/* * Copyright (c) 2018-2021 Arm Limited. * * SPDX-License-Identifier: MIT * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to * deal in the Software without restriction, including without limitation the * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or * sell copies of the Software, and to permit persons to whom the Software is * furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice shall be included in all * copies or substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. */ #ifndef ARM_COMPUTE_GRAPH_LAYERS_H #define ARM_COMPUTE_GRAPH_LAYERS_H #include "arm_compute/graph/GraphBuilder.h" #include "arm_compute/graph/Types.h" #include "arm_compute/graph/frontend/ILayer.h" #include "arm_compute/graph/frontend/IStream.h" #include "arm_compute/graph/frontend/SubStream.h" #include "arm_compute/core/utils/misc/Utility.h" #include #include namespace arm_compute { namespace graph { namespace frontend { /** Input Layer */ class InputLayer final : public ILayer { public: /** Construct an input layer. * * @param[in] desc Description of input tensor. * @param[in] accessor Accessor to get input tensor data from. */ InputLayer(TensorDescriptor desc, ITensorAccessorUPtr accessor) : _desc(desc), _accessor(std::move(accessor)) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; return GraphBuilder::add_input_node(s.graph(), common_params, _desc, std::move(_accessor)); } private: TensorDescriptor _desc; ITensorAccessorUPtr _accessor; }; /** Constant Layer */ class ConstantLayer final : public ILayer { public: /** Construct a constant layer. * * @param[in] desc Description of input tensor. * @param[in] accessor Accessor to get input tensor data from. */ ConstantLayer(TensorDescriptor desc, ITensorAccessorUPtr accessor) : _desc(desc), _accessor(std::move(accessor)) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; return GraphBuilder::add_const_node(s.graph(), common_params, _desc, std::move(_accessor)); } private: TensorDescriptor _desc; ITensorAccessorUPtr _accessor; }; /** Output Layer */ class OutputLayer final : public ILayer { public: /** Construct an output layer. * * @param[in] accessor Accessor to give output tensor data to. * @param[in] connection_idx (Optional) Input connection index */ OutputLayer(ITensorAccessorUPtr accessor, unsigned int connection_idx = 0) : _accessor(std::move(accessor)), _connection_idx(connection_idx) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), _connection_idx }; return GraphBuilder::add_output_node(s.graph(), common_params, input, std::move(_accessor)); } private: ITensorAccessorUPtr _accessor; unsigned int _connection_idx; }; /** Activation Layer */ class ActivationLayer final : public ILayer { public: /** Construct an activation layer. * * @param[in] act_info Activation information * @param[in] out_quant_info (Optional) Output quantization info */ ActivationLayer(ActivationLayerInfo act_info, const QuantizationInfo out_quant_info = QuantizationInfo()) : _act_info(act_info), _out_quant_info(std::move(out_quant_info)) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_activation_node(s.graph(), common_params, input, _act_info, std::move(_out_quant_info)); } private: ActivationLayerInfo _act_info; const QuantizationInfo _out_quant_info; }; /** ArgMinMax Layer */ class ArgMinMaxLayer final : public ILayer { public: /** Construct an activation layer. * * @param[in] op Reduction Operation: min or max * @param[in] axis Axis to perform reduction along * @param[in] out_data_type (Optional) Output tensor data type * @param[in] out_quant_info (Optional) Output quantization info */ ArgMinMaxLayer(ReductionOperation op, unsigned int axis, DataType out_data_type = DataType::UNKNOWN, const QuantizationInfo out_quant_info = QuantizationInfo()) : _op(op), _axis(axis), _out_data_type(out_data_type), _out_quant_info(std::move(out_quant_info)) { } /** Create layer and add to the given stream. * * @param[in] s Stream to add layer to. * * @return ID of the created node. */ NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_arg_min_max_node(s.graph(), common_params, input, _op, _axis, _out_data_type, std::move(_out_quant_info)); } private: ReductionOperation _op; unsigned int _axis; DataType _out_data_type; QuantizationInfo _out_quant_info; }; /** Batchnormalization Layer */ class BatchNormalizationLayer final : public ILayer { public: /** Construct a batch normalization layer. * * @param[in] mean Accessor to get mean tensor data from. * @param[in] var Accessor to get var tensor data from. * @param[in] gamma (Optional) Accessor to get gamma tensor data from. Default: nullptr. * @param[in] beta (Optional) Accessor to get beta tensor data from. Default: nullptr. * @param[in] epsilon (Optional) Epsilon value. Default: 0.001. */ BatchNormalizationLayer(ITensorAccessorUPtr mean, ITensorAccessorUPtr var, ITensorAccessorUPtr gamma = nullptr, ITensorAccessorUPtr beta = nullptr, float epsilon = 0.001f) : _mean(std::move(mean)), _var(std::move(var)), _gamma(std::move(gamma)), _beta(std::move(beta)), _epsilon(epsilon) { } NodeID create_layer(IStream &s) override { ARM_COMPUTE_ERROR_ON(_mean == nullptr); ARM_COMPUTE_ERROR_ON(_var == nullptr); NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_batch_normalization_node(s.graph(), common_params, input, _epsilon, std::move(_mean), std::move(_var), std::move(_beta), std::move(_gamma)); } private: ITensorAccessorUPtr _mean; ITensorAccessorUPtr _var; ITensorAccessorUPtr _gamma; ITensorAccessorUPtr _beta; float _epsilon; }; /** Bounding Box Transform Layer */ class BoundingBoxTransformLayer final : public ILayer { public: /** Construct a bounding box transform layer. * * @param[in] sub_stream_input Graph sub-stream for the input * @param[in] sub_stream_deltas Graph sub-stream for the deltas * @param[in] info Contains BoundingBox operation information described in @ref BoundingBoxTransformInfo. */ BoundingBoxTransformLayer(SubStream &&sub_stream_input, SubStream &&sub_stream_deltas, BoundingBoxTransformInfo info) : _ss_input(sub_stream_input), _ss_deltas(sub_stream_deltas), _bbox_info(info) { } /** Create layer and add to the given stream. * * @param[in] s Stream to add layer to. * * @return ID of the created node. */ NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { _ss_input.tail_node(), 0 }; NodeIdxPair deltas = { _ss_deltas.tail_node(), 0 }; return GraphBuilder::add_bounding_box_transform_node(s.graph(), common_params, input, deltas, _bbox_info); } private: SubStream _ss_input; SubStream _ss_deltas; BoundingBoxTransformInfo _bbox_info; }; /** Channel Shuffle Layer */ class ChannelShuffleLayer final : public ILayer { public: /** Construct a Channel Shuffle layer. * * @param[in] num_groups Number of groups */ ChannelShuffleLayer(unsigned int num_groups) : _num_groups(num_groups) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_channel_shuffle_node(s.graph(), common_params, input, _num_groups); } private: unsigned int _num_groups; }; /** Concat Layer */ class ConcatLayer final : public ILayer { public: /** Construct a concatenation layer * * @param[in] sub_stream1 First graph branch * @param[in] sub_stream2 Second graph branch * @param[in] rest_sub_streams Rest sub-graph branches */ template ConcatLayer(SubStream &&sub_stream1, SubStream &&sub_stream2, Ts &&... rest_sub_streams) : _sub_streams(), _concat_descriptor(DataLayoutDimension::CHANNEL) { _sub_streams.push_back(std::make_unique(std::move(sub_stream1))); _sub_streams.push_back(std::make_unique(std::move(sub_stream2))); utility::for_each([&](SubStream && sub_stream) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); }, std::move(rest_sub_streams)...); } /** Construct a concatenation layer * * @param[in] concat_descriptor Concat layer descriptor * @param[in] sub_stream1 First graph branch * @param[in] sub_stream2 Second graph branch * @param[in] rest_sub_streams Rest sub-graph branches */ template ConcatLayer(descriptors::ConcatLayerDescriptor concat_descriptor, SubStream &&sub_stream1, SubStream &&sub_stream2, Ts &&... rest_sub_streams) : _sub_streams(), _concat_descriptor(concat_descriptor) { _sub_streams.push_back(std::make_unique(std::move(sub_stream1))); _sub_streams.push_back(std::make_unique(std::move(sub_stream2))); utility::for_each([&](SubStream && sub_stream) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); }, std::move(rest_sub_streams)...); } /** Construct a concat layer * * @param[in] sub_stream Sub-stream */ template ConcatLayer(SubStream &&sub_stream) : _sub_streams(), _concat_descriptor(DataLayoutDimension::CHANNEL) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); } NodeID create_layer(IStream &s) override { NodeID nid = EmptyNodeID; NodeParams common_params = { name(), s.hints().target_hint }; if(_sub_streams.size() == 1 && _sub_streams.at(0) != nullptr) { nid = _sub_streams[0]->tail_node(); } else { // Collect tail nodes and concatenate std::vector nodes; for(auto &ss : _sub_streams) { if(ss && (ss->tail_node() != EmptyNodeID)) { const auto tail_node = s.graph().node(ss->tail_node()); if(tail_node != nullptr && tail_node->type() != NodeType::Output) { nodes.push_back({ ss->tail_node(), 0 }); } } } nid = GraphBuilder::add_concatenate_node(s.graph(), common_params, nodes, _concat_descriptor); } return nid; } private: std::vector> _sub_streams; descriptors::ConcatLayerDescriptor _concat_descriptor; }; /** Convolution Layer */ class ConvolutionLayer final : public ILayer { public: /** Construct a convolution layer. * * @param[in] conv_width Convolution width. * @param[in] conv_height Convolution height. * @param[in] ofm Output feature map. * @param[in] weights Accessor to get kernel weights from. * @param[in] bias Accessor to get kernel bias from. * @param[in] conv_info Padding and stride information. * @param[in] num_groups (Optional) Number of groups. Default: 1. * @param[in] weights_quant_info (Optional) Weights quantization information * @param[in] out_quant_info (Optional) Output quantization info */ ConvolutionLayer(unsigned int conv_width, unsigned int conv_height, unsigned int ofm, ITensorAccessorUPtr weights, ITensorAccessorUPtr bias, PadStrideInfo conv_info, unsigned int num_groups = 1, const QuantizationInfo weights_quant_info = QuantizationInfo(), const QuantizationInfo out_quant_info = QuantizationInfo()) : _conv_width(conv_width), _conv_height(conv_height), _ofm(ofm), _conv_info(std::move(conv_info)), _num_groups(num_groups), _weights(std::move(weights)), _bias(std::move(bias)), _weights_quant_info(std::move(weights_quant_info)), _out_quant_info(std::move(out_quant_info)) { } NodeID create_layer(IStream &s) override { NodeIdxPair input = { s.tail_node(), 0 }; NodeParams common_params = { name(), s.hints().target_hint }; return GraphBuilder::add_convolution_node(s.graph(), common_params, input, Size2D(_conv_width, _conv_height), _ofm, _conv_info, _num_groups, s.hints().convolution_method_hint, s.hints().fast_math_hint, std::move(_weights), std::move(_bias), std::move(_weights_quant_info), std::move(_out_quant_info)); } private: unsigned int _conv_width; unsigned int _conv_height; unsigned int _ofm; const PadStrideInfo _conv_info; unsigned int _num_groups; ITensorAccessorUPtr _weights; ITensorAccessorUPtr _bias; const QuantizationInfo _weights_quant_info; const QuantizationInfo _out_quant_info; }; /** Deconvolution Layer */ class DeconvolutionLayer final : public ILayer { public: /** Construct a convolution layer. * * @param[in] conv_width Convolution width. * @param[in] conv_height Convolution height. * @param[in] ofm Output feature map. * @param[in] weights Accessor to get kernel weights from. * @param[in] bias Accessor to get kernel bias from. * @param[in] deconv_info Padding and stride information. */ DeconvolutionLayer(unsigned int conv_width, unsigned int conv_height, unsigned int ofm, ITensorAccessorUPtr weights, ITensorAccessorUPtr bias, PadStrideInfo deconv_info) : _conv_width(conv_width), _conv_height(conv_height), _ofm(ofm), _deconv_info(std::move(deconv_info)), _weights(std::move(weights)), _bias(std::move(bias)) { } NodeID create_layer(IStream &s) override { NodeIdxPair input = { s.tail_node(), 0 }; NodeParams common_params = { name(), s.hints().target_hint }; return GraphBuilder::add_deconvolution_node(s.graph(), common_params, input, Size2D(_conv_width, _conv_height), _ofm, _deconv_info, std::move(_weights), std::move(_bias)); } private: unsigned int _conv_width; unsigned int _conv_height; unsigned int _ofm; const PadStrideInfo _deconv_info; ITensorAccessorUPtr _weights; ITensorAccessorUPtr _bias; }; /** Depthwise Convolution Layer */ class DepthwiseConvolutionLayer final : public ILayer { public: /** Construct a depthwise convolution layer. * * @param[in] conv_width Convolution width. * @param[in] conv_height Convolution height. * @param[in] weights Accessor to get kernel weights from. * @param[in] bias Accessor to get kernel bias from. * @param[in] conv_info Padding and stride information. * @param[in] depth_multiplier (Optional) Depth multiplier parameter. * @param[in] weights_quant_info (Optional) Quantization info used for weights * @param[in] out_quant_info (Optional) Output quantization info */ DepthwiseConvolutionLayer(unsigned int conv_width, unsigned int conv_height, ITensorAccessorUPtr weights, ITensorAccessorUPtr bias, PadStrideInfo conv_info, int depth_multiplier = 1, const QuantizationInfo weights_quant_info = QuantizationInfo(), const QuantizationInfo out_quant_info = QuantizationInfo()) : _conv_width(conv_width), _conv_height(conv_height), _conv_info(std::move(conv_info)), _weights(std::move(weights)), _bias(std::move(bias)), _depth_multiplier(depth_multiplier), _weights_quant_info(std::move(weights_quant_info)), _out_quant_info(std::move(out_quant_info)) { } NodeID create_layer(IStream &s) override { NodeIdxPair input = { s.tail_node(), 0 }; NodeParams common_params = { name(), s.hints().target_hint }; return GraphBuilder::add_depthwise_convolution_node(s.graph(), common_params, input, Size2D(_conv_width, _conv_height), _conv_info, _depth_multiplier, s.hints().depthwise_convolution_method_hint, std::move(_weights), std::move(_bias), std::move(_weights_quant_info), std::move(_out_quant_info)); } private: unsigned int _conv_width; unsigned int _conv_height; const PadStrideInfo _conv_info; ITensorAccessorUPtr _weights; ITensorAccessorUPtr _bias; int _depth_multiplier; const QuantizationInfo _weights_quant_info; const QuantizationInfo _out_quant_info; }; /** DepthToSpace Layer */ class DepthToSpaceLayer final : public ILayer { public: /** Construct an DepthToSpace layer. * * @param[in] block_shape Block size to rearranged */ DepthToSpaceLayer(int32_t block_shape) : _block_shape(block_shape) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_depth_to_space_node(s.graph(), common_params, input, _block_shape); } private: int32_t _block_shape; }; /** Dequantization Layer */ class DequantizationLayer final : public ILayer { public: /** Construct a dequantization layer. * */ DequantizationLayer() { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_dequantization_node(s.graph(), common_params, input); } }; /** DetectionOutput Layer */ class DetectionOutputLayer final : public ILayer { public: /** Construct a detection output layer. * * @param[in] sub_stream_conf Confidence graph sub-stream. * @param[in] sub_stream_prior PriorBox graph sub-stream. * @param[in] detect_info DetectionOutput parameters. */ DetectionOutputLayer(SubStream &&sub_stream_conf, SubStream &&sub_stream_prior, const DetectionOutputLayerInfo &detect_info) : _ss_conf(std::move(sub_stream_conf)), _ss_prior(std::move(sub_stream_prior)), _detect_info(detect_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input_loc = { s.tail_node(), 0 }; NodeIdxPair input_conf = { _ss_conf.tail_node(), 0 }; NodeIdxPair input_priorbox = { _ss_prior.tail_node(), 0 }; return GraphBuilder::add_detection_output_node(s.graph(), common_params, input_loc, input_conf, input_priorbox, _detect_info); } private: SubStream _ss_conf; SubStream _ss_prior; DetectionOutputLayerInfo _detect_info; }; /** DetectionOutputPostProcess Layer */ class DetectionPostProcessLayer final : public ILayer { public: /** Construct a detection output layer. * * @param[in] sub_stream_class_prediction Class prediction graph sub-stream. * @param[in] detect_info DetectionOutput parameters. * @param[in] anchors Accessor to get anchors tensor data from. * @param[in] out_quant_info (Optional) Output quantization info */ DetectionPostProcessLayer(SubStream &&sub_stream_class_prediction, DetectionPostProcessLayerInfo detect_info, ITensorAccessorUPtr anchors, const QuantizationInfo out_quant_info = QuantizationInfo()) : _sub_stream_class_prediction(std::move(sub_stream_class_prediction)), _detect_info(detect_info), _anchors(std::move(anchors)), _out_quant_info(std::move(out_quant_info)) { } NodeID create_layer(IStream &s) override { ARM_COMPUTE_ERROR_ON(_anchors == nullptr); NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input_box_encoding = { s.tail_node(), 0 }; NodeIdxPair input_class_prediction = { _sub_stream_class_prediction.tail_node(), 0 }; return GraphBuilder::add_detection_post_process_node(s.graph(), common_params, input_box_encoding, input_class_prediction, _detect_info, std::move(_anchors), std::move(_out_quant_info)); } private: SubStream _sub_stream_class_prediction; DetectionPostProcessLayerInfo _detect_info; ITensorAccessorUPtr _anchors; const QuantizationInfo _out_quant_info; }; /** Dummy Layer */ class DummyLayer final : public ILayer { public: /** Construct a dummy layer. * * @param[in] shape Output shape */ DummyLayer(TensorShape shape) : _shape(shape) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_dummy_node(s.graph(), common_params, input, _shape); } private: TensorShape _shape; }; class EltwiseLayer final : public ILayer { public: /** Construct an element-wise operation layer * * @param[in] sub_stream0 First graph sub-stream * @param[in] sub_stream1 First graph sub-stream * @param[in] op Element-wise operation to perform */ EltwiseLayer(SubStream &&sub_stream0, SubStream &&sub_stream1, EltwiseOperation op) : _ss0(std::move(sub_stream0)), _ss1(std::move(sub_stream1)), _op(op) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input0 = { _ss0.tail_node(), 0 }; NodeIdxPair input1 = { _ss1.tail_node(), 0 }; return GraphBuilder::add_elementwise_node(s.graph(), common_params, input0, input1, _op); } private: SubStream _ss0; SubStream _ss1; EltwiseOperation _op; }; /** Flatten Layer */ class FlattenLayer final : public ILayer { public: /** Construct a flatten layer. */ FlattenLayer() { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_flatten_node(s.graph(), common_params, input); } }; /** Fully Connected Layer */ class FullyConnectedLayer final : public ILayer { public: /** Construct a fully connected layer. * * @param[in] num_outputs Number of outputs. * @param[in] weights Accessor to get weights from. * @param[in] bias Accessor to get bias from. * @param[in] fc_info (Optional) Fully connected layer metadata * @param[in] weights_quant_info (Optional) Weights quantization information * @param[in] out_quant_info (Optional) Output quantization info */ FullyConnectedLayer(unsigned int num_outputs, ITensorAccessorUPtr weights, ITensorAccessorUPtr bias, const FullyConnectedLayerInfo fc_info = FullyConnectedLayerInfo(), const QuantizationInfo weights_quant_info = QuantizationInfo(), const QuantizationInfo out_quant_info = QuantizationInfo()) : _num_outputs(num_outputs), _weights(std::move(weights)), _bias(std::move(bias)), _weights_ss(nullptr), _bias_ss(nullptr), _fc_info(fc_info), _weights_quant_info(std::move(weights_quant_info)), _out_quant_info(std::move(out_quant_info)) { } /** Construct a fully connected layer. * * @param[in] num_outputs Number of outputs. * @param[in] sub_stream_weights Graph sub-stream for the weights. * @param[in] sub_stream_bias Graph sub-stream for the bias. * @param[in] fc_info (Optional) Fully connected layer metadata * @param[in] weights_quant_info (Optional) Weights quantization information * @param[in] out_quant_info (Optional) Output quantization info */ FullyConnectedLayer(unsigned int num_outputs, SubStream sub_stream_weights, SubStream sub_stream_bias, const FullyConnectedLayerInfo fc_info = FullyConnectedLayerInfo(), const QuantizationInfo weights_quant_info = QuantizationInfo(), const QuantizationInfo out_quant_info = QuantizationInfo()) : _num_outputs(num_outputs), _weights(nullptr), _bias(nullptr), _weights_ss(std::make_unique(std::move(sub_stream_weights))), _bias_ss(std::make_unique(std::move(sub_stream_bias))), _fc_info(fc_info), _weights_quant_info(std::move(weights_quant_info)), _out_quant_info(std::move(out_quant_info)) { } /** Create layer and add to the given stream. * * @param[in] s Stream to add layer to. * * @return ID of the created node. */ NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; if(_weights != nullptr) { return GraphBuilder::add_fully_connected_layer(s.graph(), common_params, input, _num_outputs, std::move(_weights), std::move(_bias), _fc_info, std::move(_weights_quant_info), std::move(_out_quant_info), s.hints().fast_math_hint); } else { ARM_COMPUTE_ERROR_ON(_weights_ss == nullptr); NodeID bias_nid = (_bias_ss == nullptr) ? EmptyNodeID : _bias_ss->tail_node(); return GraphBuilder::add_fully_connected_layer(s.graph(), common_params, input, _num_outputs, _weights_ss->tail_node(), bias_nid, _fc_info, std::move(_out_quant_info), s.hints().fast_math_hint); } } private: unsigned int _num_outputs; ITensorAccessorUPtr _weights; ITensorAccessorUPtr _bias; std::unique_ptr _weights_ss; std::unique_ptr _bias_ss; const FullyConnectedLayerInfo _fc_info; const QuantizationInfo _weights_quant_info; const QuantizationInfo _out_quant_info; }; /** Generate Proposals Layer */ class GenerateProposalsLayer final : public ILayer { public: /** Construct a generate proposals layer. * * @param[in] ss_scores Graph sub-stream for the scores. * @param[in] ss_deltas Graph sub-stream for the deltas. * @param[in] ss_anchors Graph sub-stream for the anchors. * @param[in] info Generate Proposals operation information. */ GenerateProposalsLayer(SubStream &&ss_scores, SubStream &&ss_deltas, SubStream &&ss_anchors, GenerateProposalsInfo info) : _ss_scores(std::move(ss_scores)), _ss_deltas(std::move(ss_deltas)), _ss_anchors(std::move(ss_anchors)), _info(info) { } /** Create layer and add to the given stream. * * @param[in] s Stream to add layer to. * * @return ID of the created node. */ NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair scores = { _ss_scores.tail_node(), 0 }; NodeIdxPair deltas = { _ss_deltas.tail_node(), 0 }; NodeIdxPair anchors = { _ss_anchors.tail_node(), 0 }; return GraphBuilder::add_generate_proposals_node(s.graph(), common_params, scores, deltas, anchors, _info); } private: SubStream _ss_scores; SubStream _ss_deltas; SubStream _ss_anchors; GenerateProposalsInfo _info; }; /** L2 Normalize Layer */ class L2NormalizeLayer final : public ILayer { public: /** Construct a L2 Normalize layer. * * @param[in] axis Axis to perform normalization on * @param[in] epsilon Lower bound value for the normalization */ L2NormalizeLayer(int axis, float epsilon) : _axis(axis), _epsilon(epsilon) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_l2_normalize_node(s.graph(), common_params, input, _axis, _epsilon); } private: int _axis; float _epsilon; }; /** Normalization Layer */ class NormalizationLayer final : public ILayer { public: /** Construct a normalization layer. * * @param[in] norm_info Normalization information. */ NormalizationLayer(NormalizationLayerInfo norm_info) : _norm_info(norm_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_normalization_node(s.graph(), common_params, input, _norm_info); } private: NormalizationLayerInfo _norm_info; }; /** Normalize planar YUV Layer */ class NormalizePlanarYUVLayer final : public ILayer { public: /** Construct a normalize planar YUV layer. * * @param[in] mean Accessor to get mean tensor data from. * @param[in] std Accessor to get std tensor data from. */ NormalizePlanarYUVLayer(ITensorAccessorUPtr mean, ITensorAccessorUPtr std) : _mean(std::move(mean)), _std(std::move(std)) { } NodeID create_layer(IStream &s) override { ARM_COMPUTE_ERROR_ON(_mean == nullptr); ARM_COMPUTE_ERROR_ON(_std == nullptr); NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_normalize_planar_yuv_node(s.graph(), common_params, input, std::move(_mean), std::move(_std)); } private: ITensorAccessorUPtr _mean; ITensorAccessorUPtr _std; }; /** Pad Layer */ class PadLayer final : public ILayer { public: /** Construct a pad layer. * * @param[in] padding The padding for each spatial dimension of the input tensor. The pair padding[i] * specifies the front and the end padding in the i-th dimension. * @param[in] pad_value Padding value to use. Defaults to 0. */ PadLayer(PaddingList padding, PixelValue pad_value = PixelValue()) : _padding(padding), _pad_value(pad_value) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_pad_node(s.graph(), common_params, input, _padding, _pad_value); } private: PaddingList _padding; PixelValue _pad_value; }; /** Permute Layer */ class PermuteLayer final : public ILayer { public: /** Construct a permute layer. * * @param[in] perm Permutation vector. * @param[in] layout (Optional) Data layout to assign to permuted tensor. * If UNKNOWN then the input's layout will be used. */ PermuteLayer(PermutationVector perm, DataLayout layout = DataLayout::UNKNOWN) : _perm(perm), _layout(layout) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_permute_node(s.graph(), common_params, input, _perm, _layout); } private: PermutationVector _perm; DataLayout _layout; }; /** Pooling Layer */ class PoolingLayer final : public ILayer { public: /** Construct a pooling layer. * * @param[in] pool_info Pooling information. */ PoolingLayer(PoolingLayerInfo pool_info) : _pool_info(pool_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_pooling_node(s.graph(), common_params, input, _pool_info); } private: PoolingLayerInfo _pool_info; }; /** PRelu Layer */ class PReluLayer final : public ILayer { public: /** Construct an PRelu operation layer * * @param[in] sub_stream0 First graph sub-stream * @param[in] sub_stream1 First graph sub-stream */ PReluLayer(SubStream &&sub_stream0, SubStream &&sub_stream1) : _ss0(std::move(sub_stream0)), _ss1(std::move(sub_stream1)) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { _ss0.tail_node(), 0 }; NodeIdxPair alpha = { _ss1.tail_node(), 0 }; return GraphBuilder::add_prelu_node(s.graph(), common_params, input, alpha); } private: SubStream _ss0; SubStream _ss1; }; /** Print Layer */ class PrintLayer final : public ILayer { public: /** Construct a print layer. * * Example usage to locally dequantize and print a tensor: * * Tensor *output = new Tensor(); * const auto transform = [output](ITensor *input) * { * output->allocator()->init(*input->info()); * output->info()->set_data_type(DataType::F32); * output->allocator()->allocate(); * * Window win; * win.use_tensor_dimensions(input->info()->tensor_shape()); * Iterator in(input, win); * Iterator out(output, win); * execute_window_loop(win, [&](const Coordinates &) * { * *(reinterpret_cast(out.ptr())) = dequantize_qasymm8(*in.ptr(), input->info()->quantization_info().uniform()); * }, in, out); * * return output; * }; * * graph << InputLayer(input_descriptor.set_quantization_info(in_quant_info), get_input_accessor(common_params, nullptr, false)) * << ... * << \\ CNN Layers * << ... * << PrintLayer(std::cout, IOFormatInfo(), transform) * << ... * << OutputLayer(get_output_accessor(common_params, 5)); * * @param[in] stream Output stream. * @param[in] format_info (Optional) Format info. * @param[in] transform (Optional) Input transform function. */ PrintLayer(std::ostream &stream, const IOFormatInfo &format_info = IOFormatInfo(), const std::function transform = nullptr) : _stream(stream), _format_info(format_info), _transform(transform) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_print_node(s.graph(), common_params, input, _stream, _format_info, _transform); } private: std::ostream &_stream; const IOFormatInfo &_format_info; const std::function _transform; }; /** PriorBox Layer */ class PriorBoxLayer final : public ILayer { public: /** Construct a priorbox layer. * * @param[in] sub_stream First graph sub-stream * @param[in] prior_info PriorBox parameters. */ PriorBoxLayer(SubStream &&sub_stream, const PriorBoxLayerInfo &prior_info) : _ss(std::move(sub_stream)), _prior_info(prior_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input0 = { s.tail_node(), 0 }; NodeIdxPair input1 = { _ss.tail_node(), 0 }; return GraphBuilder::add_priorbox_node(s.graph(), common_params, input0, input1, _prior_info); } private: SubStream _ss; PriorBoxLayerInfo _prior_info; }; /** Quantization Layer */ class QuantizationLayer final : public ILayer { public: /** Construct a quantization layer. * * @param[in] out_quant_info Output tensor quantization info */ QuantizationLayer(QuantizationInfo out_quant_info) : _out_quant_info(out_quant_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_quantization_node(s.graph(), common_params, input, _out_quant_info); } private: QuantizationInfo _out_quant_info; }; /** Reduction Layer */ class ReductionLayer final : public ILayer { public: /** Construct a reduction layer. * * @param[in] op Reduction operation * @param[in] axis Reduction axis * @param[in] keep_dims (Optional) Whether to keep the reduced dimension after the operation. Defaults to true. */ ReductionLayer(ReductionOperation op, unsigned int axis, bool keep_dims) : _op(op), _axis(axis), _keep_dims(keep_dims) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_reduction_operation_node(s.graph(), common_params, input, _op, _axis, _keep_dims); } private: ReductionOperation _op; unsigned int _axis; bool _keep_dims; }; /** Reorg Layer */ class ReorgLayer final : public ILayer { public: /** Construct a reorg layer. * * @param[in] stride Stride value to use for reorganizing the values in the output tensor. * It defines the spatial distance between 2 consecutive pixels in the x and y direction */ ReorgLayer(int stride) : _stride(stride) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_reorg_node(s.graph(), common_params, input, _stride); } private: int _stride; }; /** Reshape Layer */ class ReshapeLayer final : public ILayer { public: /** Construct a reshape layer. * * @param[in] shape Target shape. */ ReshapeLayer(TensorShape shape) : _shape(shape) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_reshape_node(s.graph(), common_params, input, _shape); } private: TensorShape _shape; }; /** Resize Layer */ class ResizeLayer final : public ILayer { public: ResizeLayer(InterpolationPolicy policy, float width_scale, float height_scale) : _policy(policy), _width_scale(width_scale), _height_scale(height_scale) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_resize_node(s.graph(), common_params, input, _policy, _width_scale, _height_scale); } private: InterpolationPolicy _policy; float _width_scale; float _height_scale; }; /** ROIAlign Layer */ class ROIAlignLayer final : public ILayer { public: /** Construct a RoiAlign layer. * * @param[in] sub_stream_input Graph sub-stream for the input * @param[in] sub_stream_rois Graph sub-stream for the rois * @param[in] pool_info Pooling information. */ ROIAlignLayer(SubStream &&sub_stream_input, SubStream &&sub_stream_rois, ROIPoolingLayerInfo pool_info) : _ss_input(sub_stream_input), _ss_rois(sub_stream_rois), _pool_info(pool_info) { } /** Prevent instances of this class from being copy constructed */ ROIAlignLayer(const ROIAlignLayer &) = delete; /** Prevent instances of this class from being copied */ ROIAlignLayer &operator=(const ROIAlignLayer &) = delete; NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { _ss_input.tail_node(), 0 }; NodeIdxPair rois = { _ss_rois.tail_node(), 0 }; return GraphBuilder::add_roi_align_node(s.graph(), common_params, input, rois, _pool_info); } private: SubStream _ss_input; SubStream _ss_rois; ROIPoolingLayerInfo _pool_info; }; /** Scale Layer */ class ScaleLayer final : public ILayer { public: /** Construct a scale layer. * * @param[in] mul_w Accessor to get mul weight from. * @param[in] add_w Accessor to get add weight from. */ ScaleLayer(ITensorAccessorUPtr mul_w, ITensorAccessorUPtr add_w) : _mul_w(std::move(mul_w)), _add_w(std::move(add_w)) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_scale_layer(s.graph(), common_params, input, std::move(_mul_w), std::move(_add_w)); } private: ITensorAccessorUPtr _mul_w; ITensorAccessorUPtr _add_w; }; /** Slice Layer */ class SliceLayer final : public ILayer { public: /** Construct a slice layer. * * @param[in] starts The starts of the dimensions of the input tensor to be sliced. The length must be of rank(input). * @param[in] ends The ends of the dimensions of the input tensor to be sliced. The length must be of rank(input). */ SliceLayer(Coordinates &starts, Coordinates &ends) : _starts(starts), _ends(ends) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_slice_node(s.graph(), common_params, input, _starts, _ends); } private: Coordinates _starts; Coordinates _ends; }; /** Softmax Layer */ class SoftmaxLayer final : public ILayer { public: /** Construct a softmax layer. * * @param[in] beta (Optional) Beta value. Default 1.0. */ SoftmaxLayer(float beta = 1.0f) : _beta(beta) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_softmax_node(s.graph(), common_params, input, _beta); } private: float _beta; }; /** Stack Layer */ class StackLayer final : public ILayer { public: /** Construct a concatenation layer * * @param[in] sub_stream1 First graph branch * @param[in] sub_stream2 Second graph branch * @param[in] rest_sub_streams Rest sub-graph branches */ template StackLayer(SubStream &&sub_stream1, SubStream &&sub_stream2, Ts &&... rest_sub_streams) : _sub_streams(), _axis(0) { _sub_streams.push_back(std::make_unique(std::move(sub_stream1))); _sub_streams.push_back(std::make_unique(std::move(sub_stream2))); utility::for_each([&](SubStream && sub_stream) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); }, std::move(rest_sub_streams)...); } /** Construct a concatenation layer * * @param[in] axis Stack layer axis along which to stack the inputs * @param[in] sub_stream1 First graph branch * @param[in] sub_stream2 Second graph branch * @param[in] rest_sub_streams Rest sub-graph branches */ template StackLayer(int axis, SubStream &&sub_stream1, SubStream &&sub_stream2, Ts &&... rest_sub_streams) : _sub_streams(), _axis(axis) { _sub_streams.push_back(std::make_unique(std::move(sub_stream1))); _sub_streams.push_back(std::make_unique(std::move(sub_stream2))); utility::for_each([&](SubStream && sub_stream) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); }, std::move(rest_sub_streams)...); } /** Construct a concat layer * * @param[in] sub_stream Sub-stream */ template StackLayer(SubStream &&sub_stream) : _sub_streams(), _axis(0) { _sub_streams.push_back(std::make_unique(std::move(sub_stream))); } NodeID create_layer(IStream &s) override { NodeID nid = EmptyNodeID; NodeParams common_params = { name(), s.hints().target_hint }; if(_sub_streams.size() == 1 && _sub_streams.at(0) != nullptr) { nid = _sub_streams[0]->tail_node(); } else { // Collect tail nodes and stack std::vector nodes; for(auto &ss : _sub_streams) { if(ss && (ss->tail_node() != EmptyNodeID)) { const auto tail_node = s.graph().node(ss->tail_node()); if(tail_node != nullptr && tail_node->type() != NodeType::Output) { nodes.push_back({ ss->tail_node(), 0 }); } } } nid = GraphBuilder::add_stack_node(s.graph(), common_params, nodes, _axis); } return nid; } private: std::vector> _sub_streams; int _axis; }; /** StridedSlice Layer */ class StridedSliceLayer final : public ILayer { public: /** Construct a strided slice layer. * * @param[in] starts The starts of the dimensions of the input tensor to be sliced. The length must be of rank(input). * @param[in] ends The ends of the dimensions of the input tensor to be sliced. The length must be of rank(input). * @param[in] strides The strides of the dimensions of the input tensor to be sliced. The length must be of rank(input). * @param[in] strided_slice_info Contains masks for the starts, ends and strides */ StridedSliceLayer(Coordinates &starts, Coordinates &ends, BiStrides &strides, StridedSliceLayerInfo strided_slice_info) : _starts(starts), _ends(ends), _strides(strides), _info(strided_slice_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_strided_slice_node(s.graph(), common_params, input, _starts, _ends, _strides, _info); } private: Coordinates _starts; Coordinates _ends; BiStrides _strides; StridedSliceLayerInfo _info; }; /** YOLO Layer */ class YOLOLayer final : public ILayer { public: /** Construct a YOLO layer. * * @param[in] act_info Activation info */ YOLOLayer(ActivationLayerInfo act_info) : _act_info(act_info) { } NodeID create_layer(IStream &s) override { NodeParams common_params = { name(), s.hints().target_hint }; NodeIdxPair input = { s.tail_node(), 0 }; return GraphBuilder::add_yolo_node(s.graph(), common_params, input, _act_info); } private: ActivationLayerInfo _act_info; }; } // namespace frontend } // namespace graph } // namespace arm_compute #endif /* ARM_COMPUTE_GRAPH_LAYERS_H */