| Teresa Charlin | bd22c7d | 2022-04-26 18:14:12 +0100 | [diff] [blame] | 1 | // |
| 2 | // Copyright © 2022 Arm Ltd and Contributors. All rights reserved. |
| 3 | // SPDX-License-Identifier: MIT |
| 4 | // |
| 5 | |
| 6 | #pragma once |
| 7 | |
| 8 | #include "NeonBaseWorkload.hpp" |
| 9 | |
| 10 | #include "arm_compute/runtime/Tensor.h" |
| 11 | #include "arm_compute/runtime/NEON/functions/NEGather.h" |
| 12 | #include "arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h" |
| 13 | #include "arm_compute/runtime/NEON/functions/NEReductionOperation.h" |
| 14 | #include "arm_compute/runtime/NEON/functions/NEReshapeLayer.h" |
| 15 | |
| 16 | namespace armnn |
| 17 | { |
| 18 | arm_compute::Status NeonGatherNdWorkloadValidate(const TensorInfo& input, |
| 19 | const TensorInfo& indices, |
| 20 | const TensorInfo& output); |
| 21 | |
| 22 | class NeonGatherNdWorkload : public NeonBaseWorkload<GatherNdQueueDescriptor> |
| 23 | { |
| 24 | public: |
| 25 | NeonGatherNdWorkload(const GatherNdQueueDescriptor& descriptor, const WorkloadInfo& info); |
| 26 | virtual void Execute() const override; |
| 27 | |
| 28 | private: |
| 29 | arm_compute::Tensor m_FlattenedCoeff; |
| 30 | arm_compute::Tensor m_outputMul; |
| 31 | arm_compute::Tensor m_FlattenedIndices; |
| 32 | arm_compute::Tensor m_outputGather; |
| 33 | |
| 34 | mutable arm_compute::NEPixelWiseMultiplication m_MulLayer; |
| 35 | mutable arm_compute::NEReductionOperation m_ReduceSumLayer; |
| 36 | mutable arm_compute::NEGather m_GatherLayer; |
| 37 | mutable arm_compute::NEReshapeLayer m_ReshapeLayer; |
| 38 | |
| 39 | }; |
| 40 | |
| 41 | } //namespace armnn |