blob: 77ec1edd820058f525445893f16de8a961571137 (
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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
|
/*
* Copyright (c) 2019 Samsung Electronics Co., Ltd. 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 "UnpackCommand.hpp"
#include "Support.hpp"
#include <tensorflow/core/framework/graph.pb.h>
#include <google/protobuf/io/coded_stream.h>
#include <google/protobuf/io/zero_copy_stream_impl.h>
#include <google/protobuf/text_format.h>
#include <cassert>
#include <stdexcept>
namespace
{
template <typename T> void unpack(tensorflow::TensorProto *);
template <> void unpack<float>(tensorflow::TensorProto *input_tensor)
{
const auto &input_shape = input_tensor->tensor_shape();
assert(input_shape.dim_size() <= 6);
int input_flat_size = tfkit::tf::GetElementCount(input_shape);
// Adjust where shape is not set but actual value exist
if (input_tensor->tensor_content().size() > 0 && input_flat_size == -1)
{
input_flat_size = input_tensor->tensor_content().size() / sizeof(float);
}
if (input_tensor->tensor_content().size() == 0)
{
// Do nothing as there is no tensor content to unpack
}
else if (input_tensor->tensor_content().size() == input_flat_size * sizeof(float))
{
// clang-format off
// TODO fix indentation
input_tensor->clear_float_val();
const float *tensor_content =
reinterpret_cast<const float *>(input_tensor->tensor_content().data());
for (int i = 0; i < input_flat_size; i++)
{
input_tensor->add_float_val(tensor_content[i]);
}
input_tensor->clear_tensor_content();
// clang-format on
}
else
{
throw std::runtime_error{"Number of elements mismatch in unpack<float>."};
// TODO: support for these
}
}
template <> void unpack<int32_t>(tensorflow::TensorProto *input_tensor)
{
const auto &input_shape = input_tensor->tensor_shape();
assert(input_shape.dim_size() <= 6);
int input_flat_size = tfkit::tf::GetElementCount(input_shape);
// Adjust where shape is not set but actual value exist
if (input_tensor->tensor_content().size() > 0 && input_flat_size == -1)
{
input_flat_size = input_tensor->tensor_content().size() / sizeof(int32_t);
}
if (input_tensor->tensor_content().size() == 0)
{
// Do nothing as there is no tensor content to unpack
}
else if (input_tensor->tensor_content().size() == input_flat_size * sizeof(int32_t))
{
input_tensor->clear_int_val();
const int32_t *tensor_content =
reinterpret_cast<const int32_t *>(input_tensor->tensor_content().data());
for (int i = 0; i < input_flat_size; i++)
{
input_tensor->add_int_val(tensor_content[i]);
}
input_tensor->clear_tensor_content();
}
else
{
throw std::runtime_error{"Number of elements mismatch in unpack<int32_t>."};
// TODO: support for these
}
}
template <> void unpack<int8_t>(tensorflow::TensorProto *input_tensor)
{
const auto &input_shape = input_tensor->tensor_shape();
assert(input_shape.dim_size() <= 6);
int input_flat_size = tfkit::tf::GetElementCount(input_shape);
// Adjust where shape is not set but actual value exist
if (input_tensor->tensor_content().size() > 0 && input_flat_size == -1)
{
input_flat_size = input_tensor->tensor_content().size() / sizeof(int8_t);
}
if (input_tensor->tensor_content().size() == 0)
{
// Do nothing as there is no tensor content to unpack
}
else if (input_tensor->tensor_content().size() == input_flat_size * sizeof(int8_t))
{
input_tensor->clear_int_val();
const int8_t *tensor_content =
reinterpret_cast<const int8_t *>(input_tensor->tensor_content().data());
for (int i = 0; i < input_flat_size; i++)
{
input_tensor->add_int_val(tensor_content[i]);
}
input_tensor->clear_tensor_content();
}
else
{
throw std::runtime_error{"Number of elements mismatch in unpack<int8_t>."};
// TODO: support for these
}
}
void unpack(tensorflow::GraphDef &graph_def)
{
auto nodes = graph_def.mutable_node();
for (int i = 0; i < nodes->size(); ++i)
{
tensorflow::NodeDef *n = nodes->Mutable(i);
// TODO: handle for other operators
if (n->op() == "Const")
{
const auto dtype = tfkit::tf::GetDataTypeAttr(*n, "dtype");
tensorflow::TensorProto *tensor = tfkit::tf::GetTensorAttr(*n, "value");
switch (dtype)
{
case tensorflow::DT_FLOAT:
unpack<float>(tensor);
break;
case tensorflow::DT_INT32:
unpack<int32_t>(tensor);
break;
case tensorflow::DT_INT8:
unpack<int8_t>(tensor);
break;
default:
throw std::runtime_error{"Unsupported dtype"};
}
}
}
}
} // namespace
namespace tfkit
{
int UnpackCommand::run(int argc, const char *const *argv) const
{
tensorflow::GraphDef graph_def;
CmdArguments cmdargs(argc, argv);
auto ioconfig = make_ioconfig(cmdargs);
google::protobuf::io::IstreamInputStream is{ioconfig->in()};
if (!google::protobuf::TextFormat::Parse(&is, &graph_def))
{
std::cerr << "ERROR: Failed to parse prototxt" << std::endl;
return 255;
}
// convert tensor_content to float_val
unpack(graph_def);
google::protobuf::io::OstreamOutputStream os{ioconfig->out()};
google::protobuf::TextFormat::Print(graph_def, &os);
return 0;
}
} // namespace tfkit
|
