1 /*
2 * Copyright (c) 2017-2018 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 "NormalizationLayer.h"
25
26 #include "arm_compute/core/Types.h"
27
28 namespace arm_compute
29 {
30 namespace test
31 {
32 namespace validation
33 {
34 namespace reference
35 {
36 template <typename T>
normalization_layer(const SimpleTensor<T> & src,NormalizationLayerInfo info)37 SimpleTensor<T> normalization_layer(const SimpleTensor<T> &src, NormalizationLayerInfo info)
38 {
39 // Create reference
40 SimpleTensor<T> dst{ src.shape(), src.data_type(), 1 };
41
42 // Compute reference
43 const uint32_t norm_size = info.norm_size();
44 NormType type = info.type();
45 float beta = info.beta();
46 uint32_t kappa = info.kappa();
47
48 const int cols = src.shape()[0];
49 const int rows = src.shape()[1];
50 const int depth = src.shape()[2];
51 int upper_dims = src.shape().total_size() / (cols * rows);
52
53 float coeff = info.scale_coeff();
54 int radius_cols = norm_size / 2;
55
56 // IN_MAP_1D and CROSS_MAP normalize over a single axis only
57 int radius_rows = (NormType::IN_MAP_2D == type) ? norm_size / 2 : 0;
58
59 if(info.is_cross_map())
60 {
61 // Remove also depth from upper dimensions since it is the dimension we
62 // want to use for normalization
63 upper_dims /= depth;
64
65 for(int r = 0; r < upper_dims; ++r)
66 {
67 for(int i = 0; i < rows; ++i)
68 {
69 for(int k = 0; k < cols; ++k)
70 {
71 for(int l = 0; l < depth; ++l)
72 {
73 float accumulated_scale = 0.f;
74
75 for(int j = -radius_cols; j <= radius_cols; ++j)
76 {
77 const int z = l + j;
78
79 if(z >= 0 && z < depth)
80 {
81 const T value = src[k + i * cols + z * rows * cols + r * cols * rows * depth];
82 accumulated_scale += value * value;
83 }
84 }
85
86 dst[k + i * cols + l * rows * cols + r * cols * rows * depth] = kappa + accumulated_scale * coeff;
87 }
88 }
89 }
90 }
91 }
92 else
93 {
94 for(int r = 0; r < upper_dims; ++r)
95 {
96 for(int i = 0; i < rows; ++i)
97 {
98 for(int k = 0; k < cols; ++k)
99 {
100 float accumulated_scale = 0.f;
101
102 for(int j = -radius_rows; j <= radius_rows; ++j)
103 {
104 const int y = i + j;
105 for(int l = -radius_cols; l <= radius_cols; ++l)
106 {
107 const int x = k + l;
108
109 if((x >= 0 && y >= 0) && (x < cols && y < rows))
110 {
111 const T value = src[x + y * cols + r * cols * rows];
112 accumulated_scale += value * value;
113 }
114 }
115 }
116
117 dst[k + i * cols + r * cols * rows] = kappa + accumulated_scale * coeff;
118 }
119 }
120 }
121 }
122
123 if(beta == 1.f)
124 {
125 for(int i = 0; i < dst.num_elements(); ++i)
126 {
127 dst[i] = src[i] / dst[i];
128 }
129 }
130 else if(beta == 0.5f)
131 {
132 for(int i = 0; i < dst.num_elements(); ++i)
133 {
134 dst[i] = src[i] / std::sqrt(dst[i]);
135 }
136 }
137 else
138 {
139 for(int i = 0; i < dst.num_elements(); ++i)
140 {
141 dst[i] = src[i] * std::exp(std::log(dst[i]) * -beta);
142 }
143 }
144
145 return dst;
146 }
147
148 template SimpleTensor<float> normalization_layer(const SimpleTensor<float> &src, NormalizationLayerInfo info);
149 template SimpleTensor<half> normalization_layer(const SimpleTensor<half> &src, NormalizationLayerInfo info);
150 } // namespace reference
151 } // namespace validation
152 } // namespace test
153 } // namespace arm_compute
154