-
Notifications
You must be signed in to change notification settings - Fork 56
/
yolov5.cpp
599 lines (486 loc) · 30.5 KB
/
yolov5.cpp
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
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
#include <iostream>
#include <chrono>
#include "cuda_runtime_api.h"
#include "logging.h"
#include "common.hpp"
#define USE_FP16 // comment out this if want to use FP32
#define DEVICE 0 // GPU id
#define NMS_THRESH 0.4
#define CONF_THRESH 0.5
#define BATCH_SIZE 1
#define NET s // s m l x
#define NETSTRUCT(str) createEngine_##str
#define CREATENET(net) NETSTRUCT(net)
#define STR1(x) #x
#define STR2(x) STR1(x)
// stuff we know about the network and the input/output blobs
static const int INPUT_H = Yolo::INPUT_H;
static const int INPUT_W = Yolo::INPUT_W;
static const int OUTPUT_SIZE = Yolo::MAX_OUTPUT_BBOX_COUNT * sizeof(Yolo::Detection) / sizeof(float) + 1; // we assume the yololayer outputs no more than 1000 boxes that conf >= 0.1
const char* INPUT_BLOB_NAME = "data";
const char* OUTPUT_BLOB_NAME = "prob";
static Logger gLogger;
// Creat the engine using only the API and not any parser.
ICudaEngine* createEngine_s(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt) {
INetworkDefinition* network = builder->createNetworkV2(0U);
// Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{3, INPUT_H, INPUT_W});
assert(data);
std::map<std::string, Weights> weightMap = loadWeights("../yolov5s.wts");
Weights emptywts{DataType::kFLOAT, nullptr, 0};
// yolov5 backbone
auto focus0 = focus(network, weightMap, *data, 3, 32, 3, "model.0");
auto conv1 = convBlock(network, weightMap, *focus0->getOutput(0), 64, 3, 2, 1, "model.1");
auto bottleneck_CSP2 = bottleneckCSP(network, weightMap, *conv1->getOutput(0), 64, 64, 1, true, 1, 0.5, "model.2");
auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), 128, 3, 2, 1, "model.3");
auto bottleneck_csp4 = bottleneckCSP(network, weightMap, *conv3->getOutput(0), 128, 128, 3, true, 1, 0.5, "model.4");
auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), 256, 3, 2, 1, "model.5");
auto bottleneck_csp6 = bottleneckCSP(network, weightMap, *conv5->getOutput(0), 256, 256, 3, true, 1, 0.5, "model.6");
auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), 512, 3, 2, 1, "model.7");
auto spp8 = SPP(network, weightMap, *conv7->getOutput(0), 512, 512, 5, 9, 13, "model.8");
// yolov5 head
auto bottleneck_csp9 = bottleneckCSP(network, weightMap, *spp8->getOutput(0), 512, 512, 1, false, 1, 0.5, "model.9");
auto conv10 = convBlock(network, weightMap, *bottleneck_csp9->getOutput(0), 256, 1, 1, 1, "model.10");
float *deval = reinterpret_cast<float*>(malloc(sizeof(float) * 256 * 2 * 2));
for (int i = 0; i < 256 * 2 * 2; i++) {
deval[i] = 1.0;
}
Weights deconvwts11{DataType::kFLOAT, deval, 256 * 2 * 2};
IDeconvolutionLayer* deconv11 = network->addDeconvolutionNd(*conv10->getOutput(0), 256, DimsHW{2, 2}, deconvwts11, emptywts);
deconv11->setStrideNd(DimsHW{2, 2});
deconv11->setNbGroups(256);
weightMap["deconv11"] = deconvwts11;
ITensor* inputTensors12[] = {deconv11->getOutput(0), bottleneck_csp6->getOutput(0)};
auto cat12 = network->addConcatenation(inputTensors12, 2);
auto bottleneck_csp13 = bottleneckCSP(network, weightMap, *cat12->getOutput(0), 512, 256, 1, false, 1, 0.5, "model.13");
auto conv14 = convBlock(network, weightMap, *bottleneck_csp13->getOutput(0), 128, 1, 1, 1, "model.14");
Weights deconvwts15{DataType::kFLOAT, deval, 128 * 2 * 2};
IDeconvolutionLayer* deconv15 = network->addDeconvolutionNd(*conv14->getOutput(0), 128, DimsHW{2, 2}, deconvwts15, emptywts);
deconv15->setStrideNd(DimsHW{2, 2});
deconv15->setNbGroups(128);
//weightMap["deconv15"] = deconvwts15;
ITensor* inputTensors16[] = {deconv15->getOutput(0), bottleneck_csp4->getOutput(0)};
auto cat16 = network->addConcatenation(inputTensors16, 2);
auto bottleneck_csp17 = bottleneckCSP(network, weightMap, *cat16->getOutput(0), 256, 128, 1, false, 1, 0.5, "model.17");
IConvolutionLayer* det0 = network->addConvolutionNd(*bottleneck_csp17->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{1, 1}, weightMap["model.24.m.0.weight"], weightMap["model.24.m.0.bias"]);
auto conv18 = convBlock(network, weightMap, *bottleneck_csp17->getOutput(0), 128, 3, 2, 1, "model.18");
ITensor* inputTensors19[] = {conv18->getOutput(0), conv14->getOutput(0)};
auto cat19 = network->addConcatenation(inputTensors19, 2);
auto bottleneck_csp20 = bottleneckCSP(network, weightMap, *cat19->getOutput(0), 256, 256, 1, false, 1, 0.5, "model.20");
IConvolutionLayer* det1 = network->addConvolutionNd(*bottleneck_csp20->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{1, 1}, weightMap["model.24.m.1.weight"], weightMap["model.24.m.1.bias"]);
auto conv21 = convBlock(network, weightMap, *bottleneck_csp20->getOutput(0), 256, 3, 2, 1, "model.21");
ITensor* inputTensors22[] = {conv21->getOutput(0), conv10->getOutput(0)};
auto cat22 = network->addConcatenation(inputTensors22, 2);
auto bottleneck_csp23 = bottleneckCSP(network, weightMap, *cat22->getOutput(0), 512, 512, 1, false, 1, 0.5, "model.23");
IConvolutionLayer* det2 = network->addConvolutionNd(*bottleneck_csp23->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{1, 1}, weightMap["model.24.m.2.weight"], weightMap["model.24.m.2.bias"]);
auto creator = getPluginRegistry()->getPluginCreator("YoloLayer_TRT", "1");
const PluginFieldCollection* pluginData = creator->getFieldNames();
IPluginV2 *pluginObj = creator->createPlugin("yololayer", pluginData);
ITensor* inputTensors_yolo[] = {det2->getOutput(0), det1->getOutput(0), det0->getOutput(0)};
auto yolo = network->addPluginV2(inputTensors_yolo, 3, *pluginObj);
yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
network->markOutput(*yolo->getOutput(0));
// Build engine
builder->setMaxBatchSize(maxBatchSize);
config->setMaxWorkspaceSize(16 * (1 << 20)); // 16MB
#ifdef USE_FP16
config->setFlag(BuilderFlag::kFP16);
#endif
std::cout << "Building engine, please wait for a while..." << std::endl;
ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
std::cout << "Build engine successfully!" << std::endl;
// Don't need the network any more
network->destroy();
// Release host memory
for (auto& mem : weightMap)
{
free((void*) (mem.second.values));
}
return engine;
}
ICudaEngine* createEngine_m(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt) {
INetworkDefinition* network = builder->createNetworkV2(0U);
// Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{ 3, INPUT_H, INPUT_W });
assert(data);
std::map<std::string, Weights> weightMap = loadWeights("../yolov5m.wts");
Weights emptywts{ DataType::kFLOAT, nullptr, 0 };
/* ------ yolov5 backbone------ */
auto focus0 = focus(network, weightMap, *data, 3, 48, 3, "model.0");
auto conv1 = convBlock(network, weightMap, *focus0->getOutput(0), 96, 3, 2, 1, "model.1");
auto bottleneck_CSP2 = bottleneckCSP(network, weightMap, *conv1->getOutput(0), 96, 96, 2, true, 1, 0.5, "model.2");
auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), 192, 3, 2, 1, "model.3");
auto bottleneck_csp4 = bottleneckCSP(network, weightMap, *conv3->getOutput(0), 192, 192, 6, true, 1, 0.5, "model.4");
auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), 384, 3, 2, 1, "model.5");
auto bottleneck_csp6 = bottleneckCSP(network, weightMap, *conv5->getOutput(0), 384, 384, 6, true, 1, 0.5, "model.6");
auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), 768, 3, 2, 1, "model.7");
auto spp8 = SPP(network, weightMap, *conv7->getOutput(0), 768, 768, 5, 9, 13, "model.8");
/* ------ yolov5 head ------ */
auto bottleneck_csp9 = bottleneckCSP(network, weightMap, *spp8->getOutput(0), 768, 768, 2, false, 1, 0.5, "model.9");
auto conv10 = convBlock(network, weightMap, *bottleneck_csp9->getOutput(0), 384, 1, 1, 1, "model.10");
float *deval = reinterpret_cast<float*>(malloc(sizeof(float) * 384 * 2 * 2));
for (int i = 0; i < 384 * 2 * 2; i++) {
deval[i] = 1.0;
}
Weights deconvwts11{ DataType::kFLOAT, deval, 384 * 2 * 2 };
IDeconvolutionLayer* deconv11 = network->addDeconvolutionNd(*conv10->getOutput(0), 384, DimsHW{ 2, 2 }, deconvwts11, emptywts);
deconv11->setStrideNd(DimsHW{ 2, 2 });
deconv11->setNbGroups(384);
weightMap["deconv11"] = deconvwts11;
ITensor* inputTensors12[] = { deconv11->getOutput(0), bottleneck_csp6->getOutput(0) };
auto cat12 = network->addConcatenation(inputTensors12, 2);
auto bottleneck_csp13 = bottleneckCSP(network, weightMap, *cat12->getOutput(0), 768, 384, 2, false, 1, 0.5, "model.13");
auto conv14 = convBlock(network, weightMap, *bottleneck_csp13->getOutput(0), 192, 1, 1, 1, "model.14");
Weights deconvwts15{ DataType::kFLOAT, deval, 192 * 2 * 2 };
IDeconvolutionLayer* deconv15 = network->addDeconvolutionNd(*conv14->getOutput(0), 192, DimsHW{ 2, 2 }, deconvwts15, emptywts);
deconv15->setStrideNd(DimsHW{ 2, 2 });
deconv15->setNbGroups(192);
ITensor* inputTensors16[] = { deconv15->getOutput(0), bottleneck_csp4->getOutput(0) };
auto cat16 = network->addConcatenation(inputTensors16, 2);
auto bottleneck_csp17 = bottleneckCSP(network, weightMap, *cat16->getOutput(0), 384, 192, 2, false, 1, 0.5, "model.17");
//yolo layer 0
IConvolutionLayer* det0 = network->addConvolutionNd(*bottleneck_csp17->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.0.weight"], weightMap["model.24.m.0.bias"]);
auto conv18 = convBlock(network, weightMap, *bottleneck_csp17->getOutput(0), 192, 3, 2, 1, "model.18");
ITensor* inputTensors19[] = {conv18->getOutput(0), conv14->getOutput(0)};
auto cat19 = network->addConcatenation(inputTensors19, 2);
auto bottleneck_csp20 = bottleneckCSP(network, weightMap, *cat19->getOutput(0), 384, 384, 2, false, 1, 0.5, "model.20");
//yolo layer 1
IConvolutionLayer* det1 = network->addConvolutionNd(*bottleneck_csp20->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.1.weight"], weightMap["model.24.m.1.bias"]);
auto conv21 = convBlock(network, weightMap, *bottleneck_csp20->getOutput(0), 384, 3, 2, 1, "model.21");
ITensor* inputTensors22[] = { conv21->getOutput(0), conv10->getOutput(0) };
auto cat22 = network->addConcatenation(inputTensors22, 2);
auto bottleneck_csp23 = bottleneckCSP(network, weightMap, *cat22->getOutput(0), 768, 768, 2, false, 1, 0.5, "model.23");
// yolo layer 2
IConvolutionLayer* det2 = network->addConvolutionNd(*bottleneck_csp23->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.2.weight"], weightMap["model.24.m.2.bias"]);
auto creator = getPluginRegistry()->getPluginCreator("YoloLayer_TRT", "1");
const PluginFieldCollection* pluginData = creator->getFieldNames();
IPluginV2 *pluginObj = creator->createPlugin("yololayer", pluginData);
ITensor* inputTensors_yolo[] = {det2->getOutput(0), det1->getOutput(0), det0->getOutput(0)};
auto yolo = network->addPluginV2(inputTensors_yolo, 3, *pluginObj);
yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
network->markOutput(*yolo->getOutput(0));
// Build engine
builder->setMaxBatchSize(maxBatchSize);
config->setMaxWorkspaceSize(16 * (1 << 20)); // 16MB
#ifdef USE_FP16
config->setFlag(BuilderFlag::kFP16);
#endif
std::cout << "Building engine, please wait for a while..." << std::endl;
ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
std::cout << "Build engine successfully!" << std::endl;
// Don't need the network any more
network->destroy();
// Release host memory
for (auto& mem : weightMap)
{
free((void*)(mem.second.values));
}
return engine;
}
ICudaEngine* createEngine_l(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt) {
INetworkDefinition* network = builder->createNetworkV2(0U);
// Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{ 3, INPUT_H, INPUT_W });
assert(data);
std::map<std::string, Weights> weightMap = loadWeights("../yolov5l.wts");
Weights emptywts{ DataType::kFLOAT, nullptr, 0 };
/* ------ yolov5 backbone------ */
auto focus0 = focus(network, weightMap, *data, 3, 64, 3, "model.0");
auto conv1 = convBlock(network, weightMap, *focus0->getOutput(0), 128, 3, 2, 1, "model.1");
auto bottleneck_CSP2 = bottleneckCSP(network, weightMap, *conv1->getOutput(0), 128, 128, 3, true, 1, 0.5, "model.2");
auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), 256, 3, 2, 1, "model.3");
auto bottleneck_csp4 = bottleneckCSP(network, weightMap, *conv3->getOutput(0), 256, 256, 9, true, 1, 0.5, "model.4");
auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), 512, 3, 2, 1, "model.5");
auto bottleneck_csp6 = bottleneckCSP(network, weightMap, *conv5->getOutput(0), 512, 512, 9, true, 1, 0.5, "model.6");
auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), 1024, 3, 2, 1, "model.7");
auto spp8 = SPP(network, weightMap, *conv7->getOutput(0), 1024, 1024, 5, 9, 13, "model.8");
/* ------ yolov5 head ------ */
auto bottleneck_csp9 = bottleneckCSP(network, weightMap, *spp8->getOutput(0), 1024, 1024, 3, false, 1, 0.5, "model.9");
auto conv10 = convBlock(network, weightMap, *bottleneck_csp9->getOutput(0), 512, 1, 1, 1, "model.10");
float *deval = reinterpret_cast<float*>(malloc(sizeof(float) * 512 * 2 * 2));
for (int i = 0; i < 512 * 2 * 2; i++) {
deval[i] = 1.0;
}
Weights deconvwts11{ DataType::kFLOAT, deval, 512 * 2 * 2 };
IDeconvolutionLayer* deconv11 = network->addDeconvolutionNd(*conv10->getOutput(0), 512, DimsHW{ 2, 2 }, deconvwts11, emptywts);
deconv11->setStrideNd(DimsHW{ 2, 2 });
deconv11->setNbGroups(512);
weightMap["deconv11"] = deconvwts11;
ITensor* inputTensors12[] = { deconv11->getOutput(0), bottleneck_csp6->getOutput(0) };
auto cat12 = network->addConcatenation(inputTensors12, 2);
auto bottleneck_csp13 = bottleneckCSP(network, weightMap, *cat12->getOutput(0), 1024, 512, 3, false, 1, 0.5, "model.13");
auto conv14 = convBlock(network, weightMap, *bottleneck_csp13->getOutput(0), 256, 1, 1, 1, "model.14");
Weights deconvwts15{ DataType::kFLOAT, deval, 256 * 2 * 2 };
IDeconvolutionLayer* deconv15 = network->addDeconvolutionNd(*conv14->getOutput(0), 256, DimsHW{ 2, 2 }, deconvwts15, emptywts);
deconv15->setStrideNd(DimsHW{ 2, 2 });
deconv15->setNbGroups(256);
ITensor* inputTensors16[] = {deconv15->getOutput(0), bottleneck_csp4->getOutput(0)};
auto cat16 = network->addConcatenation(inputTensors16, 2);
auto bottleneck_csp17 = bottleneckCSP(network, weightMap, *cat16->getOutput(0), 512, 256, 3, false, 1, 0.5, "model.17");
// yolo layer 0
IConvolutionLayer* det0 = network->addConvolutionNd(*bottleneck_csp17->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.0.weight"], weightMap["model.24.m.0.bias"]);
auto conv18 = convBlock(network, weightMap, *bottleneck_csp17->getOutput(0), 256, 3, 2, 1, "model.18");
ITensor* inputTensors19[] = {conv18->getOutput(0), conv14->getOutput(0)};
auto cat19 = network->addConcatenation(inputTensors19, 2);
auto bottleneck_csp20 = bottleneckCSP(network, weightMap, *cat19->getOutput(0), 512, 512, 3, false, 1, 0.5, "model.20");
//yolo layer 1
IConvolutionLayer* det1 = network->addConvolutionNd(*bottleneck_csp20->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.1.weight"], weightMap["model.24.m.1.bias"]);
auto conv21 = convBlock(network, weightMap, *bottleneck_csp20->getOutput(0), 512, 3, 2, 1, "model.21");
ITensor* inputTensors22[] = {conv21->getOutput(0), conv10->getOutput(0)};
auto cat22 = network->addConcatenation(inputTensors22, 2);
auto bottleneck_csp23 = bottleneckCSP(network, weightMap, *cat22->getOutput(0), 1024, 1024, 3, false, 1, 0.5, "model.23");
IConvolutionLayer* det2 = network->addConvolutionNd(*bottleneck_csp23->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.2.weight"], weightMap["model.24.m.2.bias"]);
auto creator = getPluginRegistry()->getPluginCreator("YoloLayer_TRT", "1");
const PluginFieldCollection* pluginData = creator->getFieldNames();
IPluginV2 *pluginObj = creator->createPlugin("yololayer", pluginData);
ITensor* inputTensors_yolo[] = {det2->getOutput(0), det1->getOutput(0), det0->getOutput(0)};
auto yolo = network->addPluginV2(inputTensors_yolo, 3, *pluginObj);
yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
network->markOutput(*yolo->getOutput(0));
// Build engine
builder->setMaxBatchSize(maxBatchSize);
config->setMaxWorkspaceSize(16 * (1 << 20)); // 16MB
#ifdef USE_FP16
config->setFlag(BuilderFlag::kFP16);
#endif
std::cout << "Building engine, please wait for a while..." << std::endl;
ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
std::cout << "Build engine successfully!" << std::endl;
// Don't need the network any more
network->destroy();
// Release host memory
for (auto& mem : weightMap)
{
free((void*)(mem.second.values));
}
return engine;
}
ICudaEngine* createEngine_x(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt) {
INetworkDefinition* network = builder->createNetworkV2(0U);
// Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{ 3, INPUT_H, INPUT_W });
assert(data);
std::map<std::string, Weights> weightMap = loadWeights("../yolov5x.wts");
Weights emptywts{ DataType::kFLOAT, nullptr, 0 };
/* ------ yolov5 backbone------ */
auto focus0 = focus(network, weightMap, *data, 3, 80, 3, "model.0");
auto conv1 = convBlock(network, weightMap, *focus0->getOutput(0), 160, 3, 2, 1, "model.1");
auto bottleneck_CSP2 = bottleneckCSP(network, weightMap, *conv1->getOutput(0), 160, 160, 4, true, 1, 0.5, "model.2");
auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), 320, 3, 2, 1, "model.3");
auto bottleneck_csp4 = bottleneckCSP(network, weightMap, *conv3->getOutput(0), 320, 320, 12, true, 1, 0.5, "model.4");
auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), 640, 3, 2, 1, "model.5");
auto bottleneck_csp6 = bottleneckCSP(network, weightMap, *conv5->getOutput(0), 640, 640, 12, true, 1, 0.5, "model.6");
auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), 1280, 3, 2, 1, "model.7");
auto spp8 = SPP(network, weightMap, *conv7->getOutput(0), 1280, 1280, 5, 9, 13, "model.8");
/* ------- yolov5 head ------- */
auto bottleneck_csp9 = bottleneckCSP(network, weightMap, *spp8->getOutput(0), 1280, 1280, 4, false, 1, 0.5, "model.9");
auto conv10 = convBlock(network, weightMap, *bottleneck_csp9->getOutput(0), 640, 1, 1, 1, "model.10");
float *deval = reinterpret_cast<float*>(malloc(sizeof(float) * 640 * 2 * 2));
for (int i = 0; i < 640 * 2 * 2; i++) {
deval[i] = 1.0;
}
Weights deconvwts11{ DataType::kFLOAT, deval, 640 * 2 * 2 };
IDeconvolutionLayer* deconv11 = network->addDeconvolutionNd(*conv10->getOutput(0), 640, DimsHW{ 2, 2 }, deconvwts11, emptywts);
deconv11->setStrideNd(DimsHW{ 2, 2 });
deconv11->setNbGroups(640);
weightMap["deconv11"] = deconvwts11;
ITensor* inputTensors12[] = { deconv11->getOutput(0), bottleneck_csp6->getOutput(0) };
auto cat12 = network->addConcatenation(inputTensors12, 2);
auto bottleneck_csp13 = bottleneckCSP(network, weightMap, *cat12->getOutput(0), 1280, 640, 4, false, 1, 0.5, "model.13");
auto conv14 = convBlock(network, weightMap, *bottleneck_csp13->getOutput(0), 320, 1, 1, 1, "model.14");
Weights deconvwts15{ DataType::kFLOAT, deval, 320 * 2 * 2 };
IDeconvolutionLayer* deconv15 = network->addDeconvolutionNd(*conv14->getOutput(0), 320, DimsHW{ 2, 2 }, deconvwts15, emptywts);
deconv15->setStrideNd(DimsHW{ 2, 2 });
deconv15->setNbGroups(320);
ITensor* inputTensors16[] = { deconv15->getOutput(0), bottleneck_csp4->getOutput(0) };
auto cat16 = network->addConcatenation(inputTensors16, 2);
auto bottleneck_csp17 = bottleneckCSP(network, weightMap, *cat16->getOutput(0), 640, 320, 4, false, 1, 0.5, "model.17");
// yolo layer 0
IConvolutionLayer* det0 = network->addConvolutionNd(*bottleneck_csp17->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.0.weight"], weightMap["model.24.m.0.bias"]);
auto conv18 = convBlock(network, weightMap, *bottleneck_csp17->getOutput(0), 320, 3, 2, 1, "model.18");
ITensor* inputTensors19[] = { conv18->getOutput(0), conv14->getOutput(0) };
auto cat19 = network->addConcatenation(inputTensors19, 2);
auto bottleneck_csp20 = bottleneckCSP(network, weightMap, *cat19->getOutput(0), 640, 640, 4, false, 1, 0.5, "model.20");
// yolo layer 1
IConvolutionLayer* det1 = network->addConvolutionNd(*bottleneck_csp20->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.1.weight"], weightMap["model.24.m.1.bias"]);
auto conv21 = convBlock(network, weightMap, *bottleneck_csp20->getOutput(0), 640, 3, 2, 1, "model.21");
ITensor* inputTensors22[] = { conv21->getOutput(0), conv10->getOutput(0) };
auto cat22 = network->addConcatenation(inputTensors22, 2);
auto bottleneck_csp23 = bottleneckCSP(network, weightMap, *cat22->getOutput(0), 1280, 1280, 4, false, 1, 0.5, "model.23");
// yolo layer 2
IConvolutionLayer* det2 = network->addConvolutionNd(*bottleneck_csp23->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.2.weight"], weightMap["model.24.m.2.bias"]);
auto creator = getPluginRegistry()->getPluginCreator("YoloLayer_TRT", "1");
const PluginFieldCollection* pluginData = creator->getFieldNames();
IPluginV2 *pluginObj = creator->createPlugin("yololayer", pluginData);
ITensor* inputTensors_yolo[] = { det2->getOutput(0), det1->getOutput(0), det0->getOutput(0) };
auto yolo = network->addPluginV2(inputTensors_yolo, 3, *pluginObj);
yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
network->markOutput(*yolo->getOutput(0));
// Build engine
builder->setMaxBatchSize(maxBatchSize);
config->setMaxWorkspaceSize(16 * (1 << 20)); // 16MB
#ifdef USE_FP16
config->setFlag(BuilderFlag::kFP16);
#endif
std::cout << "Building engine, please wait for a while..." << std::endl;
ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
std::cout << "Build engine successfully!" << std::endl;
// Don't need the network any more
network->destroy();
// Release host memory
for (auto& mem : weightMap)
{
free((void*)(mem.second.values));
}
return engine;
}
void APIToModel(unsigned int maxBatchSize, IHostMemory** modelStream) {
// Create builder
IBuilder* builder = createInferBuilder(gLogger);
IBuilderConfig* config = builder->createBuilderConfig();
// Create model to populate the network, then set the outputs and create an engine
ICudaEngine* engine = (CREATENET(NET))(maxBatchSize, builder, config, DataType::kFLOAT);
//ICudaEngine* engine = createEngine(maxBatchSize, builder, config, DataType::kFLOAT);
assert(engine != nullptr);
// Serialize the engine
(*modelStream) = engine->serialize();
// Close everything down
engine->destroy();
builder->destroy();
}
void doInference(IExecutionContext& context, float* input, float* output, int batchSize) {
const ICudaEngine& engine = context.getEngine();
// Pointers to input and output device buffers to pass to engine.
// Engine requires exactly IEngine::getNbBindings() number of buffers.
assert(engine.getNbBindings() == 2);
void* buffers[2];
// In order to bind the buffers, we need to know the names of the input and output tensors.
// Note that indices are guaranteed to be less than IEngine::getNbBindings()
const int inputIndex = engine.getBindingIndex(INPUT_BLOB_NAME);
const int outputIndex = engine.getBindingIndex(OUTPUT_BLOB_NAME);
// Create GPU buffers on device
CHECK(cudaMalloc(&buffers[inputIndex], batchSize * 3 * INPUT_H * INPUT_W * sizeof(float)));
CHECK(cudaMalloc(&buffers[outputIndex], batchSize * OUTPUT_SIZE * sizeof(float)));
// Create stream
cudaStream_t stream;
CHECK(cudaStreamCreate(&stream));
// DMA input batch data to device, infer on the batch asynchronously, and DMA output back to host
CHECK(cudaMemcpyAsync(buffers[inputIndex], input, batchSize * 3 * INPUT_H * INPUT_W * sizeof(float), cudaMemcpyHostToDevice, stream));
context.enqueue(batchSize, buffers, stream, nullptr);
CHECK(cudaMemcpyAsync(output, buffers[outputIndex], batchSize * OUTPUT_SIZE * sizeof(float), cudaMemcpyDeviceToHost, stream));
cudaStreamSynchronize(stream);
// Release stream and buffers
cudaStreamDestroy(stream);
CHECK(cudaFree(buffers[inputIndex]));
CHECK(cudaFree(buffers[outputIndex]));
}
int main(int argc, char** argv) {
cudaSetDevice(DEVICE);
// create a model using the API directly and serialize it to a stream
char *trtModelStream{nullptr};
size_t size{0};
std::string engine_name = STR2(NET);
engine_name = "yolov5" + engine_name + ".engine";
if (argc == 2 && std::string(argv[1]) == "-s") {
IHostMemory* modelStream{nullptr};
APIToModel(BATCH_SIZE, &modelStream);
assert(modelStream != nullptr);
std::ofstream p(engine_name, std::ios::binary);
if (!p) {
std::cerr << "could not open plan output file" << std::endl;
return -1;
}
p.write(reinterpret_cast<const char*>(modelStream->data()), modelStream->size());
modelStream->destroy();
return 0;
} else if (argc == 3 && std::string(argv[1]) == "-d") {
std::ifstream file(engine_name, std::ios::binary);
if (file.good()) {
file.seekg(0, file.end);
size = file.tellg();
file.seekg(0, file.beg);
trtModelStream = new char[size];
assert(trtModelStream);
file.read(trtModelStream, size);
file.close();
}
} else {
std::cerr << "arguments not right!" << std::endl;
std::cerr << "./yolov5 -s // serialize model to plan file" << std::endl;
std::cerr << "./yolov5 -d ../samples // deserialize plan file and run inference" << std::endl;
return -1;
}
std::vector<std::string> file_names;
if (read_files_in_dir(argv[2], file_names) < 0) {
std::cout << "read_files_in_dir failed." << std::endl;
return -1;
}
// prepare input data ---------------------------
static float data[BATCH_SIZE * 3 * INPUT_H * INPUT_W];
//for (int i = 0; i < 3 * INPUT_H * INPUT_W; i++)
// data[i] = 1.0;
static float prob[BATCH_SIZE * OUTPUT_SIZE];
IRuntime* runtime = createInferRuntime(gLogger);
assert(runtime != nullptr);
ICudaEngine* engine = runtime->deserializeCudaEngine(trtModelStream, size);
assert(engine != nullptr);
IExecutionContext* context = engine->createExecutionContext();
assert(context != nullptr);
delete[] trtModelStream;
int fcount = 0;
for (int f = 0; f < (int)file_names.size(); f++) {
fcount++;
if (fcount < BATCH_SIZE && f + 1 != (int)file_names.size()) continue;
for (int b = 0; b < fcount; b++) {
cv::Mat img = cv::imread(std::string(argv[2]) + "/" + file_names[f - fcount + 1 + b]);
if (img.empty()) continue;
cv::Mat pr_img = preprocess_img(img); // letterbox BGR to RGB
int i = 0;
for (int row = 0; row < INPUT_H; ++row) {
uchar* uc_pixel = pr_img.data + row * pr_img.step;
for (int col = 0; col < INPUT_W; ++col) {
data[b * 3 * INPUT_H * INPUT_W + i] = (float)uc_pixel[2] / 255.0;
data[b * 3 * INPUT_H * INPUT_W + i + INPUT_H * INPUT_W] = (float)uc_pixel[1] / 255.0;
data[b * 3 * INPUT_H * INPUT_W + i + 2 * INPUT_H * INPUT_W] = (float)uc_pixel[0] / 255.0;
uc_pixel += 3;
++i;
}
}
}
// Run inference
auto start = std::chrono::system_clock::now();
doInference(*context, data, prob, BATCH_SIZE);
auto end = std::chrono::system_clock::now();
std::cout << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count() << "ms" << std::endl;
std::vector<std::vector<Yolo::Detection>> batch_res(fcount);
for (int b = 0; b < fcount; b++) {
auto& res = batch_res[b];
nms(res, &prob[b * OUTPUT_SIZE], CONF_THRESH, NMS_THRESH);
}
for (int b = 0; b < fcount; b++) {
auto& res = batch_res[b];
//std::cout << res.size() << std::endl;
cv::Mat img = cv::imread(std::string(argv[2]) + "/" + file_names[f - fcount + 1 + b]);
for (size_t j = 0; j < res.size(); j++) {
cv::Rect r = get_rect(img, res[j].bbox);
cv::rectangle(img, r, cv::Scalar(0x27, 0xC1, 0x36), 2);
cv::putText(img, std::to_string((int)res[j].class_id), cv::Point(r.x, r.y - 1), cv::FONT_HERSHEY_PLAIN, 1.2, cv::Scalar(0xFF, 0xFF, 0xFF), 2);
}
cv::imwrite("_" + file_names[f - fcount + 1 + b], img);
}
fcount = 0;
}
// Destroy the engine
context->destroy();
engine->destroy();
runtime->destroy();
// Print histogram of the output distribution
//std::cout << "\nOutput:\n\n";
//for (unsigned int i = 0; i < OUTPUT_SIZE; i++)
//{
// std::cout << prob[i] << ", ";
// if (i % 10 == 0) std::cout << std::endl;
//}
//std::cout << std::endl;
return 0;
}