aboutsummaryrefslogtreecommitdiff
path: root/tests/validation/reference/BatchNormalizationLayer.cpp
blob: ae309d909316a769d823314693ff4090f78ece64 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
/*
 * Copyright (c) 2017-2018 ARM Limited.
 *
 * SPDX-License-Identifier: MIT
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to
 * deal in the Software without restriction, including without limitation the
 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
#include "BatchNormalizationLayer.h"

#include "ActivationLayer.h"

#include "tests/validation/FixedPoint.h"
#include "tests/validation/Helpers.h"
#include "tests/validation/reference/Permute.h"

namespace arm_compute
{
namespace test
{
namespace validation
{
namespace reference
{
// Batch Normalization Layer for fixed point type
template <typename T, typename std::enable_if<std::is_integral<T>::value, int>::type *>
SimpleTensor<T> batch_normalization_layer(const SimpleTensor<T> &src, const SimpleTensor<T> &mean, const SimpleTensor<T> &var, const SimpleTensor<T> &beta, const SimpleTensor<T> &gamma, float epsilon,
                                          ActivationLayerInfo act_info, int fixed_point_position)
{
    ARM_COMPUTE_ERROR_ON_MSG(src.data_layout() == DataLayout::NHWC, "Unsupported NHWC format");
    ARM_COMPUTE_UNUSED(act_info);
    SimpleTensor<T> result(src.shape(), src.data_type());

    const auto cols       = static_cast<int>(src.shape()[0]);
    const auto rows       = static_cast<int>(src.shape()[1]);
    const auto depth      = static_cast<int>(src.shape()[2]);
    const int  upper_dims = src.shape().total_size() / (cols * rows * depth);

    for(int r = 0; r < upper_dims; ++r)
    {
        for(int i = 0; i < depth; ++i)
        {
            for(int k = 0; k < rows; ++k)
            {
                for(int l = 0; l < cols; ++l)
                {
                    const int pos = l + k * cols + i * rows * cols + r * cols * rows * depth;

                    fixed_point_arithmetic::fixed_point<T> src_qs(src[pos], fixed_point_position, true);
                    fixed_point_arithmetic::fixed_point<T> var_qs(var[i], fixed_point_position, true);
                    fixed_point_arithmetic::fixed_point<T> mean_qs(mean[i], fixed_point_position, true);
                    fixed_point_arithmetic::fixed_point<T> beta_qs(beta[i], fixed_point_position, true);
                    fixed_point_arithmetic::fixed_point<T> gamma_qs(gamma[i], fixed_point_position, true);
                    fixed_point_arithmetic::fixed_point<T> epsilon_qs(epsilon, fixed_point_position);

                    auto denominator = fixed_point_arithmetic::inv_sqrt(var_qs + epsilon_qs);
                    auto numerator   = src_qs - mean_qs;
                    auto x_bar       = numerator * denominator;
                    x_bar            = beta_qs + x_bar * gamma_qs;
                    result[pos]      = x_bar.raw();
                }
            }
        }
    }

    return result;
}

// Batch Normalization Layer for floating point type
template <typename T, typename std::enable_if<is_floating_point<T>::value, int>::type *>
SimpleTensor<T> batch_normalization_layer(const SimpleTensor<T> &src, const SimpleTensor<T> &mean, const SimpleTensor<T> &var, const SimpleTensor<T> &beta, const SimpleTensor<T> &gamma, float epsilon,
                                          ActivationLayerInfo act_info, int fixed_point_position)
{
    ARM_COMPUTE_UNUSED(fixed_point_position);

    const bool            is_nhwc  = src.data_layout() == DataLayout::NHWC;
    const SimpleTensor<T> perm_src = (is_nhwc) ? permute(src, PermutationVector(1U, 2U, 0U)) : src;
    SimpleTensor<T>       result(perm_src.shape(), perm_src.data_type());

    const auto cols       = static_cast<int>(perm_src.shape()[0]);
    const auto rows       = static_cast<int>(perm_src.shape()[1]);
    const auto depth      = static_cast<int>(perm_src.shape()[2]);
    const int  upper_dims = perm_src.shape().total_size() / (cols * rows * depth);

    for(int r = 0; r < upper_dims; ++r)
    {
        for(int i = 0; i < depth; ++i)
        {
            for(int k = 0; k < rows; ++k)
            {
                for(int l = 0; l < cols; ++l)
                {
                    const int   pos         = l + k * cols + i * rows * cols + r * cols * rows * depth;
                    const float denominator = sqrt(var[i] + epsilon);
                    const float numerator   = perm_src[pos] - mean[i];
                    const float x_bar       = numerator / denominator;
                    result[pos]             = beta[i] + x_bar * gamma[i];
                }
            }
        }
    }

    if(act_info.enabled())
    {
        result = activation_layer(result, act_info);
    }

    if(is_nhwc)
    {
        result = permute(result, PermutationVector(2U, 0U, 1U));
    }
    return result;
}
template SimpleTensor<float> batch_normalization_layer(const SimpleTensor<float> &src, const SimpleTensor<float> &mean, const SimpleTensor<float> &var, const SimpleTensor<float> &beta,
                                                       const SimpleTensor<float> &gamma, float epsilon, ActivationLayerInfo act_info, int fixed_point_position);
template SimpleTensor<int8_t> batch_normalization_layer(const SimpleTensor<int8_t> &src, const SimpleTensor<int8_t> &mean, const SimpleTensor<int8_t> &var, const SimpleTensor<int8_t> &beta,
                                                        const SimpleTensor<int8_t> &gamma, float epsilon, ActivationLayerInfo act_info, int fixed_point_position);
template SimpleTensor<int16_t> batch_normalization_layer(const SimpleTensor<int16_t> &src, const SimpleTensor<int16_t> &mean, const SimpleTensor<int16_t> &var, const SimpleTensor<int16_t> &beta,
                                                         const SimpleTensor<int16_t> &gamma, float epsilon, ActivationLayerInfo act_info, int fixed_point_position);
template SimpleTensor<half> batch_normalization_layer(const SimpleTensor<half> &src, const SimpleTensor<half> &mean, const SimpleTensor<half> &var,
                                                      const SimpleTensor<half> &beta,
                                                      const SimpleTensor<half> &gamma, float epsilon, ActivationLayerInfo act_info, int fixed_point_position);

} // namespace reference
} // namespace validation
} // namespace test
} // namespace arm_compute