Compute Library
 21.02
NEChannelShuffleLayer.h
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2018-2020 Arm Limited.
3  *
4  * SPDX-License-Identifier: MIT
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to
8  * deal in the Software without restriction, including without limitation the
9  * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10  * sell copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in all
14  * copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
24 #ifndef ARM_COMPUTE_NECHANNELSHUFFLELAYER_H
25 #define ARM_COMPUTE_NECHANNELSHUFFLELAYER_H
26 
27 #include "arm_compute/core/Types.h"
29 
30 namespace arm_compute
31 {
32 // Forward declarations
33 class ITensor;
34 class ITensorInfo;
35 
36 /** Basic function to run @ref NEChannelShuffleLayerKernel
37  *
38  * @note The function performs a channel shuffle operation on the input tensor. Given NCHW tensor with group G, it will
39  * first divide the channels into G groups, C = (G * C'), and perform a transpose of the channel, which gives C = (C' * G).
40  * for more details see: https://arxiv.org/pdf/1707.01083.pdf
41  */
43 {
44 public:
45  /** Initialize the function
46  *
47  * @param[in] input Input tensor. Data types supported: All
48  * @param[out] output Output tensor. Data type supported: Same as @p input
49  * @param[in] num_groups Number of groups. Must be greater than 1 and the number of channels of the tensors must be a multiple of the number of groups.
50  */
51  void configure(const ITensor *input, ITensor *output, unsigned int num_groups);
52  /** Static function to check if given info will lead to a valid configuration of @ref NEChannelShuffleLayer
53  *
54  * @param[in] input Input tensor. Data types supported: All
55  * @param[out] output Output tensor. Data type supported: Same as @p input
56  * @param[in] num_groups Number of groups. Must be greater than 1 and the number of channels of the tensors must be a multiple of the number of groups.
57  *
58  * @return a status
59  */
60  static Status validate(const ITensorInfo *input, const ITensorInfo *output, unsigned int num_groups);
61 };
62 } // namespace arm_compute
63 #endif /* ARM_COMPUTE_NECHANNELSHUFFLELAYER_H */
Basic function to run NEChannelShuffleLayerKernel.
Store the tensor's metadata.
Definition: ITensorInfo.h:40
void configure(const ITensor *input, ITensor *output, unsigned int num_groups)
Initialize the function.
Status class.
Definition: Error.h:52
Interface for Neon tensor.
Definition: ITensor.h:36
Copyright (c) 2017-2021 Arm Limited.
Basic interface for functions which have a single Neon kernel and no border.
static Status validate(const ITensorInfo *input, const ITensorInfo *output, unsigned int num_groups)
Static function to check if given info will lead to a valid configuration of NEChannelShuffleLayer.
const unsigned int num_groups
Definition: Im2Col.cpp:153