diff options
Diffstat (limited to 'arm_compute/runtime/CPP/functions')
8 files changed, 99 insertions, 63 deletions
diff --git a/arm_compute/runtime/CPP/functions/CPPBoxWithNonMaximaSuppressionLimit.h b/arm_compute/runtime/CPP/functions/CPPBoxWithNonMaximaSuppressionLimit.h index 58b4bf25cc..9af4ed6208 100644 --- a/arm_compute/runtime/CPP/functions/CPPBoxWithNonMaximaSuppressionLimit.h +++ b/arm_compute/runtime/CPP/functions/CPPBoxWithNonMaximaSuppressionLimit.h @@ -61,8 +61,16 @@ public: * @param[in] keeps_size (Optional) Number of filtered indices per class tensor of size [num_classes]. Data types supported: U32. * @param[in] info (Optional) BoxNMSLimitInfo information. */ - void configure(const ITensor *scores_in, const ITensor *boxes_in, const ITensor *batch_splits_in, ITensor *scores_out, ITensor *boxes_out, ITensor *classes, - ITensor *batch_splits_out = nullptr, ITensor *keeps = nullptr, ITensor *keeps_size = nullptr, const BoxNMSLimitInfo info = BoxNMSLimitInfo()); + void configure(const ITensor *scores_in, + const ITensor *boxes_in, + const ITensor *batch_splits_in, + ITensor *scores_out, + ITensor *boxes_out, + ITensor *classes, + ITensor *batch_splits_out = nullptr, + ITensor *keeps = nullptr, + ITensor *keeps_size = nullptr, + const BoxNMSLimitInfo info = BoxNMSLimitInfo()); /** Static function to check if given info will lead to a valid configuration of @ref CPPDetectionOutputLayer * * @param[in] scores_in The scores input tensor of size [count, num_classes]. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32 @@ -81,9 +89,16 @@ public: * * @return a status */ - static Status validate(const ITensorInfo *scores_in, const ITensorInfo *boxes_in, const ITensorInfo *batch_splits_in, const ITensorInfo *scores_out, const ITensorInfo *boxes_out, - const ITensorInfo *classes, - const ITensorInfo *batch_splits_out = nullptr, const ITensorInfo *keeps = nullptr, const ITensorInfo *keeps_size = nullptr, const BoxNMSLimitInfo info = BoxNMSLimitInfo()); + static Status validate(const ITensorInfo *scores_in, + const ITensorInfo *boxes_in, + const ITensorInfo *batch_splits_in, + const ITensorInfo *scores_out, + const ITensorInfo *boxes_out, + const ITensorInfo *classes, + const ITensorInfo *batch_splits_out = nullptr, + const ITensorInfo *keeps = nullptr, + const ITensorInfo *keeps_size = nullptr, + const BoxNMSLimitInfo info = BoxNMSLimitInfo()); // Inherited methods overridden: void run() override; diff --git a/arm_compute/runtime/CPP/functions/CPPDetectionOutputLayer.h b/arm_compute/runtime/CPP/functions/CPPDetectionOutputLayer.h index f2c7ccccc5..dc8c8e76ba 100644 --- a/arm_compute/runtime/CPP/functions/CPPDetectionOutputLayer.h +++ b/arm_compute/runtime/CPP/functions/CPPDetectionOutputLayer.h @@ -24,9 +24,8 @@ #ifndef ARM_COMPUTE_CPP_DETECTION_OUTPUT_LAYER_H #define ARM_COMPUTE_CPP_DETECTION_OUTPUT_LAYER_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" namespace arm_compute { @@ -52,7 +51,11 @@ public: * * @note Output contains all the detections. Of those, only the ones selected by the valid region are valid. */ - void configure(const ITensor *input_loc, const ITensor *input_conf, const ITensor *input_priorbox, ITensor *output, DetectionOutputLayerInfo info = DetectionOutputLayerInfo()); + void configure(const ITensor *input_loc, + const ITensor *input_conf, + const ITensor *input_priorbox, + ITensor *output, + DetectionOutputLayerInfo info = DetectionOutputLayerInfo()); /** Static function to check if given info will lead to a valid configuration of @ref CPPDetectionOutputLayer * * @param[in] input_loc The mbox location input tensor info. Data types supported: F32. @@ -63,7 +66,10 @@ public: * * @return a status */ - static Status validate(const ITensorInfo *input_loc, const ITensorInfo *input_conf, const ITensorInfo *input_priorbox, const ITensorInfo *output, + static Status validate(const ITensorInfo *input_loc, + const ITensorInfo *input_conf, + const ITensorInfo *input_priorbox, + const ITensorInfo *output, DetectionOutputLayerInfo info = DetectionOutputLayerInfo()); // Inherited methods overridden: void run() override; @@ -82,12 +88,12 @@ private: int _num_priors; int _num; - std::vector<LabelBBox> _all_location_predictions; + std::vector<LabelBBox> _all_location_predictions; std::vector<std::map<int, std::vector<float>>> _all_confidence_scores; - std::vector<BBox> _all_prior_bboxes; - std::vector<std::array<float, 4>> _all_prior_variances; - std::vector<LabelBBox> _all_decode_bboxes; - std::vector<std::map<int, std::vector<int>>> _all_indices; + std::vector<BBox> _all_prior_bboxes; + std::vector<std::array<float, 4>> _all_prior_variances; + std::vector<LabelBBox> _all_decode_bboxes; + std::vector<std::map<int, std::vector<int>>> _all_indices; }; } // namespace arm_compute #endif /* ARM_COMPUTE_CPP_DETECTION_OUTPUT_LAYER_H */ diff --git a/arm_compute/runtime/CPP/functions/CPPDetectionPostProcessLayer.h b/arm_compute/runtime/CPP/functions/CPPDetectionPostProcessLayer.h index 94248ff314..a40e4f9ecb 100644 --- a/arm_compute/runtime/CPP/functions/CPPDetectionPostProcessLayer.h +++ b/arm_compute/runtime/CPP/functions/CPPDetectionPostProcessLayer.h @@ -24,10 +24,9 @@ #ifndef ARM_COMPUTE_CPP_DETECTION_POSTPROCESS_H #define ARM_COMPUTE_CPP_DETECTION_POSTPROCESS_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" #include "arm_compute/runtime/CPP/functions/CPPNonMaximumSuppression.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" #include "arm_compute/runtime/IMemoryManager.h" #include "arm_compute/runtime/MemoryGroup.h" #include "arm_compute/runtime/Tensor.h" @@ -65,8 +64,14 @@ public: * * @note Output contains all the detections. Of those, only the ones selected by the valid region are valid. */ - void configure(const ITensor *input_box_encoding, const ITensor *input_score, const ITensor *input_anchors, - ITensor *output_boxes, ITensor *output_classes, ITensor *output_scores, ITensor *num_detection, DetectionPostProcessLayerInfo info = DetectionPostProcessLayerInfo()); + void configure(const ITensor *input_box_encoding, + const ITensor *input_score, + const ITensor *input_anchors, + ITensor *output_boxes, + ITensor *output_classes, + ITensor *output_scores, + ITensor *num_detection, + DetectionPostProcessLayerInfo info = DetectionPostProcessLayerInfo()); /** Static function to check if given info will lead to a valid configuration of @ref CPPDetectionPostProcessLayer * * @param[in] input_box_encoding The bounding box input tensor info. Data types supported: F32/QASYMM8/QASYMM8_SIGNED. @@ -80,8 +85,13 @@ public: * * @return a status */ - static Status validate(const ITensorInfo *input_box_encoding, const ITensorInfo *input_class_score, const ITensorInfo *input_anchors, - ITensorInfo *output_boxes, ITensorInfo *output_classes, ITensorInfo *output_scores, ITensorInfo *num_detection, + static Status validate(const ITensorInfo *input_box_encoding, + const ITensorInfo *input_class_score, + const ITensorInfo *input_anchors, + ITensorInfo *output_boxes, + ITensorInfo *output_classes, + ITensorInfo *output_scores, + ITensorInfo *num_detection, DetectionPostProcessLayerInfo info = DetectionPostProcessLayerInfo()); // Inherited methods overridden: void run() override; diff --git a/arm_compute/runtime/CPP/functions/CPPNonMaximumSuppression.h b/arm_compute/runtime/CPP/functions/CPPNonMaximumSuppression.h index 71c44a8bd1..af6afc6029 100644 --- a/arm_compute/runtime/CPP/functions/CPPNonMaximumSuppression.h +++ b/arm_compute/runtime/CPP/functions/CPPNonMaximumSuppression.h @@ -24,9 +24,8 @@ #ifndef ARM_COMPUTE_CPP_NONMAXIMUMSUPPRESSION_LAYER_H #define ARM_COMPUTE_CPP_NONMAXIMUMSUPPRESSION_LAYER_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" namespace arm_compute { @@ -48,7 +47,12 @@ public: * @param[in] nms_threshold The threshold used in non maximum suppression. * */ - void configure(const ITensor *bboxes, const ITensor *scores, ITensor *indices, unsigned int max_output_size, const float score_threshold, const float nms_threshold); + void configure(const ITensor *bboxes, + const ITensor *scores, + ITensor *indices, + unsigned int max_output_size, + const float score_threshold, + const float nms_threshold); /** Static function to check if given arguments will lead to a valid configuration of @ref CPPNonMaximumSuppression * @@ -60,8 +64,12 @@ public: * @param[in] nms_threshold The threshold used in non maximum suppression. * */ - static Status validate(const ITensorInfo *bboxes, const ITensorInfo *scores, const ITensorInfo *indices, unsigned int max_output_size, - const float score_threshold, const float nms_threshold); + static Status validate(const ITensorInfo *bboxes, + const ITensorInfo *scores, + const ITensorInfo *indices, + unsigned int max_output_size, + const float score_threshold, + const float nms_threshold); }; } // namespace arm_compute #endif /* ARM_COMPUTE_CPP_NONMAXIMUMSUPPRESSION_LAYER_H */ diff --git a/arm_compute/runtime/CPP/functions/CPPPermute.h b/arm_compute/runtime/CPP/functions/CPPPermute.h index 85c1502324..232da41b8e 100644 --- a/arm_compute/runtime/CPP/functions/CPPPermute.h +++ b/arm_compute/runtime/CPP/functions/CPPPermute.h @@ -24,9 +24,8 @@ #ifndef ARM_COMPUTE_CPPPERMUTE_H #define ARM_COMPUTE_CPPPERMUTE_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" namespace arm_compute { @@ -53,5 +52,5 @@ public: */ static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PermutationVector &perm); }; -} +} // namespace arm_compute #endif /* ARM_COMPUTE_CPPPERMUTE_H */ diff --git a/arm_compute/runtime/CPP/functions/CPPSplit.h b/arm_compute/runtime/CPP/functions/CPPSplit.h index b2b4d07c86..9be081f5bb 100644 --- a/arm_compute/runtime/CPP/functions/CPPSplit.h +++ b/arm_compute/runtime/CPP/functions/CPPSplit.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2020 Arm Limited. + * Copyright (c) 2020-2021,2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -29,9 +29,6 @@ #include "arm_compute/core/TensorInfo.h" #include "arm_compute/core/Types.h" #include "arm_compute/core/utils/misc/ShapeCalculator.h" - -#include "support/ToolchainSupport.h" - #include "arm_compute/runtime/IFunction.h" namespace arm_compute @@ -41,8 +38,7 @@ template <typename SliceType, typename TensorInterfaceType = ITensor> class CPPSplit : public IFunction { public: - CPPSplit() - : _outputs_vector(), _slice_functions(), _num_outputs(0) + CPPSplit() : _outputs_vector(), _slice_functions(), _num_outputs(0) { } /** Static function to check if given info will lead to a valid configuration of @ref CPPSplit @@ -66,14 +62,16 @@ public: unsigned int total_output_shape_size = 0; // Sum the output sizes and fall back to evenly-sized splits if any are zero - const bool using_split_shapes = std::none_of(outputs.begin(), outputs.end(), [&total_output_shape_size](ITensorInfo * info) - { - unsigned int output_shape_size = info->tensor_shape().total_size(); - total_output_shape_size += output_shape_size; - return output_shape_size == 0; - }); - - if(using_split_shapes) + const bool using_split_shapes = std::none_of(outputs.begin(), outputs.end(), + [&total_output_shape_size](ITensorInfo *info) + { + unsigned int output_shape_size = + info->tensor_shape().total_size(); + total_output_shape_size += output_shape_size; + return output_shape_size == 0; + }); + + if (using_split_shapes) { ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape().total_size() != total_output_shape_size); } @@ -85,10 +83,10 @@ public: // Validate output tensors unsigned int axis_offset = 0; - for(const auto &output : outputs) + for (const auto &output : outputs) { ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(output); - if(using_split_shapes) + if (using_split_shapes) { output_shape = output->tensor_shape(); ARM_COMPUTE_RETURN_ERROR_ON(output_shape.total_size() == 0); @@ -99,14 +97,14 @@ public: // Start/End coordinates Coordinates start_coords; Coordinates end_coords; - for(unsigned int d = 0; d < output_shape.num_dimensions(); ++d) + for (unsigned int d = 0; d < output_shape.num_dimensions(); ++d) { end_coords.set(d, -1); } // Output auto inizialitation if not yet initialized TensorInfo tmp_output_info = *output->clone(); - if(tmp_output_info.tensor_shape().total_size() == 0) + if (tmp_output_info.tensor_shape().total_size() == 0) { tmp_output_info = input->clone()->set_is_resizable(true).set_tensor_shape(output_shape); } @@ -130,7 +128,8 @@ public: * from the split dimension. * @param[in] axis Axis on which to split the input. */ - void configure(const TensorInterfaceType *input, const std::vector<TensorInterfaceType *> &outputs, unsigned int axis) + void + configure(const TensorInterfaceType *input, const std::vector<TensorInterfaceType *> &outputs, unsigned int axis) { // Create Slice functions _num_outputs = outputs.size(); @@ -138,17 +137,16 @@ public: // Extract output tensor info std::vector<ITensorInfo *> outputs_info; - for(auto &output : outputs) + for (auto &output : outputs) { ARM_COMPUTE_ERROR_ON_NULLPTR(output); outputs_info.emplace_back(output->info()); } // If any of the outputs have a zero size, fall-back to using evenly-sized output splits - const bool outputs_have_sizes = std::none_of(outputs_info.begin(), outputs_info.end(), [](ITensorInfo * info) - { - return info->tensor_shape().total_size() == 0; - }); + const bool outputs_have_sizes = + std::none_of(outputs_info.begin(), outputs_info.end(), + [](ITensorInfo *info) { return info->tensor_shape().total_size() == 0; }); // Validate ARM_COMPUTE_ERROR_THROW_ON(CPPSplit::validate(input->info(), outputs_info, axis)); @@ -156,12 +154,13 @@ public: unsigned int axis_offset = 0; unsigned int i = 0; - for(const auto &output_info : outputs_info) + for (const auto &output_info : outputs_info) { // Get output shape - TensorShape output_shape = (outputs_have_sizes ? - output_info->tensor_shape() : - arm_compute::misc::shape_calculator::compute_split_shape(input->info(), axis, _num_outputs)); + TensorShape output_shape = + (outputs_have_sizes + ? output_info->tensor_shape() + : arm_compute::misc::shape_calculator::compute_split_shape(input->info(), axis, _num_outputs)); const size_t axis_split_step = output_shape[axis]; @@ -169,7 +168,7 @@ public: Coordinates start_coords; Coordinates end_coords; - for(unsigned int d = 0; d < output_shape.num_dimensions(); ++d) + for (unsigned int d = 0; d < output_shape.num_dimensions(); ++d) { end_coords.set(d, -1); } diff --git a/arm_compute/runtime/CPP/functions/CPPTopKV.h b/arm_compute/runtime/CPP/functions/CPPTopKV.h index 2f63084056..232cbb3067 100644 --- a/arm_compute/runtime/CPP/functions/CPPTopKV.h +++ b/arm_compute/runtime/CPP/functions/CPPTopKV.h @@ -24,9 +24,8 @@ #ifndef ARM_COMPUTE_CPPTOPKV_H #define ARM_COMPUTE_CPPTOPKV_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" namespace arm_compute { @@ -54,7 +53,8 @@ public: * * @return a status */ - static Status validate(const ITensorInfo *predictions, const ITensorInfo *targets, ITensorInfo *output, const unsigned int k); + static Status + validate(const ITensorInfo *predictions, const ITensorInfo *targets, ITensorInfo *output, const unsigned int k); }; } // namespace arm_compute #endif /* ARM_COMPUTE_CPPTOPKV_H */ diff --git a/arm_compute/runtime/CPP/functions/CPPUpsample.h b/arm_compute/runtime/CPP/functions/CPPUpsample.h index b97d4d1cc1..3b0f997b17 100644 --- a/arm_compute/runtime/CPP/functions/CPPUpsample.h +++ b/arm_compute/runtime/CPP/functions/CPPUpsample.h @@ -24,9 +24,8 @@ #ifndef ARM_COMPUTE_CPPUPSAMPLE_H #define ARM_COMPUTE_CPPUPSAMPLE_H -#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" - #include "arm_compute/core/Types.h" +#include "arm_compute/runtime/CPP/ICPPSimpleFunction.h" namespace arm_compute { @@ -44,5 +43,5 @@ public: */ void configure(const ITensor *input, ITensor *output, const PadStrideInfo &info); }; -} +} // namespace arm_compute #endif /* ARM_COMPUTE_CPPUPSAMPLE_H */ |