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
203
204
205
206
|
#include <iostream>
#include <sstream>
#include <string>
/**
* The tracer.cpp generates a binary that accepts multiple Torch Mobile Model(s)
* (with bytecode.pkl), each of which has at least 1 bundled
* input. This binary then feeds the bundled input(s) into each corresponding
* model and executes it using the lite interpreter.
*
* Both root operators as well as called operators are recorded and saved
* into a YAML file (whose path is provided on the command line).
*
* Note: Root operators may include primary and other operators that
* are not invoked using the dispatcher, and hence they may not show
* up in the Traced Operator list.
*
*/
#include <ATen/core/dispatch/ObservedOperators.h>
#include <torch/csrc/autograd/grad_mode.h>
#include <torch/csrc/jit/mobile/import.h>
#include <torch/csrc/jit/mobile/model_tracer/KernelDTypeTracer.h>
#include <torch/csrc/jit/mobile/model_tracer/MobileModelRunner.h>
#include <torch/csrc/jit/mobile/model_tracer/OperatorCallTracer.h>
#include <torch/csrc/jit/mobile/model_tracer/TensorUtils.h>
#include <torch/csrc/jit/mobile/model_tracer/TracerRunner.h>
#include <torch/csrc/jit/mobile/module.h>
#include <torch/csrc/jit/mobile/parse_operators.h>
#include <torch/script.h>
typedef std::map<std::string, std::set<std::string>> kt_type;
C10_DEFINE_string(
model_input_path,
"",
"A comma separated list of path(s) to the input model file(s) (.ptl).");
C10_DEFINE_string(
build_yaml_path,
"",
"The path of the output YAML file containing traced operator information.");
#define REQUIRE_STRING_ARG(name) \
if (FLAGS_##name.empty()) { \
std::cerr << "You must specify the flag --" #name "\n"; \
return 1; \
}
#define REQUIRE_INT_ARG(name) \
if (FLAGS_##name == -1) { \
std::cerr << "You must specify the flag --" #name "\n"; \
return 1; \
}
void printOpYAML(
std::ostream& out,
int indent,
const std::string& op_name,
bool is_used_for_training,
bool is_root_operator,
bool include_all_overloads) {
out << std::string(indent, ' ') << op_name << ":" << std::endl;
out << std::string(indent + 2, ' ')
<< "is_used_for_training: " << (is_used_for_training ? "true" : "false")
<< std::endl;
out << std::string(indent + 2, ' ')
<< "is_root_operator: " << (is_root_operator ? "true" : "false")
<< std::endl;
out << std::string(indent + 2, ' ')
<< "include_all_overloads: " << (include_all_overloads ? "true" : "false")
<< std::endl;
}
void printOpsYAML(
std::ostream& out,
const std::set<std::string>& operator_list,
bool is_used_for_training,
bool is_root_operator,
bool include_all_overloads) {
for (auto& it : operator_list) {
printOpYAML(out, 2, it, false, is_root_operator, false);
}
}
void printDTypeYAML(
std::ostream& out,
int indent,
const std::string& kernel_tag_name,
const std::set<std::string> dtypes) {
std::string indent_str = std::string(indent, ' ');
out << indent_str << kernel_tag_name << ":" << std::endl;
for (auto& dtype : dtypes) {
out << indent_str << "- " << dtype << std::endl;
}
}
void printDTypesYAML(
std::ostream& out,
const torch::jit::mobile::KernelDTypeTracer::kernel_tags_type&
kernel_tags) {
for (auto& it : kernel_tags) {
printDTypeYAML(out, 2, it.first, it.second);
}
}
void printCustomClassesYAML(
std::ostream& out,
const torch::jit::mobile::CustomClassTracer::custom_classes_type&
loaded_classes) {
for (auto& class_name : loaded_classes) {
out << "- " << class_name << std::endl;
}
}
/**
* Runs multiple PyTorch lite interpreter models, and additionally writes
* out a list of root and called operators, kernel dtypes, and loaded/used
* TorchBind custom classes.
*/
int main(int argc, char* argv[]) {
if (!c10::ParseCommandLineFlags(&argc, &argv)) {
std::cerr << "Failed to parse command line flags!" << std::endl;
return 1;
}
REQUIRE_STRING_ARG(model_input_path);
REQUIRE_STRING_ARG(build_yaml_path);
std::istringstream sin(FLAGS_model_input_path);
std::ofstream yaml_out(FLAGS_build_yaml_path);
std::cout << "Output: " << FLAGS_build_yaml_path << std::endl;
torch::jit::mobile::TracerResult tracer_result;
std::vector<std::string> model_input_paths;
for (std::string model_input_path;
std::getline(sin, model_input_path, ',');) {
std::cout << "Processing: " << model_input_path << std::endl;
model_input_paths.push_back(model_input_path);
}
try {
tracer_result = torch::jit::mobile::trace_run(model_input_paths);
} catch (std::exception& ex) {
std::cerr
<< "ModelTracer has not been able to load the module for the following reasons:\n"
<< ex.what()
<< "\nPlease consider opening an issue at https://github.com/pytorch/pytorch/issues "
<< "with the detailed error message." << std::endl;
throw ex;
}
if (tracer_result.traced_operators.size() <=
torch::jit::mobile::always_included_traced_ops.size()) {
std::cerr
<< c10::str(
"Error traced_operators size: ",
tracer_result.traced_operators.size(),
". Expected the traced operator list to be bigger then the default size ",
torch::jit::mobile::always_included_traced_ops.size(),
". Please report a bug in PyTorch.")
<< std::endl;
}
// If the op exist in both traced_ops and root_ops, leave it in root_ops only
for (const auto& root_op : tracer_result.root_ops) {
if (tracer_result.traced_operators.find(root_op) !=
tracer_result.traced_operators.end()) {
tracer_result.traced_operators.erase(root_op);
}
}
yaml_out << "include_all_non_op_selectives: false" << std::endl;
yaml_out << "build_features: []" << std::endl;
yaml_out << "operators:" << std::endl;
printOpsYAML(
yaml_out,
tracer_result.root_ops,
false /* is_used_for_training */,
true /* is_root_operator */,
false /* include_all_overloads */);
printOpsYAML(
yaml_out,
tracer_result.traced_operators,
false /* is_used_for_training */,
false /* is_root_operator */,
false /* include_all_overloads */);
yaml_out << "kernel_metadata:";
if (tracer_result.called_kernel_tags.empty()) {
yaml_out << " []";
}
yaml_out << std::endl;
printDTypesYAML(yaml_out, tracer_result.called_kernel_tags);
yaml_out << "custom_classes:";
if (tracer_result.loaded_classes.empty()) {
yaml_out << " []";
}
yaml_out << std::endl;
printCustomClassesYAML(yaml_out, tracer_result.loaded_classes);
return 0;
}
|
