-
Notifications
You must be signed in to change notification settings - Fork 1
/
cmd_common.h
239 lines (202 loc) · 7.5 KB
/
cmd_common.h
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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
#pragma once
#include "absl/flags/flag.h"
#include "absl/flags/parse.h"
#include "absl/flags/declare.h"
#include "absl/flags/usage.h"
#include "absl/log/log.h"
#include "absl/log/initialize.h"
#include "cuda_fp16.h"
#include "nn-scaler.h"
#include "infer_engine.h"
#include "reformat/reformat.h"
#include "layers.h"
#include "image_io.h"
#include "logging_trt.h"
#ifdef _WIN32
#define NOMINMAX
#define WIN32_LEAN_AND_MEAN
#include <Windows.h>
#endif
ABSL_DECLARE_FLAG(int, stderrthreshold);
ABSL_DECLARE_FLAG(bool, log_prefix);
ABSL_FLAG(std::string, model_path, "models", "path to the folder to save model files");
InferenceSession *session = nullptr;
int using_io = 0;
pixel_importer_cpu *importer_cpu = nullptr;
pixel_exporter_cpu *exporter_cpu = nullptr;
pixel_importer_gpu<float> *importer_gpu = nullptr;
pixel_exporter_gpu<float> *exporter_gpu = nullptr;
pixel_importer_gpu<half> *importer_gpu_fp16 = nullptr;
pixel_exporter_gpu<half> *exporter_gpu_fp16 = nullptr;
int32_t h_scale, w_scale;
#if defined(__GNUC__)
extern "C" __attribute__((weak)) int32_t getInferLibVersion() noexcept {
return NV_TENSORRT_VERSION;
}
#elif defined(_MSC_VER)
extern "C" int32_t getInferLibVersion_UseHeader() noexcept {
return NV_TENSORRT_VERSION;
}
#pragma comment(linker, "/alternatename:getInferLibVersion=getInferLibVersion_UseHeader")
#endif
static Logger gLogger;
ABSL_FLAG(bool, fp16, false, "use FP16 processing, allow FP16 in engine");
ABSL_FLAG(bool, int8, false, "allow INT8 in engine");
ABSL_FLAG(bool, force_precision, false, "Force precision config in model");
ABSL_FLAG(bool, external, false, "use external algorithms from cuDNN and cuBLAS");
ABSL_FLAG(bool, low_mem, false, "tweak configs to reduce memory consumption");
ABSL_FLAG(int32_t, aux_stream, -1, "Auxiliary streams to use");
ABSL_FLAG(std::string, reformatter, "auto", "reformatter used to import and export pixels: cpu, gpu, auto");
ABSL_FLAG(uint32_t, tile_width, 512, "tile width");
ABSL_FLAG(uint32_t, tile_height, 512, "tile height");
ABSL_FLAG(uint32_t, tile_pad, 16, "tile pad border to reduce tile block discontinuity");
ABSL_FLAG(uint32_t, extend_grace, 0, "grace limit to not split another tile");
ABSL_FLAG(uint32_t, alignment, 1, "model input alignment requirement");
ABSL_FLAG(bool, cuda_lazy_load, true, "enable CUDA lazying load.");
void setup_session(bool handle_alpha) {
auto model_path = absl::GetFlag(FLAGS_model_path);
auto tile_width = absl::GetFlag(FLAGS_tile_width);
auto tile_height = absl::GetFlag(FLAGS_tile_height);
auto tile_pad = absl::GetFlag(FLAGS_tile_pad);
auto extend_grace = absl::GetFlag(FLAGS_extend_grace);
auto alignment = absl::GetFlag(FLAGS_alignment);
if (!exists(std::filesystem::path(model_path))) {
LOG(QFATAL) << "model path " << std::quoted(model_path) << " not exist.";
}
if (tile_width == 0 || tile_height == 0) {
LOG(QFATAL) << "Invalid tile size.";
}
if (tile_pad >= tile_width || tile_pad >= tile_height) {
LOG(QFATAL) << "Invalid tile pad size.";
}
if (extend_grace >= (tile_width - tile_pad)
|| extend_grace >= (tile_height - tile_pad)) {
LOG(QFATAL) << "Invalid tile extend grace.";
}
if (alignment == 0 || tile_width % alignment != 0 || tile_height % alignment != 0
|| tile_pad % alignment != 0 || extend_grace % alignment != 0) {
LOG(QFATAL) << "Invalid tile alignment.";
}
// ----------------------------------
// Lazy load
if (absl::GetFlag(FLAGS_cuda_lazy_load)) {
#ifdef _WIN32
SetEnvironmentVariableW(L"CUDA_MODULE_LOADING", L"LAZY");
#else
setenv("CUDA_MODULE_LOADING", "LAZY", 1);
#endif
}
// ----------------------------------
// IO
auto err = init_image_io();
if (!err.empty()) {
LOG(QFATAL) << "Failed init Image IO: " << err;
}
// ----------------------------------
// Layers
#if NV_TENSORRT_MAJOR < 9
plugins::register_resize_plugin();
#endif
// ----------------------------------
// Engine
auto max_width = absl::GetFlag(FLAGS_tile_width) + absl::GetFlag(FLAGS_extend_grace);
auto max_height = absl::GetFlag(FLAGS_tile_height) + absl::GetFlag(FLAGS_extend_grace);
InferenceContext ctx{
{
{int(std::min(
std::max(
absl::GetFlag(FLAGS_extend_grace) + absl::GetFlag(FLAGS_tile_pad),
absl::GetFlag(FLAGS_alignment)
),
MinDimension)),
int(absl::GetFlag(FLAGS_tile_width)),
int(max_width)},
{int(std::min(
std::max(
absl::GetFlag(FLAGS_extend_grace) + absl::GetFlag(FLAGS_tile_pad),
absl::GetFlag(FLAGS_alignment)
),
MinDimension)),
int(absl::GetFlag(FLAGS_tile_height)),
int(max_height)},
1,
absl::GetFlag(FLAGS_aux_stream),
absl::GetFlag(FLAGS_fp16),
absl::GetFlag(FLAGS_int8),
absl::GetFlag(FLAGS_force_precision),
absl::GetFlag(FLAGS_external),
absl::GetFlag(FLAGS_low_mem),
},
gLogger,
absl::GetFlag(FLAGS_model_path)
};
if (!ctx.has_file()) {
LOG(INFO) << "Building optimized engine for current tile config. This may take some time. "
"Some errors may occur, but as long as there are no fatal ones, this will be fine.";
err = OptimizationContext(ctx.config, gLogger, absl::GetFlag(FLAGS_model_path)).optimize();
if (!err.empty()) {
LOG(QFATAL) << "Failed building optimized engine: " << err;
}
}
err = ctx.load_engine();
if (!err.empty()) {
LOG(QFATAL) << "Failed loading engine: " << err;
}
session = new InferenceSession(ctx);
session->config(1, absl::GetFlag(FLAGS_tile_height), absl::GetFlag(FLAGS_tile_width));
err = session->init();
if (!err.empty()) {
LOG(QFATAL) << "Failed initialize context: " << err;
}
err = session->allocation();
if (!err.empty()) {
LOG(QFATAL) << "Failed allocate memory for context: " << err;
}
std::tie(h_scale, w_scale) = session->detect_scale();
if (h_scale == -1 || w_scale == -1) {
LOG(QFATAL) << "Bad model, can't detect scale ratio.";
}
if (h_scale != w_scale) {
LOG(QFATAL) << "different width and height scale ratio unimplemented.";
}
// ------------------------------
// Import & Export
auto max_size = size_t(max_width) * max_height;
if (absl::GetFlag(FLAGS_reformatter) == "auto") {
absl::SetFlag(&FLAGS_reformatter, absl::GetFlag(FLAGS_fp16) ? "gpu" : "cpu");
}
if (absl::GetFlag(FLAGS_fp16) && absl::GetFlag(FLAGS_reformatter) == "cpu") {
LOG(QFATAL) << "CPU reformatter can not handle FP16.";
}
if (absl::GetFlag(FLAGS_reformatter) == "cpu") {
importer_cpu = new pixel_importer_cpu(max_size, handle_alpha);
exporter_cpu = new pixel_exporter_cpu(h_scale * w_scale * max_size, handle_alpha);
using_io = 0;
}
else if (absl::GetFlag(FLAGS_reformatter) == "gpu") {
if (absl::GetFlag(FLAGS_fp16)) {
importer_gpu_fp16 = new pixel_importer_gpu<half>(max_size, handle_alpha);
exporter_gpu_fp16 =
new pixel_exporter_gpu<half>(h_scale * w_scale * max_size, handle_alpha);
using_io = 2;
}
else {
importer_gpu = new pixel_importer_gpu<float>(max_size, handle_alpha);
exporter_gpu =
new pixel_exporter_gpu<float>(h_scale * w_scale * max_size, handle_alpha);
using_io = 1;
}
}
else {
LOG(QFATAL) << "Unknown reformatter.";
}
}
struct runner {
chan works;
std::thread pipeline;
runner() : works{}, pipeline{launch_pipeline, std::ref(works), nullptr} {};
~runner() {
works.close();
pipeline.join();
}
};