shuffle_channels.hpp

// Copyright (C) 2018-2021 Intel Corporation
// SPDX-License-Identifier: Apache-2.0
//
 
#pragma once
 
#include <memory>
 
#include "ngraph/node.hpp"
#include "ngraph/op/util/fused_op.hpp"
 
namespace ngraph
{
    namespace op
    {
        namespace v0
        {
            /// \brief Permutes data in the channel dimension of the input
            class NGRAPH_API ShuffleChannels : public Op
            {
            public:
                NGRAPH_RTTI_DECLARATION;
 
                ShuffleChannels() = default;
                /// \brief Constructs a ShuffleChannels node.
                ///
                /// \param data - Node producing the input tensor
                /// \param axis - channel dimension index in the data tensor. A negative value means
                ///               that the index should be calculated from the back of the input
                ///               data
                ///               shape.
                /// \param group - number of group the channel dimension specified by axis should
                /// be
                ///                 split into
                ShuffleChannels(const Output<Node>& data,
                                const int64_t axis = 1,
                                const int64_t group = 1);
 
                bool visit_attributes(AttributeVisitor& visitor) override;
                size_t get_zero_based_axis() const;
 
                virtual void validate_and_infer_types() override;
 
                virtual std::shared_ptr<Node>
                    clone_with_new_inputs(const OutputVector& new_args) const override;
 
                int64_t get_axis() const { return m_axis; }
                int64_t get_group() const { return m_group; }
                bool evaluate(const HostTensorVector& outputs,
                              const HostTensorVector& inputs) const override;
                bool has_evaluate() const override;
 
            private:
                /// \brief Generates a shape required to permute the data
                ///
                /// \param data_shape - Shape of the original input data tensor
                /// \return A 4D tensor to be used to reshape the input data before shuffling it
                Shape get_pre_shuffle_shape(const Shape& data_shape) const;
                bool evaluate_shuffle_channels(const HostTensorVector& outputs,
                                               const HostTensorVector& inputs) const;
 
                int64_t m_axis;
                int64_t m_group;
            };
        } // namespace v0
        using v0::ShuffleChannels;
    } // namespace op
} // namespace ngraph