1 /*
2 * Copyright (c) 2019-2020 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 #include "DepthToSpaceLayer.h"
25
26 #include "tests/validation/Helpers.h"
27
28 namespace arm_compute
29 {
30 namespace test
31 {
32 namespace validation
33 {
34 namespace reference
35 {
36 // Batch to Space
37 template <typename T>
depth_to_space(const SimpleTensor<T> & src,const TensorShape & dst_shape,int32_t block_shape)38 SimpleTensor<T> depth_to_space(const SimpleTensor<T> &src, const TensorShape &dst_shape, int32_t block_shape)
39 {
40 ARM_COMPUTE_ERROR_ON(block_shape <= 0);
41 SimpleTensor<T> result(dst_shape, src.data_type());
42
43 const auto width_in = static_cast<int>(src.shape()[0]);
44 const auto height_in = static_cast<int>(src.shape()[1]);
45 const auto channel_in = static_cast<int>(src.shape()[2]);
46 const auto batch_in = static_cast<int>(src.shape()[3]);
47 const int r = channel_in / (block_shape * block_shape);
48 #if defined(_OPENMP)
49 #pragma omp parallel for collapse(4)
50 #endif /* _OPENMP */
51 for(int b = 0; b < batch_in; ++b)
52 {
53 for(int z = 0; z < channel_in; ++z)
54 {
55 for(int y = 0; y < height_in; ++y)
56 {
57 for(int x = 0; x < width_in; ++x)
58 {
59 const int out_x = (block_shape * x + (z / r) % block_shape);
60 const int out_y = (block_shape * y + (z / r) / block_shape);
61 const int out_pos = out_x + dst_shape[0] * out_y + (z % r) * dst_shape[0] * dst_shape[1] + b * dst_shape[0] * dst_shape[1] * dst_shape[2];
62 const int in_pos = x + width_in * y + z * width_in * height_in + b * width_in * height_in * channel_in;
63 result[out_pos] = src[in_pos];
64 }
65 }
66 }
67 }
68
69 return result;
70 }
71 template SimpleTensor<float> depth_to_space(const SimpleTensor<float> &src, const TensorShape &dst_shape, int32_t block_shape);
72 template SimpleTensor<half> depth_to_space(const SimpleTensor<half> &src, const TensorShape &dst_shape, int32_t block_shape);
73 } // namespace reference
74 } // namespace validation
75 } // namespace test
76 } // namespace arm_compute
77