx));
ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(channel_idx) != scale_out_info.dimension(channel_idx));
- if(do_upsampling)
+ if (do_upsampling)
{
const PadStrideInfo conv_info(1, 1, 0, 0, 0, 0, DimensionRoundingType::CEIL);
- ARM_COMPUTE_RETURN_ON_ERROR(NEConvolutionLayer::validate(&scale_out_info, weights, bias, output, conv_info, weights_info, Size2D(1U, 1U), ActivationLayerInfo(), enable_fast_math));
+ ARM_COMPUTE_RETURN_ON_ERROR(NEConvolutionLayer::validate(&scale_out_info, weights, bias, output, conv_info,
+ weights_info, Size2D(1U, 1U), ActivationLayerInfo(),
+ enable_fast_math));
}
else
{
- const PadStrideInfo conv_info(1, 1, upsample_info.pad_left(), upsample_info.pad_right(), upsample_info.pad_top(), upsample_info.pad_bottom(), DimensionRoundingType::CEIL);
- ARM_COMPUTE_RETURN_ON_ERROR(NEConvolutionLayer::validate(input, weights, bias, output, conv_info, weights_info, Size2D(1U, 1U), ActivationLayerInfo(), enable_fast_math));
+ const PadStrideInfo conv_info(1, 1, upsample_info.pad_left(), upsample_info.pad_right(),
+ upsample_info.pad_top(), upsample_info.pad_bottom(), DimensionRoundingType::CEIL);
+ ARM_COMPUTE_RETURN_ON_ERROR(NEConvolutionLayer::validate(input, weights, bias, output, conv_info, weights_info,
+ Size2D(1U, 1U), ActivationLayerInfo(),
+ enable_fast_math));
}
return Status{};
}
-void NEDeconvolutionLayer::configure(ITensor *input, const ITensor *weights, const ITensor *bias, ITensor *output, const PadStrideInfo &info, bool enable_fast_math, const WeightsInfo &weights_info)
+void NEDeconvolutionLayer::configure(ITensor *input,
+ const ITensor *weights,
+ const ITensor *bias,
+ ITensor *output,
+ const PadStrideInfo &info,
+ bool enable_fast_math,
+ const WeightsInfo &weights_info)
{
// Perform validation step
ARM_COMPUTE_ERROR_ON_NULLPTR(input, weights, output);
- ARM_COMPUTE_ERROR_THROW_ON(NEDeconvolutionLayer::validate(input->info(), weights->info(), (bias == nullptr) ? nullptr : bias->info(), output->info(), info, enable_fast_math, weights_info));
+ ARM_COMPUTE_ERROR_THROW_ON(NEDeconvolutionLayer::validate(input->info(), weights->info(),
+ (bias == nullptr) ? nullptr : bias->info(),
+ output->info(), info, enable_fast_math, weights_info));
ARM_COMPUTE_LOG_PARAMS(input, weights, bias, output, info, enable_fast_math, weights_info);
const DataLayout data_layout = input->info()->data_layout();
const unsigned int width_idx = get_data_layout_dimension_index(data_layout, DataLayoutDimension::WIDTH);
const unsigned int height_idx = get_data_layout_dimension_index(data_layout, DataLayoutDimension::HEIGHT);
- auto out_dims = deconvolution_output_dimensions(input->info()->dimension(width_idx), input->info()->dimension(height_idx),
- weights->info()->dimension(width_idx), weights->info()->dimension(height_idx), info);
+ auto out_dims = deconvolution_output_dimensions(
+ input->info()->dimension(width_idx), input->info()->dimension(height_idx),
+ weights->info()->dimension(width_idx), weights->info()->dimension(height_idx), info);
const TensorShape output_shape = compute_deconvolution_output_shape(out_dims, *input->info(), *weights->info());
@@ -191,7 +222,8 @@
const unsigned int stride_y = info.stride().second;
// Output auto initialization if not yet initialized
- auto_init_if_empty(*output->info(), output_shape, 1, input->info()->data_type(), input->info()->quantization_info());
+ auto_init_if_empty(*output->info(), output_shape, 1, input->info()->data_type(),
+ input->info()->quantization_info());
_flip_axis.allocator()->init(TensorInfo(TensorShape(2U), 1, DataType::U32));
@@ -199,12 +231,11 @@
_flip_weights.configure(weights, &_weights_flipped, &_flip_axis);
// setup the function to convolve the upscaled output
- uint32_t deconv_pad_x = 0;
- uint32_t deconv_pad_y = 0;
- const TensorShape scale_out_shape = compute_deconvolution_upsampled_shape(*input->info(), *weights->info(),
- stride_x, stride_y,
- out_dims, deconv_pad_x, deconv_pad_y);
- const PadStrideInfo upsample_info = compute_upsample_info(info, deconv_pad_x, deconv_pad_y);
+ uint32_t deconv_pad_x = 0;
+ uint32_t deconv_pad_y = 0;
+ const TensorShape scale_out_shape = compute_deconvolution_upsampled_shape(
+ *input->info(), *weights->info(), stride_x, stride_y, out_dims, deconv_pad_x, deconv_pad_y);
+ const PadStrideInfo upsample_info = compute_upsample_info(info, deconv_pad_x, deconv_pad_y);
// Do not perform upsampling when the operation uses unit stride in all dimensions
_do_upsampling = stride_x != 1 || stride_y != 1;
@@ -216,12 +247,12 @@
axis_data[1] = static_cast<uint32_t>(height_idx);
// Setup convolution and upsampling, if needed
- if(_do_upsampling)
+ if (_do_upsampling)
{
_memory_group.manage(&_scaled_output);
const PadStrideInfo conv_info(1, 1, 0, 0, 0, 0, DimensionRoundingType::CEIL);
- TensorInfo scale_out_info(scale_out_shape, 1, input->info()->data_type(), input->info()->quantization_info());
+ TensorInfo scale_out_info(scale_out_shape, 1, input->info()->data_type(), input->info()->quantization_info());
scale_out_info.set_data_layout(data_layout);
_scaled_output.allocator()->init(scale_out_info);
@@ -229,14 +260,17 @@
// The padding amount can be given as input to the convolution layer.
_upsample_f.configure(input, &_scaled_output, upsample_info);
- _conv_f.configure(&_scaled_output, &_weights_flipped, bias, output, conv_info, weights_info, Size2D(1U, 1U), ActivationLayerInfo(), enable_fast_math);
+ _conv_f.configure(&_scaled_output, &_weights_flipped, bias, output, conv_info, weights_info, Size2D(1U, 1U),
+ ActivationLayerInfo(), enable_fast_math);
_scaled_output.allocator()->allocate();
}
else
{
- const PadStrideInfo conv_info(1, 1, upsample_info.pad_left(), upsample_info.pad_right(), upsample_info.pad_top(), upsample_info.pad_bottom(), DimensionRoundingType::CEIL);
- _conv_f.configure(input, &_weights_flipped, bias, output, conv_info, weights_info, Size2D(1U, 1U), ActivationLayerInfo(), enable_fast_math);
+ const PadStrideInfo conv_info(1, 1, upsample_info.pad_left(), upsample_info.pad_right(),
+ upsample_info.pad_top(), upsample_info.pad_bottom(), DimensionRoundingType::CEIL);
+ _conv_f.configure(input, &_weights_flipped, bias, output, conv_info, weights_info, Size2D(1U, 1U),
+ ActivationLayerInfo(), enable_fast_math);
}
}
@@ -246,7 +280,7 @@
MemoryGroupResourceScope scope_mg(_memory_group);
- if(_do_upsampling)
+ if (_do_upsampling)
{
_upsample_f.run();
}
@@ -255,7 +289,7 @@
void NEDeconvolutionLayer::prepare()
{
- if(!_is_prepared)
+ if (!_is_prepared)
{
ARM_COMPUTE_ERROR_ON(!_original_weights->is_used());