Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel
Resources  
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Explore Pricing
Product  
Data Engine LLMs Platform Enterprise
Connect to Dagshub Discord channel
Register
Login

Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git

Watch 4
Issues 62 Pull Requests 44 Discussions 0

#367 fix: Request correct hydra-core

Merged
Ghost merged 1 commits into Deci-AI:master from deci-ai:hotfix/ALG-000_hydra-req
Conversation 0 Files changed 1 Commits 1
File Compare
super-gradients
/
docs
/
super_gradients.training.datasets.detection_datasets.html
View left file View right file
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
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629 
  1. <!DOCTYPE html>
    2. 

  2. <html class="writer-html5" lang="en" >
    3. 

  3. <head>
    4. 

  4.   <meta charset="utf-8" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
    5. 

  5. 

  6.   <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    7. 

  7.   <title>super_gradients.training.datasets.detection_datasets package &mdash; SuperGradients 1.0 documentation</title>
    8. 

  8.       <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
    9. 

  9.       <link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
    10. 

  10.       <link rel="stylesheet" href="_static/graphviz.css" type="text/css" />
    11. 

  11.   <!--[if lt IE 9]>
    12. 

  12.     <script src="_static/js/html5shiv.min.js"></script>
    13. 

  13.   <![endif]-->
    14. 

  14.   
    15. 

  15.         <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
    16. 

  16.         <script src="_static/jquery.js"></script>
    17. 

  17.         <script src="_static/underscore.js"></script>
    18. 

  18.         <script src="_static/doctools.js"></script>
    19. 

  19.     <script src="_static/js/theme.js"></script>
    20. 

  20.     <link rel="index" title="Index" href="genindex.html" />
    21. 

  21.     <link rel="search" title="Search" href="search.html" /> 
    22. 

  22. </head>
    23. 

  23. 

  24. <body class="wy-body-for-nav"> 
    25. 

  25.   <div class="wy-grid-for-nav">
    26. 

  26.     <nav data-toggle="wy-nav-shift" class="wy-nav-side">
    27. 

  27.       <div class="wy-side-scroll">
    28. 

  28.         <div class="wy-side-nav-search" >
    29. 

  29.             <a href="index.html" class="icon icon-home"> SuperGradients
    30. 

  30.           </a>
    31. 

  31. <div role="search">
    32. 

  32.   <form id="rtd-search-form" class="wy-form" action="search.html" method="get">
    33. 

  33.     <input type="text" name="q" placeholder="Search docs" />
    34. 

  34.     <input type="hidden" name="check_keywords" value="yes" />
    35. 

  35.     <input type="hidden" name="area" value="default" />
    36. 

  36.   </form>
    37. 

  37. </div>
    38. 

  38.         </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
    39. 

  39.               <p class="caption"><span class="caption-text">Welcome To SuperGradients</span></p>
    40. 

  40. <ul>
    41. 

  41. <li class="toctree-l1"><a class="reference internal" href="welcome.html">Fill our 4-question quick survey! We will raffle free SuperGradients swag between those who will participate -&gt; Fill Survey</a></li>
    42. 

  42. <li class="toctree-l1"><a class="reference internal" href="welcome.html#supergradients">SuperGradients</a></li>
    43. 

  43. </ul>
    44. 

  44. <p class="caption"><span class="caption-text">Technical Documentation</span></p>
    45. 

  45. <ul>
    46. 

  46. <li class="toctree-l1"><a class="reference internal" href="super_gradients.common.html">Common package</a></li>
    47. 

  47. <li class="toctree-l1"><a class="reference internal" href="super_gradients.training.html">Training package</a></li>
    48. 

  48. </ul>
    49. 

  49. <p class="caption"><span class="caption-text">User Guide</span></p>
    50. 

  50. <ul>
    51. 

  51. <li class="toctree-l1"><a class="reference internal" href="user_guide.html">What is SuperGradients?</a></li>
    52. 

  52. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#introducing-the-supergradients-library">Introducing the SuperGradients library</a></li>
    53. 

  53. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#installation">Installation</a></li>
    54. 

  54. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#integrating-your-training-code-complete-walkthrough">Integrating Your Training Code - Complete Walkthrough</a></li>
    55. 

  55. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#training-parameters">Training Parameters</a></li>
    56. 

  56. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#logs-and-checkpoints">Logs and Checkpoints</a></li>
    57. 

  57. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#dataset-parameters">Dataset Parameters</a></li>
    58. 

  58. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#network-architectures">Network Architectures</a></li>
    59. 

  59. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#pretrained-models">Pretrained Models</a></li>
    60. 

  60. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#how-to-reproduce-our-training-recipes">How To Reproduce Our Training Recipes</a></li>
    61. 

  61. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#professional-tools-integration">Professional Tools Integration</a></li>
    62. 

  62. <li class="toctree-l1"><a class="reference internal" href="user_guide.html#supergradients-faq">SuperGradients FAQ</a></li>
    63. 

  63. </ul>
    64. 

  64. 

  65.         </div>
    66. 

  66.       </div>
    67. 

  67.     </nav>
    68. 

  68. 

  69.     <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
    70. 

  70.           <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
    71. 

  71.           <a href="index.html">SuperGradients</a>
    72. 

  72.       </nav>
    73. 

  73. 

  74.       <div class="wy-nav-content">
    75. 

  75.         <div class="rst-content">
    76. 

  76.           <div role="navigation" aria-label="Page navigation">
    77. 

  77.   <ul class="wy-breadcrumbs">
    78. 

  78.       <li><a href="index.html" class="icon icon-home"></a> &raquo;</li>
    79. 

  79.       <li>super_gradients.training.datasets.detection_datasets package</li>
    80. 

  80.       <li class="wy-breadcrumbs-aside">
    81. 

  81.             <a href="_sources/super_gradients.training.datasets.detection_datasets.rst.txt" rel="nofollow"> View page source</a>
    82. 

  82.       </li>
    83. 

  83.   </ul>
    84. 

  84.   <hr/>
    85. 

  85. </div>
    86. 

  86.           <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
    87. 

  87.            <div itemprop="articleBody">
    88. 

  88.              
    89. 

  89.   <section id="super-gradients-training-datasets-detection-datasets-package">
    90. 

  90. <h1>super_gradients.training.datasets.detection_datasets package<a class="headerlink" href="#super-gradients-training-datasets-detection-datasets-package" title="Permalink to this headline"></a></h1>
    91. 

  91. <section id="submodules">
    92. 

  92. <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline"></a></h2>
    93. 

  93. </section>
    94. 

  94. <section id="module-super_gradients.training.datasets.detection_datasets.coco_detection">
    95. 

  95. <span id="super-gradients-training-datasets-detection-datasets-coco-detection-module"></span><h2>super_gradients.training.datasets.detection_datasets.coco_detection module<a class="headerlink" href="#module-super_gradients.training.datasets.detection_datasets.coco_detection" title="Permalink to this headline"></a></h2>
    96. 

  96. <dl class="py class">
    97. 

  97. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset">
    98. 

  98. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.coco_detection.</span></span><span class="sig-name descname"><span class="pre">COCODetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_size</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">json_file</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'instances_train2017.json'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'images/train2017'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_dir_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tight_box_rotation</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">list</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_crowd</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset" title="Permalink to this definition"></a></dt>
    99. 

  99. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    100. 

  100. <p>Detection dataset COCO implementation</p>
    101. 

  101. <dl class="py method">
    102. 

  102. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_resized_img">
    103. 

  103. <span class="sig-name descname"><span class="pre">load_resized_img</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_resized_img"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_resized_img" title="Permalink to this definition"></a></dt>
    104. 

  104. <dd><p>Loads image at index, and resizes it to self.input_dim</p>
    105. 

  105. <dl class="field-list simple">
    106. 

  106. <dt class="field-odd">Parameters</dt>
    107. 

  107. <dd class="field-odd"><p><strong>index</strong> – index to load the image from</p>
    108. 

  108. </dd>
    109. 

  109. <dt class="field-even">Returns</dt>
    110. 

  110. <dd class="field-even"><p>resized_img</p>
    111. 

  111. </dd>
    112. 

  112. </dl>
    113. 

  113. </dd></dl>
    114. 

  114. 

  115. <dl class="py method">
    116. 

  116. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_sample">
    117. 

  117. <span class="sig-name descname"><span class="pre">load_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_sample" title="Permalink to this definition"></a></dt>
    118. 

  118. <dd><dl>
    119. 

  119. <dt>Loads sample at self.ids[index] as dictionary that holds:</dt><dd><p>“image”: Image resized to self.input_dim
    120. 

  120. “target”: Detection ground truth, np.array shaped (num_targets, 5), format is [class,x1,y1,x2,y2] with</p>
    121. 

  121. <blockquote>
    122. 

  122. <div><p>image coordinates.</p>
    123. 

  123. </div></blockquote>
    124. 

  124. <p>“target_seg”: Segmentation map convex hull derived detection target.
    125. 

  125. “info”: Original shape (height,width).
    126. 

  126. “id”: COCO image id</p>
    127. 

  127. </dd>
    128. 

  128. </dl>
    129. 

  129. <dl class="field-list simple">
    130. 

  130. <dt class="field-odd">Parameters</dt>
    131. 

  131. <dd class="field-odd"><p><strong>index</strong> – Sample index</p>
    132. 

  132. </dd>
    133. 

  133. <dt class="field-even">Returns</dt>
    134. 

  134. <dd class="field-even"><p>sample as described above</p>
    135. 

  135. </dd>
    136. 

  136. </dl>
    137. 

  137. </dd></dl>
    138. 

  138. 

  139. <dl class="py method">
    140. 

  140. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_image">
    141. 

  141. <span class="sig-name descname"><span class="pre">load_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.load_image" title="Permalink to this definition"></a></dt>
    142. 

  142. <dd><p>Loads image at index with its original resolution
    143. 

  143. :param index: index in self.annotations
    144. 

  144. :return: image (np.array)</p>
    145. 

  145. </dd></dl>
    146. 

  146. 

  147. <dl class="py method">
    148. 

  148. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.apply_transforms">
    149. 

  149. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.COCODetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    150. 

  150. <dd><p>Applies self.transforms sequentially to sample</p>
    151. 

  151. <dl class="simple">
    152. 

  152. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_targets” will load
    153. 

  153. only additional samples with objects in them.</p>
    154. 

  154. </dd>
    155. 

  155. </dl>
    156. 

  156. <dl class="field-list simple">
    157. 

  157. <dt class="field-odd">Parameters</dt>
    158. 

  158. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.load_sample)</p>
    159. 

  159. </dd>
    160. 

  160. <dt class="field-even">Returns</dt>
    161. 

  161. <dd class="field-even"><p>Transformed sample</p>
    162. 

  162. </dd>
    163. 

  163. </dl>
    164. 

  164. </dd></dl>
    165. 

  165. 

  166. </dd></dl>
    167. 

  167. 

  168. <dl class="py function">
    169. 

  169. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.coco_detection.remove_useless_info">
    170. 

  170. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.coco_detection.</span></span><span class="sig-name descname"><span class="pre">remove_useless_info</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">coco</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_seg_info</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#remove_useless_info"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.coco_detection.remove_useless_info" title="Permalink to this definition"></a></dt>
    171. 

  171. <dd><p>Remove useless info in coco dataset. COCO object is modified inplace.
    172. 

  172. This function is mainly used for saving memory (save about 30% mem).</p>
    173. 

  173. </dd></dl>
    174. 

  174. 

  175. </section>
    176. 

  176. <section id="module-super_gradients.training.datasets.detection_datasets.detection_dataset">
    177. 

  177. <span id="super-gradients-training-datasets-detection-datasets-detection-dataset-module"></span><h2>super_gradients.training.datasets.detection_datasets.detection_dataset module<a class="headerlink" href="#module-super_gradients.training.datasets.detection_datasets.detection_dataset" title="Permalink to this headline"></a></h2>
    178. 

  178. <dl class="py class">
    179. 

  179. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset">
    180. 

  180. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.detection_dataset.</span></span><span class="sig-name descname"><span class="pre">DetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_dim</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">original_target_format</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="super_gradients.training.utils.html#super_gradients.training.utils.detection_utils.DetectionTargetsFormat" title="super_gradients.training.utils.detection_utils.DetectionTargetsFormat"><span class="pre">super_gradients.training.utils.detection_utils.DetectionTargetsFormat</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_num_samples</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">super_gradients.training.transforms.transforms.DetectionTransform</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_classes_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_inclusion_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_empty_annotations</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset" title="Permalink to this definition"></a></dt>
    181. 

  181. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    182. 

  182. <p>Detection dataset.</p>
    183. 

  183. <p>This is a boilerplate class to facilitate the implementation of datasets.</p>
    184. 

  184. <dl>
    185. 

  185. <dt>HOW TO CREATE A DATASET THAT INHERITS FROM DetectionDataSet ?</dt><dd><ul class="simple">
    186. 

  186. <li><p>Inherit from DetectionDataSet</p></li>
    187. 

  187. <li><p>implement the method self._load_annotation to return at least the fields “target” and “img_path”</p></li>
    188. 

  188. <li><dl class="simple">
    189. 

  189. <dt>Call super().__init__ with the required params.</dt><dd><dl class="simple">
    190. 

  190. <dt>//!super().__init__ will call self._load_annotation, so make sure that every required</dt><dd><p>attributes are set up before calling super().__init__ (ideally just call it last)</p>
    191. 

  191. </dd>
    192. 

  192. </dl>
    193. 

  193. </dd>
    194. 

  194. </dl>
    195. 

  195. </li>
    196. 

  196. </ul>
    197. 

  197. </dd>
    198. 

  198. <dt>WORKFLOW:</dt><dd><ul class="simple">
    199. 

  199. <li><dl class="simple">
    200. 

  200. <dt>On instantiation:</dt><dd><ul>
    201. 

  201. <li><p>All annotations are cached. If class_inclusion_list was specified, there is also subclassing at this step.</p></li>
    202. 

  202. <li><p>If cache is True, the images are also cached</p></li>
    203. 

  203. </ul>
    204. 

  204. </dd>
    205. 

  205. </dl>
    206. 

  206. </li>
    207. 

  207. <li><dl class="simple">
    208. 

  208. <dt>On call (__getitem__) for a specific image index:</dt><dd><ul>
    209. 

  209. <li><p>The image and annotations are grouped together in a dict called SAMPLE</p></li>
    210. 

  210. <li><p>the sample is processed according to th transform</p></li>
    211. 

  211. <li><p>Only the specified fields are returned by __getitem__</p></li>
    212. 

  212. </ul>
    213. 

  213. </dd>
    214. 

  214. </dl>
    215. 

  215. </li>
    216. 

  216. </ul>
    217. 

  217. </dd>
    218. 

  218. <dt>TERMINOLOGY</dt><dd><ul>
    219. 

  219. <li><p>TARGET:       Groundtruth, made of bboxes. The format can vary from one dataset to another</p></li>
    220. 

  220. <li><dl class="simple">
    221. 

  221. <dt>ANNOTATION:   Combination of targets (groundtruth) and metadata of the image, but without the image itself.</dt><dd><p>&gt; Has to include the fields “target” and “img_path”
    222. 

  222. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    223. 

  223. </dd>
    224. 

  224. </dl>
    225. 

  225. </li>
    226. 

  226. <li><dl class="simple">
    227. 

  227. <dt>SAMPLE:       Outout of the dataset:</dt><dd><p>&gt; Has to include the fields “target” and “image”
    228. 

  228. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    229. 

  229. </dd>
    230. 

  230. </dl>
    231. 

  231. </li>
    232. 

  232. <li><p>INDEX:        Refers to the index in the dataset.</p></li>
    233. 

  233. <li><dl>
    234. 

  234. <dt>SAMPLE ID:    Refers to the id of sample before droping any annotaion.</dt><dd><p>Let’s imagine a situation where the downloaded data is made of 120 images but 20 were drop
    235. 

  235. because they had no annotation. In that case:</p>
    236. 

  236. <blockquote>
    237. 

  237. <div><p>&gt; We have 120 samples so sample_id will be between 0 and 119
    238. 

  238. &gt; But only 100 will be indexed so index will be between 0 and 99
    239. 

  239. &gt; Therefore, we also have len(self) = 100</p>
    240. 

  240. </div></blockquote>
    241. 

  241. </dd>
    242. 

  242. </dl>
    243. 

  243. </li>
    244. 

  244. </ul>
    245. 

  245. </dd>
    246. 

  246. </dl>
    247. 

  247. <dl class="py method">
    248. 

  248. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_item">
    249. 

  249. <span class="sig-name descname"><span class="pre">get_random_item</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_item"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_item" title="Permalink to this definition"></a></dt>
    250. 

  250. <dd></dd></dl>
    251. 

  251. 

  252. <dl class="py method">
    253. 

  253. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_sample">
    254. 

  254. <span class="sig-name descname"><span class="pre">get_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_sample" title="Permalink to this definition"></a></dt>
    255. 

  255. <dd><p>Get raw sample, before any transform (beside subclassing).
    256. 

  256. :param index:   Image index
    257. 

  257. :return:        Sample, i.e. a dictionary including at least “image” and “target”</p>
    258. 

  258. </dd></dl>
    259. 

  259. 

  260. <dl class="py method">
    261. 

  261. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_resized_image">
    262. 

  262. <span class="sig-name descname"><span class="pre">get_resized_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">numpy.ndarray</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_resized_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_resized_image" title="Permalink to this definition"></a></dt>
    263. 

  263. <dd><p>Get the resized image at a specific sample_id, either from cache or by loading from disk, based on self.cached_imgs
    264. 

  264. :param index:  Image index
    265. 

  265. :return:       Resized image</p>
    266. 

  266. </dd></dl>
    267. 

  267. 

  268. <dl class="py method">
    269. 

  269. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.apply_transforms">
    270. 

  270. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    271. 

  271. <dd><p>Applies self.transforms sequentially to sample</p>
    272. 

  272. <dl class="simple">
    273. 

  273. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_annotations” will load
    274. 

  274. only additional samples with objects in them.</p>
    275. 

  275. </dd>
    276. 

  276. </dl>
    277. 

  277. <dl class="field-list simple">
    278. 

  278. <dt class="field-odd">Parameters</dt>
    279. 

  279. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.get_sample)</p>
    280. 

  280. </dd>
    281. 

  281. <dt class="field-even">Returns</dt>
    282. 

  282. <dd class="field-even"><p>Transformed sample</p>
    283. 

  283. </dd>
    284. 

  284. </dl>
    285. 

  285. </dd></dl>
    286. 

  286. 

  287. <dl class="py method">
    288. 

  288. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_samples">
    289. 

  289. <span class="sig-name descname"><span class="pre">get_random_samples</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">count</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_samples"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_samples" title="Permalink to this definition"></a></dt>
    290. 

  290. <dd><p>Load random samples.</p>
    291. 

  291. <dl class="field-list simple">
    292. 

  292. <dt class="field-odd">Parameters</dt>
    293. 

  293. <dd class="field-odd"><ul class="simple">
    294. 

  294. <li><p><strong>count</strong> – The number of samples wanted</p></li>
    295. 

  295. <li><p><strong>non_empty_annotations_only</strong> – If true, only return samples with at least 1 annotation</p></li>
    296. 

  296. </ul>
    297. 

  297. </dd>
    298. 

  298. <dt class="field-even">Returns</dt>
    299. 

  299. <dd class="field-even"><p>A list of samples satisfying input params</p>
    300. 

  300. </dd>
    301. 

  301. </dl>
    302. 

  302. </dd></dl>
    303. 

  303. 

  304. <dl class="py method">
    305. 

  305. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_sample">
    306. 

  306. <span class="sig-name descname"><span class="pre">get_random_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.get_random_sample" title="Permalink to this definition"></a></dt>
    307. 

  307. <dd></dd></dl>
    308. 

  308. 

  309. <dl class="py property">
    310. 

  310. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.output_target_format">
    311. 

  311. <em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">output_target_format</span></span><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.output_target_format" title="Permalink to this definition"></a></dt>
    312. 

  312. <dd></dd></dl>
    313. 

  313. 

  314. <dl class="py method">
    315. 

  315. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.plot">
    316. 

  316. <span class="sig-name descname"><span class="pre">plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_samples_per_plot</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_plots</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_transformed_data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.plot"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.detection_dataset.DetectionDataset.plot" title="Permalink to this definition"></a></dt>
    317. 

  317. <dd><p>Combine samples of images with bbox into plots and display the result.</p>
    318. 

  318. <dl class="field-list simple">
    319. 

  319. <dt class="field-odd">Parameters</dt>
    320. 

  320. <dd class="field-odd"><ul class="simple">
    321. 

  321. <li><p><strong>max_samples_per_plot</strong> – Maximum number of images to be displayed per plot</p></li>
    322. 

  322. <li><p><strong>n_plots</strong> – Number of plots to display (each plot being a combination of img with bbox)</p></li>
    323. 

  323. <li><p><strong>plot_transformed_data</strong> – If True, the plot will be over samples after applying transforms (i.e. on __getitem__).
    324. 

  324. If False, the plot will be over the raw samples (i.e. on get_sample)</p></li>
    325. 

  325. </ul>
    326. 

  326. </dd>
    327. 

  327. <dt class="field-even">Returns</dt>
    328. 

  328. <dd class="field-even"><p></p>
    329. 

  329. </dd>
    330. 

  330. </dl>
    331. 

  331. </dd></dl>
    332. 

  332. 

  333. </dd></dl>
    334. 

  334. 

  335. </section>
    336. 

  336. <section id="module-super_gradients.training.datasets.detection_datasets.pascal_voc_detection">
    337. 

  337. <span id="super-gradients-training-datasets-detection-datasets-pascal-voc-detection-module"></span><h2>super_gradients.training.datasets.detection_datasets.pascal_voc_detection module<a class="headerlink" href="#module-super_gradients.training.datasets.detection_datasets.pascal_voc_detection" title="Permalink to this headline"></a></h2>
    338. 

  338. <dl class="py class">
    339. 

  339. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.pascal_voc_detection.PascalVOCDetectionDataset">
    340. 

  340. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.pascal_voc_detection.</span></span><span class="sig-name descname"><span class="pre">PascalVOCDetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">images_sub_directory</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.pascal_voc_detection.PascalVOCDetectionDataset" title="Permalink to this definition"></a></dt>
    341. 

  341. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    342. 

  342. <p>Dataset for Pascal VOC object detection</p>
    343. 

  343. <dl class="py method">
    344. 

  344. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.pascal_voc_detection.PascalVOCDetectionDataset.download">
    345. 

  345. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">download</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset.download"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.pascal_voc_detection.PascalVOCDetectionDataset.download" title="Permalink to this definition"></a></dt>
    346. 

  346. <dd><p>Download Pascal dataset in XYXY_LABEL format.</p>
    347. 

  347. <p>Data extracted form <a class="reference external" href="http://host.robots.ox.ac.uk/pascal/VOC/">http://host.robots.ox.ac.uk/pascal/VOC/</a></p>
    348. 

  348. </dd></dl>
    349. 

  349. 

  350. </dd></dl>
    351. 

  351. 

  352. </section>
    353. 

  353. <section id="module-super_gradients.training.datasets.detection_datasets">
    354. 

  354. <span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-super_gradients.training.datasets.detection_datasets" title="Permalink to this headline"></a></h2>
    355. 

  355. <dl class="py class">
    356. 

  356. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.COCODetectionDataset">
    357. 

  357. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.</span></span><span class="sig-name descname"><span class="pre">COCODetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_size</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">json_file</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'instances_train2017.json'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'images/train2017'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_dir_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tight_box_rotation</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">list</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_crowd</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.COCODetectionDataset" title="Permalink to this definition"></a></dt>
    358. 

  358. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    359. 

  359. <p>Detection dataset COCO implementation</p>
    360. 

  360. <dl class="py method">
    361. 

  361. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_resized_img">
    362. 

  362. <span class="sig-name descname"><span class="pre">load_resized_img</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_resized_img"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_resized_img" title="Permalink to this definition"></a></dt>
    363. 

  363. <dd><p>Loads image at index, and resizes it to self.input_dim</p>
    364. 

  364. <dl class="field-list simple">
    365. 

  365. <dt class="field-odd">Parameters</dt>
    366. 

  366. <dd class="field-odd"><p><strong>index</strong> – index to load the image from</p>
    367. 

  367. </dd>
    368. 

  368. <dt class="field-even">Returns</dt>
    369. 

  369. <dd class="field-even"><p>resized_img</p>
    370. 

  370. </dd>
    371. 

  371. </dl>
    372. 

  372. </dd></dl>
    373. 

  373. 

  374. <dl class="py method">
    375. 

  375. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_sample">
    376. 

  376. <span class="sig-name descname"><span class="pre">load_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_sample" title="Permalink to this definition"></a></dt>
    377. 

  377. <dd><dl>
    378. 

  378. <dt>Loads sample at self.ids[index] as dictionary that holds:</dt><dd><p>“image”: Image resized to self.input_dim
    379. 

  379. “target”: Detection ground truth, np.array shaped (num_targets, 5), format is [class,x1,y1,x2,y2] with</p>
    380. 

  380. <blockquote>
    381. 

  381. <div><p>image coordinates.</p>
    382. 

  382. </div></blockquote>
    383. 

  383. <p>“target_seg”: Segmentation map convex hull derived detection target.
    384. 

  384. “info”: Original shape (height,width).
    385. 

  385. “id”: COCO image id</p>
    386. 

  386. </dd>
    387. 

  387. </dl>
    388. 

  388. <dl class="field-list simple">
    389. 

  389. <dt class="field-odd">Parameters</dt>
    390. 

  390. <dd class="field-odd"><p><strong>index</strong> – Sample index</p>
    391. 

  391. </dd>
    392. 

  392. <dt class="field-even">Returns</dt>
    393. 

  393. <dd class="field-even"><p>sample as described above</p>
    394. 

  394. </dd>
    395. 

  395. </dl>
    396. 

  396. </dd></dl>
    397. 

  397. 

  398. <dl class="py method">
    399. 

  399. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_image">
    400. 

  400. <span class="sig-name descname"><span class="pre">load_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.COCODetectionDataset.load_image" title="Permalink to this definition"></a></dt>
    401. 

  401. <dd><p>Loads image at index with its original resolution
    402. 

  402. :param index: index in self.annotations
    403. 

  403. :return: image (np.array)</p>
    404. 

  404. </dd></dl>
    405. 

  405. 

  406. <dl class="py method">
    407. 

  407. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.COCODetectionDataset.apply_transforms">
    408. 

  408. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.COCODetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    409. 

  409. <dd><p>Applies self.transforms sequentially to sample</p>
    410. 

  410. <dl class="simple">
    411. 

  411. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_targets” will load
    412. 

  412. only additional samples with objects in them.</p>
    413. 

  413. </dd>
    414. 

  414. </dl>
    415. 

  415. <dl class="field-list simple">
    416. 

  416. <dt class="field-odd">Parameters</dt>
    417. 

  417. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.load_sample)</p>
    418. 

  418. </dd>
    419. 

  419. <dt class="field-even">Returns</dt>
    420. 

  420. <dd class="field-even"><p>Transformed sample</p>
    421. 

  421. </dd>
    422. 

  422. </dl>
    423. 

  423. </dd></dl>
    424. 

  424. 

  425. </dd></dl>
    426. 

  426. 

  427. <dl class="py class">
    428. 

  428. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset">
    429. 

  429. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.</span></span><span class="sig-name descname"><span class="pre">DetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_dim</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">original_target_format</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="super_gradients.training.utils.html#super_gradients.training.utils.detection_utils.DetectionTargetsFormat" title="super_gradients.training.utils.detection_utils.DetectionTargetsFormat"><span class="pre">super_gradients.training.utils.detection_utils.DetectionTargetsFormat</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_num_samples</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">super_gradients.training.transforms.transforms.DetectionTransform</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_classes_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_inclusion_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_empty_annotations</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset" title="Permalink to this definition"></a></dt>
    430. 

  430. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    431. 

  431. <p>Detection dataset.</p>
    432. 

  432. <p>This is a boilerplate class to facilitate the implementation of datasets.</p>
    433. 

  433. <dl>
    434. 

  434. <dt>HOW TO CREATE A DATASET THAT INHERITS FROM DetectionDataSet ?</dt><dd><ul class="simple">
    435. 

  435. <li><p>Inherit from DetectionDataSet</p></li>
    436. 

  436. <li><p>implement the method self._load_annotation to return at least the fields “target” and “img_path”</p></li>
    437. 

  437. <li><dl class="simple">
    438. 

  438. <dt>Call super().__init__ with the required params.</dt><dd><dl class="simple">
    439. 

  439. <dt>//!super().__init__ will call self._load_annotation, so make sure that every required</dt><dd><p>attributes are set up before calling super().__init__ (ideally just call it last)</p>
    440. 

  440. </dd>
    441. 

  441. </dl>
    442. 

  442. </dd>
    443. 

  443. </dl>
    444. 

  444. </li>
    445. 

  445. </ul>
    446. 

  446. </dd>
    447. 

  447. <dt>WORKFLOW:</dt><dd><ul class="simple">
    448. 

  448. <li><dl class="simple">
    449. 

  449. <dt>On instantiation:</dt><dd><ul>
    450. 

  450. <li><p>All annotations are cached. If class_inclusion_list was specified, there is also subclassing at this step.</p></li>
    451. 

  451. <li><p>If cache is True, the images are also cached</p></li>
    452. 

  452. </ul>
    453. 

  453. </dd>
    454. 

  454. </dl>
    455. 

  455. </li>
    456. 

  456. <li><dl class="simple">
    457. 

  457. <dt>On call (__getitem__) for a specific image index:</dt><dd><ul>
    458. 

  458. <li><p>The image and annotations are grouped together in a dict called SAMPLE</p></li>
    459. 

  459. <li><p>the sample is processed according to th transform</p></li>
    460. 

  460. <li><p>Only the specified fields are returned by __getitem__</p></li>
    461. 

  461. </ul>
    462. 

  462. </dd>
    463. 

  463. </dl>
    464. 

  464. </li>
    465. 

  465. </ul>
    466. 

  466. </dd>
    467. 

  467. <dt>TERMINOLOGY</dt><dd><ul>
    468. 

  468. <li><p>TARGET:       Groundtruth, made of bboxes. The format can vary from one dataset to another</p></li>
    469. 

  469. <li><dl class="simple">
    470. 

  470. <dt>ANNOTATION:   Combination of targets (groundtruth) and metadata of the image, but without the image itself.</dt><dd><p>&gt; Has to include the fields “target” and “img_path”
    471. 

  471. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    472. 

  472. </dd>
    473. 

  473. </dl>
    474. 

  474. </li>
    475. 

  475. <li><dl class="simple">
    476. 

  476. <dt>SAMPLE:       Outout of the dataset:</dt><dd><p>&gt; Has to include the fields “target” and “image”
    477. 

  477. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    478. 

  478. </dd>
    479. 

  479. </dl>
    480. 

  480. </li>
    481. 

  481. <li><p>INDEX:        Refers to the index in the dataset.</p></li>
    482. 

  482. <li><dl>
    483. 

  483. <dt>SAMPLE ID:    Refers to the id of sample before droping any annotaion.</dt><dd><p>Let’s imagine a situation where the downloaded data is made of 120 images but 20 were drop
    484. 

  484. because they had no annotation. In that case:</p>
    485. 

  485. <blockquote>
    486. 

  486. <div><p>&gt; We have 120 samples so sample_id will be between 0 and 119
    487. 

  487. &gt; But only 100 will be indexed so index will be between 0 and 99
    488. 

  488. &gt; Therefore, we also have len(self) = 100</p>
    489. 

  489. </div></blockquote>
    490. 

  490. </dd>
    491. 

  491. </dl>
    492. 

  492. </li>
    493. 

  493. </ul>
    494. 

  494. </dd>
    495. 

  495. </dl>
    496. 

  496. <dl class="py method">
    497. 

  497. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_item">
    498. 

  498. <span class="sig-name descname"><span class="pre">get_random_item</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_item"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_item" title="Permalink to this definition"></a></dt>
    499. 

  499. <dd></dd></dl>
    500. 

  500. 

  501. <dl class="py method">
    502. 

  502. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.get_sample">
    503. 

  503. <span class="sig-name descname"><span class="pre">get_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.get_sample" title="Permalink to this definition"></a></dt>
    504. 

  504. <dd><p>Get raw sample, before any transform (beside subclassing).
    505. 

  505. :param index:   Image index
    506. 

  506. :return:        Sample, i.e. a dictionary including at least “image” and “target”</p>
    507. 

  507. </dd></dl>
    508. 

  508. 

  509. <dl class="py method">
    510. 

  510. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.get_resized_image">
    511. 

  511. <span class="sig-name descname"><span class="pre">get_resized_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">numpy.ndarray</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_resized_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.get_resized_image" title="Permalink to this definition"></a></dt>
    512. 

  512. <dd><p>Get the resized image at a specific sample_id, either from cache or by loading from disk, based on self.cached_imgs
    513. 

  513. :param index:  Image index
    514. 

  514. :return:       Resized image</p>
    515. 

  515. </dd></dl>
    516. 

  516. 

  517. <dl class="py method">
    518. 

  518. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.apply_transforms">
    519. 

  519. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    520. 

  520. <dd><p>Applies self.transforms sequentially to sample</p>
    521. 

  521. <dl class="simple">
    522. 

  522. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_annotations” will load
    523. 

  523. only additional samples with objects in them.</p>
    524. 

  524. </dd>
    525. 

  525. </dl>
    526. 

  526. <dl class="field-list simple">
    527. 

  527. <dt class="field-odd">Parameters</dt>
    528. 

  528. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.get_sample)</p>
    529. 

  529. </dd>
    530. 

  530. <dt class="field-even">Returns</dt>
    531. 

  531. <dd class="field-even"><p>Transformed sample</p>
    532. 

  532. </dd>
    533. 

  533. </dl>
    534. 

  534. </dd></dl>
    535. 

  535. 

  536. <dl class="py method">
    537. 

  537. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_samples">
    538. 

  538. <span class="sig-name descname"><span class="pre">get_random_samples</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">count</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_samples"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_samples" title="Permalink to this definition"></a></dt>
    539. 

  539. <dd><p>Load random samples.</p>
    540. 

  540. <dl class="field-list simple">
    541. 

  541. <dt class="field-odd">Parameters</dt>
    542. 

  542. <dd class="field-odd"><ul class="simple">
    543. 

  543. <li><p><strong>count</strong> – The number of samples wanted</p></li>
    544. 

  544. <li><p><strong>non_empty_annotations_only</strong> – If true, only return samples with at least 1 annotation</p></li>
    545. 

  545. </ul>
    546. 

  546. </dd>
    547. 

  547. <dt class="field-even">Returns</dt>
    548. 

  548. <dd class="field-even"><p>A list of samples satisfying input params</p>
    549. 

  549. </dd>
    550. 

  550. </dl>
    551. 

  551. </dd></dl>
    552. 

  552. 

  553. <dl class="py method">
    554. 

  554. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_sample">
    555. 

  555. <span class="sig-name descname"><span class="pre">get_random_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.get_random_sample" title="Permalink to this definition"></a></dt>
    556. 

  556. <dd></dd></dl>
    557. 

  557. 

  558. <dl class="py property">
    559. 

  559. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.output_target_format">
    560. 

  560. <em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">output_target_format</span></span><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.output_target_format" title="Permalink to this definition"></a></dt>
    561. 

  561. <dd></dd></dl>
    562. 

  562. 

  563. <dl class="py method">
    564. 

  564. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.DetectionDataset.plot">
    565. 

  565. <span class="sig-name descname"><span class="pre">plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_samples_per_plot</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_plots</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_transformed_data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.plot"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.DetectionDataset.plot" title="Permalink to this definition"></a></dt>
    566. 

  566. <dd><p>Combine samples of images with bbox into plots and display the result.</p>
    567. 

  567. <dl class="field-list simple">
    568. 

  568. <dt class="field-odd">Parameters</dt>
    569. 

  569. <dd class="field-odd"><ul class="simple">
    570. 

  570. <li><p><strong>max_samples_per_plot</strong> – Maximum number of images to be displayed per plot</p></li>
    571. 

  571. <li><p><strong>n_plots</strong> – Number of plots to display (each plot being a combination of img with bbox)</p></li>
    572. 

  572. <li><p><strong>plot_transformed_data</strong> – If True, the plot will be over samples after applying transforms (i.e. on __getitem__).
    573. 

  573. If False, the plot will be over the raw samples (i.e. on get_sample)</p></li>
    574. 

  574. </ul>
    575. 

  575. </dd>
    576. 

  576. <dt class="field-even">Returns</dt>
    577. 

  577. <dd class="field-even"><p></p>
    578. 

  578. </dd>
    579. 

  579. </dl>
    580. 

  580. </dd></dl>
    581. 

  581. 

  582. </dd></dl>
    583. 

  583. 

  584. <dl class="py class">
    585. 

  585. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.PascalVOCDetectionDataset">
    586. 

  586. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.detection_datasets.</span></span><span class="sig-name descname"><span class="pre">PascalVOCDetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">images_sub_directory</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.PascalVOCDetectionDataset" title="Permalink to this definition"></a></dt>
    587. 

  587. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Generic</span></code>[<code class="xref py py-obj docutils literal notranslate"><span class="pre">torch.utils.data.dataset.T_co</span></code>]</p>
    588. 

  588. <p>Dataset for Pascal VOC object detection</p>
    589. 

  589. <dl class="py method">
    590. 

  590. <dt class="sig sig-object py" id="super_gradients.training.datasets.detection_datasets.PascalVOCDetectionDataset.download">
    591. 

  591. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">download</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset.download"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.detection_datasets.PascalVOCDetectionDataset.download" title="Permalink to this definition"></a></dt>
    592. 

  592. <dd><p>Download Pascal dataset in XYXY_LABEL format.</p>
    593. 

  593. <p>Data extracted form <a class="reference external" href="http://host.robots.ox.ac.uk/pascal/VOC/">http://host.robots.ox.ac.uk/pascal/VOC/</a></p>
    594. 

  594. </dd></dl>
    595. 

  595. 

  596. </dd></dl>
    597. 

  597. 

  598. </section>
    599. 

  599. </section>
    600. 

  600. 

  601. 

  602.            </div>
    603. 

  603.           </div>
    604. 

  604.           <footer>
    605. 

  605. 

  606.   <hr/>
    607. 

  607. 

  608.   <div role="contentinfo">
    609. 

  609.     <p>&#169; Copyright 2021, SuperGradients team.</p>
    610. 

  610.   </div>
    611. 

  611. 

  612.   Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    613. 

  613.     <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    614. 

  614.     provided by <a href="https://readthedocs.org">Read the Docs</a>.
    615. 

  615.    
    616. 

  616. 

  617. </footer>
    618. 

  618.         </div>
    619. 

  619.       </div>
    620. 

  620.     </section>
    621. 

  621.   </div>
    622. 

  622.   <script>
    623. 

  623.       jQuery(function () {
    624. 

  624.           SphinxRtdTheme.Navigation.enable(true);
    625. 

  625.       });
    626. 

  626.   </script> 
    627. 

  627. 

  628. </body>
    629. 

  629. </html>
    630.
Discard
Prev
Next
Tip!

Press p or to see the previous file or, n or to see the next file