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
|
/*
* Copyright (c) 2023 Samsung Electronics Co., Ltd. All Rights Reserved
* Copyright 2018 The TensorFlow Authors. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "kernels/FloorMod.h"
#include "kernels/Utils.h"
#include <tensorflow/lite/kernels/internal/reference/binary_function.h>
#include <cmath>
namespace
{
template <typename T> T FloorDivFunc(T input1, T input2)
{
struct FloatMod
{
float operator()(const float lhs, const float rhs) const { return std::fmod(lhs, rhs); }
};
using ModFunc =
typename std::conditional<std::is_integral<T>::value, std::modulus<T>, FloatMod>::type;
ModFunc mod_func;
T trunc_mod = mod_func(input1, input2);
return (trunc_mod != 0) && ((input2 < 0) != (trunc_mod < 0)) ? (trunc_mod + input2) : trunc_mod;
}
} // namespace
namespace luci_interpreter
{
namespace kernels
{
FloorMod::FloorMod(const Tensor *x, const Tensor *y, Tensor *output) : Kernel({x, y}, {output}) {}
void FloorMod::configure()
{
LUCI_INTERPRETER_CHECK(x()->element_type() == output()->element_type());
LUCI_INTERPRETER_CHECK(y()->element_type() == output()->element_type());
output()->resize(calculateShapeForBroadcast(x()->shape(), y()->shape()));
}
void FloorMod::execute() const
{
switch (x()->element_type())
{
case DataType::FLOAT32:
evalFloat();
break;
case DataType::S8:
evalInteger<int8_t>();
break;
case DataType::S16:
evalInteger<int16_t>();
break;
case DataType::S32:
evalInteger<int32_t>();
break;
case DataType::S64:
evalInteger<int64_t>();
break;
default:
throw std::runtime_error("Unsupported type.");
}
}
void FloorMod::evalFloat() const
{
const auto x_data = getTensorData<float>(x());
const auto y_data = getTensorData<float>(y());
if (x()->shape() != y()->shape())
{
tflite::reference_ops::BroadcastBinaryFunction4DSlow<float, float, float>(
getTensorShape(x()), x_data, getTensorShape(y()), y_data, getTensorShape(output()),
getTensorData<float>(output()), FloorDivFunc<float>);
}
else
{
tflite::reference_ops::BinaryFunction<float, float, float>(
getTensorShape(x()), x_data, getTensorShape(y()), y_data, getTensorShape(output()),
getTensorData<float>(output()), FloorDivFunc<float>);
}
}
template <typename T> void FloorMod::evalInteger() const
{
const auto x_data = getTensorData<T>(x());
const auto y_data = getTensorData<T>(y());
// Check the denominator
const auto y_data_type = y()->element_type();
if (y_data_type == DataType::S8 || y_data_type == DataType::S16 || y_data_type == DataType::S32 ||
y_data_type == DataType::S64)
{
for (int i = 0; i < getTensorShape(y()).FlatSize(); ++i)
{
LUCI_INTERPRETER_CHECK(y_data[i] != 0);
}
}
if (x()->shape() != y()->shape())
{
tflite::reference_ops::BroadcastBinaryFunction4DSlow<T, T, T>(
getTensorShape(x()), x_data, getTensorShape(y()), y_data, getTensorShape(output()),
getTensorData<T>(output()), FloorDivFunc<T>);
}
else
{
tflite::reference_ops::BinaryFunction<T, T, T>(getTensorShape(x()), x_data, getTensorShape(y()),
y_data, getTensorShape(output()),
getTensorData<T>(output()), FloorDivFunc<T>);
}
}
} // namespace kernels
} // namespace luci_interpreter
|
