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
|
/* Copyright 2016 Google Inc. 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 "tensorflow/core/example/example_parser_configuration.h"
#include <vector>
#include "tensorflow/core/example/example.pb.h"
#include "tensorflow/core/example/feature.pb_text.h"
#include "tensorflow/core/framework/numeric_op.h"
#include "tensorflow/core/framework/register_types.h"
#include "tensorflow/core/lib/core/errors.h"
#include "tensorflow/core/lib/strings/strcat.h"
#include "tensorflow/core/platform/logging.h"
#include "tensorflow/core/platform/protobuf.h"
namespace tensorflow {
Status FindNodeIndexByName(const tensorflow::GraphDef& graph,
const string& node_name, int* node_idx) {
for (int i = 0; i < graph.node_size(); ++i) {
const auto& node = graph.node(i);
if (node.name() == node_name) {
*node_idx = i;
return Status::OK();
}
}
return errors::InvalidArgument(node_name, " not found in GraphDef");
}
Status ExtractExampleParserConfiguration(
const tensorflow::GraphDef& graph, const string& node_name,
tensorflow::Session* session,
std::vector<FixedLenFeature>* fixed_len_features,
std::vector<VarLenFeature>* var_len_features) {
int node_idx;
TF_RETURN_IF_ERROR(FindNodeIndexByName(graph, node_name, &node_idx));
const auto& node = graph.node(node_idx);
if (node.op() != "ParseExample") {
return errors::InvalidArgument(node_name, " node is not a ParseExample op");
}
auto& attr_map = node.attr();
auto num_sparse = attr_map.at("Nsparse").i();
auto num_dense = attr_map.at("Ndense").i();
fixed_len_features->resize(num_dense);
var_len_features->resize(num_sparse);
auto tdense = attr_map.at("Tdense");
auto dense_shapes = attr_map.at("dense_shapes");
auto sparse_types = attr_map.at("sparse_types");
// Consistency check attributes.
if (tdense.list().type_size() != num_dense) {
return errors::InvalidArgument("Node attr Tdense has ",
tdense.list().type_size(),
" elements != Ndense attr: ", num_dense);
}
if (dense_shapes.list().shape_size() != num_dense) {
return errors::InvalidArgument("Node attr dense_shapes has ",
dense_shapes.list().shape_size(),
" elements != Ndense attr: ", num_dense);
}
if (sparse_types.list().type_size() != num_sparse) {
return errors::InvalidArgument("Node attr sparse_types has ",
sparse_types.list().type_size(),
" elements != NSparse attr: ", num_sparse);
}
for (int i = 0; i < tdense.list().type_size(); ++i) {
(*fixed_len_features)[i].dtype = tdense.list().type(i);
// Convert TensorShapeProto to TensorShape.
(*fixed_len_features)[i].shape = TensorShape(dense_shapes.list().shape(i));
}
for (int i = 0; i < sparse_types.list().type_size(); ++i) {
(*var_len_features)[i].dtype = sparse_types.list().type(i);
}
// We must fetch the configuration input tensors to the ParseExample op.
// Skipping index = 0, which is the serialized proto input.
std::vector<string> fetch_names(node.input_size() - 1);
for (int i = 1; i < node.input_size(); ++i) {
fetch_names[i - 1] = node.input(i);
}
std::vector<Tensor> op_input_tensors;
TF_RETURN_IF_ERROR(session->Run({}, // no_inputs,
fetch_names, {}, // no target_node_names,
&op_input_tensors));
// The input tensors are laid out sequentially in a flat manner.
// Here are the various start offsets.
int sparse_keys_start = 1;
int dense_keys_start = sparse_keys_start + num_sparse;
int dense_defaults_start = dense_keys_start + num_dense;
for (int i = 0; i < num_sparse; ++i) {
int input_idx = sparse_keys_start + i;
(*var_len_features)[i].key = op_input_tensors[input_idx].scalar<string>()();
}
for (int i = 0; i < num_dense; ++i) {
FixedLenFeature& config = (*fixed_len_features)[i];
int dense_keys_offset = dense_keys_start + i;
config.key = op_input_tensors[dense_keys_offset].scalar<string>()();
int defaults_offset = dense_defaults_start + i;
config.default_value = op_input_tensors[defaults_offset];
}
// The output tensors are laid out sequentially in a flat manner.
// Here are the various start offsets.
int sparse_indices_output_start = 0;
int sparse_values_output_start = sparse_indices_output_start + num_sparse;
int sparse_shapes_output_start = sparse_values_output_start + num_sparse;
int dense_values_output_start = sparse_shapes_output_start + num_sparse;
string node_output_prefix = strings::StrCat(node_name, ":");
for (int i = 0; i < num_sparse; ++i) {
VarLenFeature& config = (*var_len_features)[i];
int indices_offset = sparse_indices_output_start + i;
config.indices_output_tensor_name =
strings::StrCat(node_output_prefix, indices_offset);
int values_offset = sparse_values_output_start + i;
config.values_output_tensor_name =
strings::StrCat(node_output_prefix, values_offset);
int shapes_offset = sparse_shapes_output_start + i;
config.shapes_output_tensor_name =
strings::StrCat(node_output_prefix, shapes_offset);
}
for (int i = 0; i < num_dense; ++i) {
int output_idx = dense_values_output_start + i;
(*fixed_len_features)[i].values_output_tensor_name =
strings::StrCat(node_output_prefix, output_idx);
}
return Status::OK();
}
} // namespace tensorflow
|
