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

super_gradients.training.datasets.html 105 KB
History Raw

You have to be logged in to leave a comment. Sign In
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
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726 
  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 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.   <!--[if lt IE 9]>
    11. 

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

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

  13.   
    14. 

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

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

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

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

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

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

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

  21.     <link rel="next" title="super_gradients.training.datasets.classification_datasets package" href="super_gradients.training.datasets.classification_datasets.html" />
    22. 

  22.     <link rel="prev" title="super_gradients.training package" href="super_gradients.training.html" /> 
    23. 

  23. </head>
    24. 

  24. 

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

  26.   <div class="wy-grid-for-nav">
    27. 

  27.     <nav data-toggle="wy-nav-shift" class="wy-nav-side">
    28. 

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

  29.         <div class="wy-side-nav-search" >
    30. 

  30.             <a href="index.html" class="icon icon-home"> SuperGradients
    31. 

  31.           </a>
    32. 

  32. <div role="search">
    33. 

  33.   <form id="rtd-search-form" class="wy-form" action="search.html" method="get">
    34. 

  34.     <input type="text" name="q" placeholder="Search docs" />
    35. 

  35.     <input type="hidden" name="check_keywords" value="yes" />
    36. 

  36.     <input type="hidden" name="area" value="default" />
    37. 

  37.   </form>
    38. 

  38. </div>
    39. 

  39.         </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
    40. 

  40.               <p class="caption"><span class="caption-text">Contents:</span></p>
    41. 

  41. <ul class="current">
    42. 

  42. <li class="toctree-l1"><a class="reference internal" href="intro.html">Introduction</a></li>
    43. 

  43. <li class="toctree-l1"><a class="reference internal" href="intro.html#installation">Installation</a></li>
    44. 

  44. <li class="toctree-l1 current"><a class="reference internal" href="super_gradients.training.html">super_gradients.training package</a><ul class="current">
    45. 

  45. <li class="toctree-l2 current"><a class="reference internal" href="super_gradients.training.html#subpackages">Subpackages</a><ul class="current">
    46. 

  46. <li class="toctree-l3 current"><a class="current reference internal" href="#">super_gradients.training.datasets package</a><ul>
    47. 

  47. <li class="toctree-l4"><a class="reference internal" href="#subpackages">Subpackages</a></li>
    48. 

  48. <li class="toctree-l4"><a class="reference internal" href="#submodules">Submodules</a></li>
    49. 

  49. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.all_datasets">super_gradients.training.datasets.all_datasets module</a></li>
    50. 

  50. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.auto_augment">super_gradients.training.datasets.auto_augment module</a></li>
    51. 

  51. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.data_augmentation">super_gradients.training.datasets.data_augmentation module</a></li>
    52. 

  52. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.datasets_conf">super_gradients.training.datasets.datasets_conf module</a></li>
    53. 

  53. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.datasets_utils">super_gradients.training.datasets.datasets_utils module</a></li>
    54. 

  54. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.mixup">super_gradients.training.datasets.mixup module</a></li>
    55. 

  55. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets.sg_dataset">super_gradients.training.datasets.sg_dataset module</a></li>
    56. 

  56. <li class="toctree-l4"><a class="reference internal" href="#module-super_gradients.training.datasets">Module contents</a></li>
    57. 

  57. </ul>
    58. 

  58. </li>
    59. 

  59. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.exceptions.html">super_gradients.training.exceptions package</a></li>
    60. 

  60. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.legacy.html">super_gradients.training.legacy package</a></li>
    61. 

  61. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.losses.html">super_gradients.training.losses package</a></li>
    62. 

  62. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.metrics.html">super_gradients.training.metrics package</a></li>
    63. 

  63. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.models.html">super_gradients.training.models package</a></li>
    64. 

  64. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.sg_model.html">super_gradients.training.sg_model package</a></li>
    65. 

  65. <li class="toctree-l3"><a class="reference internal" href="super_gradients.training.utils.html">super_gradients.training.utils package</a></li>
    66. 

  66. </ul>
    67. 

  67. </li>
    68. 

  68. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.html#submodules">Submodules</a></li>
    69. 

  69. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.html#module-super_gradients.training.params">super_gradients.training.params module</a></li>
    70. 

  70. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.html#module-super_gradients.training.pretrained_models">super_gradients.training.pretrained_models module</a></li>
    71. 

  71. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.html#module-super_gradients.training">Module contents</a></li>
    72. 

  72. </ul>
    73. 

  73. </li>
    74. 

  74. </ul>
    75. 

  75. 

  76.         </div>
    77. 

  77.       </div>
    78. 

  78.     </nav>
    79. 

  79. 

  80.     <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
    81. 

  81.           <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
    82. 

  82.           <a href="index.html">SuperGradients</a>
    83. 

  83.       </nav>
    84. 

  84. 

  85.       <div class="wy-nav-content">
    86. 

  86.         <div class="rst-content">
    87. 

  87.           <div role="navigation" aria-label="Page navigation">
    88. 

  88.   <ul class="wy-breadcrumbs">
    89. 

  89.       <li><a href="index.html" class="icon icon-home"></a> &raquo;</li>
    90. 

  90.           <li><a href="super_gradients.training.html">super_gradients.training package</a> &raquo;</li>
    91. 

  91.       <li>super_gradients.training.datasets package</li>
    92. 

  92.       <li class="wy-breadcrumbs-aside">
    93. 

  93.             <a href="_sources/super_gradients.training.datasets.rst.txt" rel="nofollow"> View page source</a>
    94. 

  94.       </li>
    95. 

  95.   </ul>
    96. 

  96.   <hr/>
    97. 

  97. </div>
    98. 

  98.           <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
    99. 

  99.            <div itemprop="articleBody">
    100. 

  100.              
    101. 

  101.   <section id="super-gradients-training-datasets-package">
    102. 

  102. <h1>super_gradients.training.datasets package<a class="headerlink" href="#super-gradients-training-datasets-package" title="Permalink to this headline"></a></h1>
    103. 

  103. <section id="subpackages">
    104. 

  104. <h2>Subpackages<a class="headerlink" href="#subpackages" title="Permalink to this headline"></a></h2>
    105. 

  105. <div class="toctree-wrapper compound">
    106. 

  106. <ul>
    107. 

  107. <li class="toctree-l1"><a class="reference internal" href="super_gradients.training.datasets.classification_datasets.html">super_gradients.training.datasets.classification_datasets package</a><ul>
    108. 

  108. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.classification_datasets.html#module-super_gradients.training.datasets.classification_datasets">Module contents</a></li>
    109. 

  109. </ul>
    110. 

  110. </li>
    111. 

  111. <li class="toctree-l1"><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html">super_gradients.training.datasets.dataset_interfaces package</a><ul>
    112. 

  112. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#submodules">Submodules</a></li>
    113. 

  113. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#module-super_gradients.training.datasets.dataset_interfaces.dataset_interface">super_gradients.training.datasets.dataset_interfaces.dataset_interface module</a></li>
    114. 

  114. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#module-super_gradients.training.datasets.dataset_interfaces">Module contents</a></li>
    115. 

  115. </ul>
    116. 

  116. </li>
    117. 

  117. <li class="toctree-l1"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html">super_gradients.training.datasets.detection_datasets package</a><ul>
    118. 

  118. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#submodules">Submodules</a></li>
    119. 

  119. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#module-super_gradients.training.datasets.detection_datasets.coco_detection">super_gradients.training.datasets.detection_datasets.coco_detection module</a></li>
    120. 

  120. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#module-super_gradients.training.datasets.detection_datasets.detection_dataset">super_gradients.training.datasets.detection_datasets.detection_dataset module</a></li>
    121. 

  121. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#module-super_gradients.training.datasets.detection_datasets.pascal_voc_detection">super_gradients.training.datasets.detection_datasets.pascal_voc_detection module</a></li>
    122. 

  122. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#module-super_gradients.training.datasets.detection_datasets">Module contents</a></li>
    123. 

  123. </ul>
    124. 

  124. </li>
    125. 

  125. <li class="toctree-l1"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html">super_gradients.training.datasets.segmentation_datasets package</a><ul>
    126. 

  126. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#submodules">Submodules</a></li>
    127. 

  127. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets.cityscape_segmentation">super_gradients.training.datasets.segmentation_datasets.cityscape_segmentation module</a></li>
    128. 

  128. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets.coco_segmentation">super_gradients.training.datasets.segmentation_datasets.coco_segmentation module</a></li>
    129. 

  129. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets.pascal_aug_segmentation">super_gradients.training.datasets.segmentation_datasets.pascal_aug_segmentation module</a></li>
    130. 

  130. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets.pascal_voc_segmentation">super_gradients.training.datasets.segmentation_datasets.pascal_voc_segmentation module</a></li>
    131. 

  131. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets.segmentation_dataset">super_gradients.training.datasets.segmentation_datasets.segmentation_dataset module</a></li>
    132. 

  132. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#module-super_gradients.training.datasets.segmentation_datasets">Module contents</a></li>
    133. 

  133. </ul>
    134. 

  134. </li>
    135. 

  135. </ul>
    136. 

  136. </div>
    137. 

  137. </section>
    138. 

  138. <section id="submodules">
    139. 

  139. <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline"></a></h2>
    140. 

  140. </section>
    141. 

  141. <section id="module-super_gradients.training.datasets.all_datasets">
    142. 

  142. <span id="super-gradients-training-datasets-all-datasets-module"></span><h2>super_gradients.training.datasets.all_datasets module<a class="headerlink" href="#module-super_gradients.training.datasets.all_datasets" title="Permalink to this headline"></a></h2>
    143. 

  143. <dl class="py exception">
    144. 

  144. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.DataSetDoesNotExistException">
    145. 

  145. <em class="property"><span class="pre">exception</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.all_datasets.</span></span><span class="sig-name descname"><span class="pre">DataSetDoesNotExistException</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/all_datasets.html#DataSetDoesNotExistException"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.DataSetDoesNotExistException" title="Permalink to this definition"></a></dt>
    146. 

  146. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Exception</span></code></p>
    147. 

  147. <p>The requested dataset does not exist, or is not implemented.</p>
    148. 

  148. </dd></dl>
    149. 

  149. 

  150. <dl class="py class">
    151. 

  151. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets">
    152. 

  152. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.all_datasets.</span></span><span class="sig-name descname"><span class="pre">SgLibraryDatasets</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/all_datasets.html#SgLibraryDatasets"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets" title="Permalink to this definition"></a></dt>
    153. 

  153. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    154. 

  154. <p>Holds all of the different library dataset dictionaries, by DL Task mapping</p>
    155. 

  155. <blockquote>
    156. 

  156. <div><dl class="simple">
    157. 

  157. <dt>Attributes:</dt><dd><p>CLASSIFICATION          Dictionary of Classification Data sets
    158. 

  158. OBJECT_DETECTION        Dictionary of Object Detection Data sets
    159. 

  159. SEMANTIC_SEGMENTATION   Dictionary of Semantic Segmentation Data sets</p>
    160. 

  160. </dd>
    161. 

  161. </dl>
    162. 

  162. </div></blockquote>
    163. 

  163. <dl class="py attribute">
    164. 

  164. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.CLASSIFICATION">
    165. 

  165. <span class="sig-name descname"><span class="pre">CLASSIFICATION</span></span><em class="property"> <span class="pre">=</span> <span class="pre">{'cifar_10':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.Cifar10DatasetInterface'&gt;,</span> <span class="pre">'cifar_100':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.Cifar100DatasetInterface'&gt;,</span> <span class="pre">'classification_dataset':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.ClassificationDatasetInterface'&gt;,</span> <span class="pre">'imagenet':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.ImageNetDatasetInterface'&gt;,</span> <span class="pre">'library_dataset':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.LibraryDatasetInterface'&gt;,</span> <span class="pre">'test_dataset':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface'&gt;,</span> <span class="pre">'tiny_imagenet':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.TinyImageNetDatasetInterface'&gt;}</span></em><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.CLASSIFICATION" title="Permalink to this definition"></a></dt>
    166. 

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

  167. 

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

  169. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.OBJECT_DETECTION">
    170. 

  170. <span class="sig-name descname"><span class="pre">OBJECT_DETECTION</span></span><em class="property"> <span class="pre">=</span> <span class="pre">{'coco':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCoDetectionDatasetInterface'&gt;,</span> <span class="pre">'coco2014':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCo2014DetectionDatasetInterface'&gt;}</span></em><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.OBJECT_DETECTION" title="Permalink to this definition"></a></dt>
    171. 

  171. <dd></dd></dl>
    172. 

  172. 

  173. <dl class="py attribute">
    174. 

  174. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.SEMANTIC_SEGMENTATION">
    175. 

  175. <span class="sig-name descname"><span class="pre">SEMANTIC_SEGMENTATION</span></span><em class="property"> <span class="pre">=</span> <span class="pre">{'coco':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCoSegmentationDatasetInterface'&gt;,</span> <span class="pre">'pascal_aug':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.PascalAUG2012SegmentationDataSetInterface'&gt;,</span> <span class="pre">'pascal_voc':</span> <span class="pre">&lt;class</span> <span class="pre">'super_gradients.training.datasets.dataset_interfaces.dataset_interface.PascalVOC2012SegmentationDataSetInterface'&gt;}</span></em><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.SEMANTIC_SEGMENTATION" title="Permalink to this definition"></a></dt>
    176. 

  176. <dd></dd></dl>
    177. 

  177. 

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

  179. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.get_all_available_datasets">
    180. 

  180. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">get_all_available_datasets</span></span><span class="sig-paren">(</span><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">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><a class="reference internal" href="_modules/super_gradients/training/datasets/all_datasets.html#SgLibraryDatasets.get_all_available_datasets"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.get_all_available_datasets" title="Permalink to this definition"></a></dt>
    181. 

  181. <dd><p>Gets all the available datasets.</p>
    182. 

  182. </dd></dl>
    183. 

  183. 

  184. <dl class="py method">
    185. 

  185. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.get_dataset">
    186. 

  186. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">get_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dl_task</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">dataset_name</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> &#x2192; <span class="pre">Type</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></a><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/all_datasets.html#SgLibraryDatasets.get_dataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.get_dataset" title="Permalink to this definition"></a></dt>
    187. 

  187. <dd><p>Get’s a dataset with a given name for a given deep learning task.
    188. 

  188. examp:
    189. 

  189. &gt;&gt;&gt; SgLibraryDatasets.get_dataset(dl_task=’classification’, dataset_name=’cifar_100’)
    190. 

  190. &gt;&gt;&gt; &lt;Cifar100DatasetInterface instance&gt;</p>
    191. 

  191. </dd></dl>
    192. 

  192. 

  193. </dd></dl>
    194. 

  194. 

  195. </section>
    196. 

  196. <section id="module-super_gradients.training.datasets.auto_augment">
    197. 

  197. <span id="super-gradients-training-datasets-auto-augment-module"></span><h2>super_gradients.training.datasets.auto_augment module<a class="headerlink" href="#module-super_gradients.training.datasets.auto_augment" title="Permalink to this headline"></a></h2>
    198. 

  198. <p>RandAugment
    199. 

  199. RandAugment is a variant of AutoAugment which randomly selects transformations</p>
    200. 

  200. <blockquote>
    201. 

  201. <div><p>from AutoAugment to be applied on an image.</p>
    202. 

  202. </div></blockquote>
    203. 

  203. <dl class="simple">
    204. 

  204. <dt>RandomAugmentation Implementation adapted from:</dt><dd><p><a class="reference external" href="https://github.com/rwightman/pytorch-image-models/blob/master/timm/data/auto_augment.py">https://github.com/rwightman/pytorch-image-models/blob/master/timm/data/auto_augment.py</a></p>
    205. 

  205. </dd>
    206. 

  206. <dt>Papers:</dt><dd><p>RandAugment: Practical automated data augmentation… - <a class="reference external" href="https://arxiv.org/abs/1909.13719">https://arxiv.org/abs/1909.13719</a></p>
    207. 

  207. </dd>
    208. 

  208. </dl>
    209. 

  209. <dl class="py class">
    210. 

  210. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.AugmentOp">
    211. 

  211. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">AugmentOp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prob</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">magnitude</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hparams</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/auto_augment.html#AugmentOp"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.AugmentOp" title="Permalink to this definition"></a></dt>
    212. 

  212. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    213. 

  213. <p>single auto augment operations</p>
    214. 

  214. </dd></dl>
    215. 

  215. 

  216. <dl class="py class">
    217. 

  217. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.RandAugment">
    218. 

  218. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">RandAugment</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ops</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_layers</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">choice_weights</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/auto_augment.html#RandAugment"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.RandAugment" title="Permalink to this definition"></a></dt>
    219. 

  219. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    220. 

  220. <p>Random auto augment class, will select auto augment transforms according to probability weights for each op</p>
    221. 

  221. </dd></dl>
    222. 

  222. 

  223. <dl class="py function">
    224. 

  224. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.auto_contrast">
    225. 

  225. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">auto_contrast</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#auto_contrast"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.auto_contrast" title="Permalink to this definition"></a></dt>
    226. 

  226. <dd></dd></dl>
    227. 

  227. 

  228. <dl class="py function">
    229. 

  229. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.brightness">
    230. 

  230. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">brightness</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#brightness"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.brightness" title="Permalink to this definition"></a></dt>
    231. 

  231. <dd></dd></dl>
    232. 

  232. 

  233. <dl class="py function">
    234. 

  234. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.color">
    235. 

  235. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">color</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#color"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.color" title="Permalink to this definition"></a></dt>
    236. 

  236. <dd></dd></dl>
    237. 

  237. 

  238. <dl class="py function">
    239. 

  239. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.contrast">
    240. 

  240. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">contrast</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#contrast"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.contrast" title="Permalink to this definition"></a></dt>
    241. 

  241. <dd></dd></dl>
    242. 

  242. 

  243. <dl class="py function">
    244. 

  244. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.equalize">
    245. 

  245. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">equalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#equalize"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.equalize" title="Permalink to this definition"></a></dt>
    246. 

  246. <dd></dd></dl>
    247. 

  247. 

  248. <dl class="py function">
    249. 

  249. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.invert">
    250. 

  250. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">invert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#invert"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.invert" title="Permalink to this definition"></a></dt>
    251. 

  251. <dd></dd></dl>
    252. 

  252. 

  253. <dl class="py function">
    254. 

  254. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.posterize">
    255. 

  255. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">posterize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bits_to_keep</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#posterize"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.posterize" title="Permalink to this definition"></a></dt>
    256. 

  256. <dd></dd></dl>
    257. 

  257. 

  258. <dl class="py function">
    259. 

  259. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.rand_augment_ops">
    260. 

  260. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">rand_augment_ops</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">magnitude</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hparams</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="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/auto_augment.html#rand_augment_ops"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.rand_augment_ops" title="Permalink to this definition"></a></dt>
    261. 

  261. <dd></dd></dl>
    262. 

  262. 

  263. <dl class="py function">
    264. 

  264. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.rand_augment_transform">
    265. 

  265. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">rand_augment_transform</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config_str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hparams</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#rand_augment_transform"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.rand_augment_transform" title="Permalink to this definition"></a></dt>
    266. 

  266. <dd><p>Create a RandAugment transform</p>
    267. 

  267. <dl class="field-list simple">
    268. 

  268. <dt class="field-odd">Parameters</dt>
    269. 

  269. <dd class="field-odd"><p><strong>config_str</strong> – String defining configuration of random augmentation. Consists of multiple sections separated by</p>
    270. 

  270. </dd>
    271. 

  271. </dl>
    272. 

  272. <p>dashes (‘-‘). The first section defines the specific variant of rand augment (currently only ‘rand’). The remaining
    273. 

  273. sections, not order sepecific determine</p>
    274. 

  274. <blockquote>
    275. 

  275. <div><p>‘m’ - integer magnitude of rand augment
    276. 

  276. ‘n’ - integer num layers (number of transform ops selected per image)
    277. 

  277. ‘w’ - integer probabiliy weight index (index of a set of weights to influence choice of op)
    278. 

  278. ‘mstd’ -  float std deviation of magnitude noise applied
    279. 

  279. ‘inc’ - integer (bool), use augmentations that increase in severity with magnitude (default: 0)</p>
    280. 

  280. </div></blockquote>
    281. 

  281. <p>Ex ‘rand-m9-n3-mstd0.5’ results in RandAugment with magnitude 9, num_layers 3, magnitude_std 0.5
    282. 

  282. ‘rand-mstd1-w0’ results in magnitude_std 1.0, weights 0, default magnitude of 10 and num_layers 2</p>
    283. 

  283. <dl class="field-list simple">
    284. 

  284. <dt class="field-odd">Parameters</dt>
    285. 

  285. <dd class="field-odd"><p><strong>hparams</strong> – Other hparams (kwargs) for the RandAugmentation scheme</p>
    286. 

  286. </dd>
    287. 

  287. <dt class="field-even">Returns</dt>
    288. 

  288. <dd class="field-even"><p>A PyTorch compatible Transform</p>
    289. 

  289. </dd>
    290. 

  290. </dl>
    291. 

  291. </dd></dl>
    292. 

  292. 

  293. <dl class="py function">
    294. 

  294. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.rotate">
    295. 

  295. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">rotate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">degrees</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/auto_augment.html#rotate"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.rotate" title="Permalink to this definition"></a></dt>
    296. 

  296. <dd></dd></dl>
    297. 

  297. 

  298. <dl class="py function">
    299. 

  299. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.sharpness">
    300. 

  300. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">sharpness</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#sharpness"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.sharpness" title="Permalink to this definition"></a></dt>
    301. 

  301. <dd></dd></dl>
    302. 

  302. 

  303. <dl class="py function">
    304. 

  304. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.shear_x">
    305. 

  305. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">shear_x</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</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/auto_augment.html#shear_x"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.shear_x" title="Permalink to this definition"></a></dt>
    306. 

  306. <dd></dd></dl>
    307. 

  307. 

  308. <dl class="py function">
    309. 

  309. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.shear_y">
    310. 

  310. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">shear_y</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">factor</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/auto_augment.html#shear_y"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.shear_y" title="Permalink to this definition"></a></dt>
    311. 

  311. <dd></dd></dl>
    312. 

  312. 

  313. <dl class="py function">
    314. 

  314. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.solarize">
    315. 

  315. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">solarize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#solarize"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.solarize" title="Permalink to this definition"></a></dt>
    316. 

  316. <dd></dd></dl>
    317. 

  317. 

  318. <dl class="py function">
    319. 

  319. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.solarize_add">
    320. 

  320. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">solarize_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">add</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">__</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/auto_augment.html#solarize_add"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.solarize_add" title="Permalink to this definition"></a></dt>
    321. 

  321. <dd></dd></dl>
    322. 

  322. 

  323. <dl class="py function">
    324. 

  324. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_x_abs">
    325. 

  325. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">translate_x_abs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pixels</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/auto_augment.html#translate_x_abs"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.translate_x_abs" title="Permalink to this definition"></a></dt>
    326. 

  326. <dd></dd></dl>
    327. 

  327. 

  328. <dl class="py function">
    329. 

  329. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_x_rel">
    330. 

  330. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">translate_x_rel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pct</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/auto_augment.html#translate_x_rel"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.translate_x_rel" title="Permalink to this definition"></a></dt>
    331. 

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

  332. 

  333. <dl class="py function">
    334. 

  334. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_y_abs">
    335. 

  335. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">translate_y_abs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pixels</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/auto_augment.html#translate_y_abs"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.translate_y_abs" title="Permalink to this definition"></a></dt>
    336. 

  336. <dd></dd></dl>
    337. 

  337. 

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

  339. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_y_rel">
    340. 

  340. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.auto_augment.</span></span><span class="sig-name descname"><span class="pre">translate_y_rel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pct</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/auto_augment.html#translate_y_rel"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.auto_augment.translate_y_rel" title="Permalink to this definition"></a></dt>
    341. 

  341. <dd></dd></dl>
    342. 

  342. 

  343. </section>
    344. 

  344. <section id="module-super_gradients.training.datasets.data_augmentation">
    345. 

  345. <span id="super-gradients-training-datasets-data-augmentation-module"></span><h2>super_gradients.training.datasets.data_augmentation module<a class="headerlink" href="#module-super_gradients.training.datasets.data_augmentation" title="Permalink to this headline"></a></h2>
    346. 

  346. <dl class="py class">
    347. 

  347. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation">
    348. 

  348. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.data_augmentation.</span></span><span class="sig-name descname"><span class="pre">DataAugmentation</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/data_augmentation.html#DataAugmentation"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.DataAugmentation" title="Permalink to this definition"></a></dt>
    349. 

  349. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    350. 

  350. <dl class="py method">
    351. 

  351. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.cutout">
    352. 

  352. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">cutout</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mask_size</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p</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">cutout_inside</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">mask_color</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0,</span> <span class="pre">0,</span> <span class="pre">0)</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/data_augmentation.html#DataAugmentation.cutout"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.DataAugmentation.cutout" title="Permalink to this definition"></a></dt>
    353. 

  353. <dd></dd></dl>
    354. 

  354. 

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

  356. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.normalize">
    357. 

  357. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">normalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mean</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">std</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/data_augmentation.html#DataAugmentation.normalize"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.DataAugmentation.normalize" title="Permalink to this definition"></a></dt>
    358. 

  358. <dd></dd></dl>
    359. 

  359. 

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

  361. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.to_tensor">
    362. 

  362. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">to_tensor</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/data_augmentation.html#DataAugmentation.to_tensor"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.DataAugmentation.to_tensor" title="Permalink to this definition"></a></dt>
    363. 

  363. <dd></dd></dl>
    364. 

  364. 

  365. </dd></dl>
    366. 

  366. 

  367. <dl class="py class">
    368. 

  368. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.Lighting">
    369. 

  369. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.data_augmentation.</span></span><span class="sig-name descname"><span class="pre">Lighting</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alphastd</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eigval</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">tensor([0.2175,</span> <span class="pre">0.0188,</span> <span class="pre">0.0045])</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eigvec</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">tensor([[-</span> <span class="pre">0.5675,</span> <span class="pre">0.7192,</span> <span class="pre">0.4009],</span> <span class="pre">[-</span> <span class="pre">0.5808,</span> <span class="pre">-</span> <span class="pre">0.0045,</span> <span class="pre">-</span> <span class="pre">0.814],</span> <span class="pre">[-</span> <span class="pre">0.5836,</span> <span class="pre">-</span> <span class="pre">0.6948,</span> <span class="pre">0.4203]])</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/data_augmentation.html#Lighting"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.Lighting" title="Permalink to this definition"></a></dt>
    370. 

  370. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    371. 

  371. <p>Lighting noise(AlexNet - style PCA - based noise)
    372. 

  372. Taken from fastai Imagenet training - <a class="reference external" href="https://github.com/fastai/imagenet-fast/blob/faa0f9dfc9e8e058ffd07a248724bf384f526fae/imagenet_nv/fastai_imagenet.py#L103">https://github.com/fastai/imagenet-fast/blob/faa0f9dfc9e8e058ffd07a248724bf384f526fae/imagenet_nv/fastai_imagenet.py#L103</a>
    373. 

  373. To use:</p>
    374. 

  374. <blockquote>
    375. 

  375. <div><ul class="simple">
    376. 

  376. <li><p>training_params = {“imagenet_pca_aug”: 0.1}</p></li>
    377. 

  377. <li><p>Default training_params arg is 0.0 (“don’t use”)</p></li>
    378. 

  378. <li><p>0.1 is that default in the original paper</p></li>
    379. 

  379. </ul>
    380. 

  380. </div></blockquote>
    381. 

  381. </dd></dl>
    382. 

  382. 

  383. <dl class="py class">
    384. 

  384. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.RandomErase">
    385. 

  385. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.data_augmentation.</span></span><span class="sig-name descname"><span class="pre">RandomErase</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">probability</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</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/data_augmentation.html#RandomErase"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.RandomErase" title="Permalink to this definition"></a></dt>
    386. 

  386. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchvision.transforms.transforms.RandomErasing</span></code></p>
    387. 

  387. <p>A simple class that translates the parameters supported in SuperGradient’s code base</p>
    388. 

  388. <dl class="py attribute">
    389. 

  389. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.RandomErase.training">
    390. 

  390. <span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.datasets.data_augmentation.RandomErase.training" title="Permalink to this definition"></a></dt>
    391. 

  391. <dd></dd></dl>
    392. 

  392. 

  393. </dd></dl>
    394. 

  394. 

  395. </section>
    396. 

  396. <section id="module-super_gradients.training.datasets.datasets_conf">
    397. 

  397. <span id="super-gradients-training-datasets-datasets-conf-module"></span><h2>super_gradients.training.datasets.datasets_conf module<a class="headerlink" href="#module-super_gradients.training.datasets.datasets_conf" title="Permalink to this headline"></a></h2>
    398. 

  398. </section>
    399. 

  399. <section id="module-super_gradients.training.datasets.datasets_utils">
    400. 

  400. <span id="super-gradients-training-datasets-datasets-utils-module"></span><h2>super_gradients.training.datasets.datasets_utils module<a class="headerlink" href="#module-super_gradients.training.datasets.datasets_utils" title="Permalink to this headline"></a></h2>
    401. 

  401. <dl class="py class">
    402. 

  402. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.AbstractCollateFunction">
    403. 

  403. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">AbstractCollateFunction</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#AbstractCollateFunction"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.AbstractCollateFunction" title="Permalink to this definition"></a></dt>
    404. 

  404. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">abc.ABC</span></code></p>
    405. 

  405. <p>A collate function (for torch DataLoader)</p>
    406. 

  406. </dd></dl>
    407. 

  407. 

  408. <dl class="py class">
    409. 

  409. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.AtomicInteger">
    410. 

  410. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">AtomicInteger</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</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">0</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#AtomicInteger"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.AtomicInteger" title="Permalink to this definition"></a></dt>
    411. 

  411. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    412. 

  412. </dd></dl>
    413. 

  413. 

  414. <dl class="py class">
    415. 

  415. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.ComposedCollateFunction">
    416. 

  416. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">ComposedCollateFunction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">functions</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">list</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#ComposedCollateFunction"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.ComposedCollateFunction" title="Permalink to this definition"></a></dt>
    417. 

  417. <dd><p>Bases: <a class="reference internal" href="#super_gradients.training.datasets.datasets_utils.AbstractCollateFunction" title="super_gradients.training.datasets.datasets_utils.AbstractCollateFunction"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.datasets_utils.AbstractCollateFunction</span></code></a></p>
    418. 

  418. <p>A function (for torch DataLoader) which executes a sequence of sub collate functions</p>
    419. 

  419. </dd></dl>
    420. 

  420. 

  421. <dl class="py class">
    422. 

  422. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger">
    423. 

  423. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">DatasetStatisticsTensorboardLogger</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">writer</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.utils.tensorboard.writer.SummaryWriter</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">summary_params</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">dict</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">{'max_batches':</span> <span class="pre">30,</span> <span class="pre">'plot_anchors_coverage':</span> <span class="pre">True,</span> <span class="pre">'plot_box_size_distribution':</span> <span class="pre">True,</span> <span class="pre">'plot_class_distribution':</span> <span class="pre">True,</span> <span class="pre">'sample_images':</span> <span class="pre">32}</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#DatasetStatisticsTensorboardLogger"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger" title="Permalink to this definition"></a></dt>
    424. 

  424. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    425. 

  425. <dl class="py attribute">
    426. 

  426. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.DEFAULT_SUMMARY_PARAMS">
    427. 

  427. <span class="sig-name descname"><span class="pre">DEFAULT_SUMMARY_PARAMS</span></span><em class="property"> <span class="pre">=</span> <span class="pre">{'max_batches':</span> <span class="pre">30,</span> <span class="pre">'plot_anchors_coverage':</span> <span class="pre">True,</span> <span class="pre">'plot_box_size_distribution':</span> <span class="pre">True,</span> <span class="pre">'plot_class_distribution':</span> <span class="pre">True,</span> <span class="pre">'sample_images':</span> <span class="pre">32}</span></em><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.DEFAULT_SUMMARY_PARAMS" title="Permalink to this definition"></a></dt>
    428. 

  428. <dd></dd></dl>
    429. 

  429. 

  430. <dl class="py method">
    431. 

  431. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.analyze">
    432. 

  432. <span class="sig-name descname"><span class="pre">analyze</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_loader</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.utils.data.dataloader.DataLoader</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">title</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">anchors</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> <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/datasets_utils.html#DatasetStatisticsTensorboardLogger.analyze"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.analyze" title="Permalink to this definition"></a></dt>
    433. 

  433. <dd><dl class="field-list simple">
    434. 

  434. <dt class="field-odd">Parameters</dt>
    435. 

  435. <dd class="field-odd"><ul class="simple">
    436. 

  436. <li><p><strong>data_loader</strong> – the dataset data loader</p></li>
    437. 

  437. <li><p><strong>dataset_params</strong> – the dataset parameters</p></li>
    438. 

  438. <li><p><strong>title</strong> – the title for this dataset (i.e. Coco 2017 test set)</p></li>
    439. 

  439. <li><p><strong>anchors</strong> – the list of anchors used by the model. applicable only for detection datasets</p></li>
    440. 

  440. </ul>
    441. 

  441. </dd>
    442. 

  442. </dl>
    443. 

  443. </dd></dl>
    444. 

  444. 

  445. <dl class="py attribute">
    446. 

  446. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.logger">
    447. 

  447. <span class="sig-name descname"><span class="pre">logger</span></span><em class="property"> <span class="pre">=</span> <span class="pre">&lt;Logger</span> <span class="pre">super_gradients.training.datasets.datasets_utils</span> <span class="pre">(INFO)&gt;</span></em><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.logger" title="Permalink to this definition"></a></dt>
    448. 

  448. <dd></dd></dl>
    449. 

  449. 

  450. </dd></dl>
    451. 

  451. 

  452. <dl class="py class">
    453. 

  453. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.MultiScaleCollateFunction">
    454. 

  454. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">MultiScaleCollateFunction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">target_size</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">min_image_size</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">max_image_size</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">image_size_steps</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">32</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">change_frequency</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">10</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#MultiScaleCollateFunction"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.MultiScaleCollateFunction" title="Permalink to this definition"></a></dt>
    455. 

  455. <dd><p>Bases: <a class="reference internal" href="#super_gradients.training.datasets.datasets_utils.AbstractCollateFunction" title="super_gradients.training.datasets.datasets_utils.AbstractCollateFunction"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.datasets_utils.AbstractCollateFunction</span></code></a></p>
    456. 

  456. <p>a collate function to implement multi-scale data augmentation
    457. 

  457. according to <a class="reference external" href="https://arxiv.org/pdf/1612.08242.pdf">https://arxiv.org/pdf/1612.08242.pdf</a></p>
    458. 

  458. </dd></dl>
    459. 

  459. 

  460. <dl class="py class">
    461. 

  461. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation">
    462. 

  462. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">RandomResizedCropAndInterpolation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.08,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.75,</span> <span class="pre">1.3333333333333333)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interpolation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'default'</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#RandomResizedCropAndInterpolation"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation" title="Permalink to this definition"></a></dt>
    463. 

  463. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchvision.transforms.transforms.RandomResizedCrop</span></code></p>
    464. 

  464. <p>Crop the given PIL Image to random size and aspect ratio with explicitly chosen or random interpolation.</p>
    465. 

  465. <p>A crop of random size (default: of 0.08 to 1.0) of the original size and a random
    466. 

  466. aspect ratio (default: of 3/4 to 4/3) of the original aspect ratio is made. This crop
    467. 

  467. is finally resized to given size.
    468. 

  468. This is popularly used to train the Inception networks.</p>
    469. 

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

  470. <dt>Args:</dt><dd><p>size: expected output size of each edge
    471. 

  471. scale: range of size of the origin size cropped
    472. 

  472. ratio: range of aspect ratio of the origin aspect ratio cropped
    473. 

  473. interpolation: Default: PIL.Image.BILINEAR</p>
    474. 

  474. </dd>
    475. 

  475. </dl>
    476. 

  476. <dl class="py method">
    477. 

  477. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.forward">
    478. 

  478. <span class="sig-name descname"><span class="pre">forward</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#RandomResizedCropAndInterpolation.forward"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.forward" title="Permalink to this definition"></a></dt>
    479. 

  479. <dd><dl class="simple">
    480. 

  480. <dt>Args:</dt><dd><p>img (PIL Image): Image to be cropped and resized.</p>
    481. 

  481. </dd>
    482. 

  482. <dt>Returns:</dt><dd><p>PIL Image: Randomly cropped and resized image.</p>
    483. 

  483. </dd>
    484. 

  484. </dl>
    485. 

  485. </dd></dl>
    486. 

  486. 

  487. <dl class="py attribute">
    488. 

  488. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.training">
    489. 

  489. <span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.training" title="Permalink to this definition"></a></dt>
    490. 

  490. <dd></dd></dl>
    491. 

  491. 

  492. </dd></dl>
    493. 

  493. 

  494. <dl class="py function">
    495. 

  495. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.get_color_augmentation">
    496. 

  496. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">get_color_augmentation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">rand_augment_config_string</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">color_jitter</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">crop_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">224</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_mean</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">[0.485,</span> <span class="pre">0.456,</span> <span class="pre">0.406]</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#get_color_augmentation"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.get_color_augmentation" title="Permalink to this definition"></a></dt>
    497. 

  497. <dd><p>Returns color augmentation class. As these augmentation cannot work on top one another, only one is returned according to rand_augment_config_string
    498. 

  498. :param rand_augment_config_string: string which defines the auto augment configurations. If none, color jitter will be returned. For possibile values see auto_augment.py
    499. 

  499. :param color_jitter: tuple for color jitter value.
    500. 

  500. :param crop_size: relevant only for auto augment
    501. 

  501. :param img_mean: relevant only for auto augment
    502. 

  502. :return: RandAugment transform or ColorJitter</p>
    503. 

  503. </dd></dl>
    504. 

  504. 

  505. <dl class="py function">
    506. 

  506. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.get_mean_and_std">
    507. 

  507. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">get_mean_and_std</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#get_mean_and_std"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.get_mean_and_std" title="Permalink to this definition"></a></dt>
    508. 

  508. <dd><p>Compute the mean and std value of dataset.</p>
    509. 

  509. </dd></dl>
    510. 

  510. 

  511. <dl class="py function">
    512. 

  512. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.get_mean_and_std_torch">
    513. 

  513. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">get_mean_and_std_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</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">dataloader</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">num_workers</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">RandomResizeSize</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">224</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#get_mean_and_std_torch"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.get_mean_and_std_torch" title="Permalink to this definition"></a></dt>
    514. 

  514. <dd><p>A function for getting the mean and std of large datasets using pytorch dataloader and gpu functionality.</p>
    515. 

  515. <dl class="field-list simple">
    516. 

  516. <dt class="field-odd">Parameters</dt>
    517. 

  517. <dd class="field-odd"><ul class="simple">
    518. 

  518. <li><p><strong>data_dir</strong> – String, path to none-library dataset folder. For example “/data/Imagenette” or “/data/TinyImagenet”</p></li>
    519. 

  519. <li><p><strong>dataloader</strong> – a torch DataLoader, as it would feed the data into the trainer (including transforms etc).</p></li>
    520. 

  520. <li><p><strong>RandomResizeSize</strong> – Int, the size of the RandomResizeCrop as it appears in the DataInterface (for example, for Imagenet,</p></li>
    521. 

  521. </ul>
    522. 

  522. </dd>
    523. 

  523. </dl>
    524. 

  524. <p>this value should be 224).
    525. 

  525. :return: 2 lists,mean and std, each one of len 3 (1 for each channel)</p>
    526. 

  526. </dd></dl>
    527. 

  527. 

  528. </section>
    529. 

  529. <section id="module-super_gradients.training.datasets.mixup">
    530. 

  530. <span id="super-gradients-training-datasets-mixup-module"></span><h2>super_gradients.training.datasets.mixup module<a class="headerlink" href="#module-super_gradients.training.datasets.mixup" title="Permalink to this headline"></a></h2>
    531. 

  531. <p>Mixup and Cutmix</p>
    532. 

  532. <p>Papers:
    533. 

  533. mixup: Beyond Empirical Risk Minimization (<a class="reference external" href="https://arxiv.org/abs/1710.09412">https://arxiv.org/abs/1710.09412</a>)</p>
    534. 

  534. <p>CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features (<a class="reference external" href="https://arxiv.org/abs/1905.04899">https://arxiv.org/abs/1905.04899</a>)</p>
    535. 

  535. <p>Code Reference:
    536. 

  536. CutMix: <a class="reference external" href="https://github.com/clovaai/CutMix-PyTorch">https://github.com/clovaai/CutMix-PyTorch</a>
    537. 

  537. CutMix by timm: <a class="reference external" href="https://github.com/rwightman/pytorch-image-models/timm">https://github.com/rwightman/pytorch-image-models/timm</a></p>
    538. 

  538. <dl class="py class">
    539. 

  539. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.CollateMixup">
    540. 

  540. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">CollateMixup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mixup_alpha</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cutmix_alpha</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cutmix_minmax</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">float</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">prob</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">switch_prob</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mode</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">'batch'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">correct_lam</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">label_smoothing</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_classes</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">1000</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#CollateMixup"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.CollateMixup" title="Permalink to this definition"></a></dt>
    541. 

  541. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    542. 

  542. <p>Collate with Mixup/Cutmix that applies different params to each element or whole batch
    543. 

  543. A Mixup impl that’s performed while collating the batches.</p>
    544. 

  544. </dd></dl>
    545. 

  545. 

  546. <dl class="py function">
    547. 

  547. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.cutmix_bbox_and_lam">
    548. 

  548. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">cutmix_bbox_and_lam</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_shape</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">lam</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ratio_minmax</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">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">tuple</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</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">correct_lam</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">count</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><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#cutmix_bbox_and_lam"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.cutmix_bbox_and_lam" title="Permalink to this definition"></a></dt>
    549. 

  549. <dd><p>Generate bbox and apply lambda correction.</p>
    550. 

  550. </dd></dl>
    551. 

  551. 

  552. <dl class="py function">
    553. 

  553. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.mixup_target">
    554. 

  554. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">mixup_target</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_classes</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">lam</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">smoothing</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</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">'cuda'</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#mixup_target"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.mixup_target" title="Permalink to this definition"></a></dt>
    555. 

  555. <dd><p>generate a smooth target (label) two-hot tensor to support the mixed images with different labels
    556. 

  556. :param target: the targets tensor
    557. 

  557. :param num_classes: number of classes (to set the final tensor size)
    558. 

  558. :param lam: percentage of label a range [0, 1] in the mixing
    559. 

  559. :param smoothing: the smoothing multiplier
    560. 

  560. :param device: usable device [‘cuda’, ‘cpu’]
    561. 

  561. :return:</p>
    562. 

  562. </dd></dl>
    563. 

  563. 

  564. <dl class="py function">
    565. 

  565. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.one_hot">
    566. 

  566. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">one_hot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_classes</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">on_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">off_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'cuda'</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#one_hot"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.one_hot" title="Permalink to this definition"></a></dt>
    567. 

  567. <dd></dd></dl>
    568. 

  568. 

  569. <dl class="py function">
    570. 

  570. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.rand_bbox">
    571. 

  571. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">rand_bbox</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_shape</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">lam</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">margin</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.0</span></span></em>, <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">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><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#rand_bbox"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.rand_bbox" title="Permalink to this definition"></a></dt>
    572. 

  572. <dd><p>Standard CutMix bounding-box
    573. 

  573. Generates a random square bbox based on lambda value. This impl includes
    574. 

  574. support for enforcing a border margin as percent of bbox dimensions.</p>
    575. 

  575. <dl class="field-list simple">
    576. 

  576. <dt class="field-odd">Parameters</dt>
    577. 

  577. <dd class="field-odd"><ul class="simple">
    578. 

  578. <li><p><strong>img_shape</strong> – Image shape as tuple</p></li>
    579. 

  579. <li><p><strong>lam</strong> – Cutmix lambda value</p></li>
    580. 

  580. <li><p><strong>margin</strong> – Percentage of bbox dimension to enforce as margin (reduce amount of box outside image)</p></li>
    581. 

  581. <li><p><strong>count</strong> – Number of bbox to generate</p></li>
    582. 

  582. </ul>
    583. 

  583. </dd>
    584. 

  584. </dl>
    585. 

  585. </dd></dl>
    586. 

  586. 

  587. <dl class="py function">
    588. 

  588. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.rand_bbox_minmax">
    589. 

  589. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.mixup.</span></span><span class="sig-name descname"><span class="pre">rand_bbox_minmax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_shape</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">minmax</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">tuple</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span></em>, <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">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><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/mixup.html#rand_bbox_minmax"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.mixup.rand_bbox_minmax" title="Permalink to this definition"></a></dt>
    590. 

  590. <dd><p>Min-Max CutMix bounding-box
    591. 

  591. Inspired by Darknet cutmix impl, generates a random rectangular bbox
    592. 

  592. based on min/max percent values applied to each dimension of the input image.</p>
    593. 

  593. <p>Typical defaults for minmax are usually in the  .2-.3 for min and .8-.9 range for max.</p>
    594. 

  594. <dl class="field-list simple">
    595. 

  595. <dt class="field-odd">Parameters</dt>
    596. 

  596. <dd class="field-odd"><ul class="simple">
    597. 

  597. <li><p><strong>img_shape</strong> – Image shape as tuple</p></li>
    598. 

  598. <li><p><strong>minmax</strong> – Min and max bbox ratios (as percent of image size)</p></li>
    599. 

  599. <li><p><strong>count</strong> – Number of bbox to generate</p></li>
    600. 

  600. </ul>
    601. 

  601. </dd>
    602. 

  602. </dl>
    603. 

  603. </dd></dl>
    604. 

  604. 

  605. </section>
    606. 

  606. <section id="module-super_gradients.training.datasets.sg_dataset">
    607. 

  607. <span id="super-gradients-training-datasets-sg-dataset-module"></span><h2>super_gradients.training.datasets.sg_dataset module<a class="headerlink" href="#module-super_gradients.training.datasets.sg_dataset" title="Permalink to this headline"></a></h2>
    608. 

  608. <dl class="py class">
    609. 

  609. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset">
    610. 

  610. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.sg_dataset.</span></span><span class="sig-name descname"><span class="pre">BaseSgVisionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">root:</span> <span class="pre">str</span></em>, <em class="sig-param"><span class="pre">sample_loader:</span> <span class="pre">Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">default_loader&gt;</span></em>, <em class="sig-param"><span class="pre">target_loader:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">collate_fn:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">valid_sample_extensions:</span> <span class="pre">tuple</span> <span class="pre">=</span> <span class="pre">('.jpg'</span></em>, <em class="sig-param"><span class="pre">'.jpeg'</span></em>, <em class="sig-param"><span class="pre">'.png'</span></em>, <em class="sig-param"><span class="pre">'.ppm'</span></em>, <em class="sig-param"><span class="pre">'.bmp'</span></em>, <em class="sig-param"><span class="pre">'.pgm'</span></em>, <em class="sig-param"><span class="pre">'.tif'</span></em>, <em class="sig-param"><span class="pre">'.tiff'</span></em>, <em class="sig-param"><span class="pre">'.webp')</span></em>, <em class="sig-param"><span class="pre">sample_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">target_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/sg_dataset.html#BaseSgVisionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset" title="Permalink to this definition"></a></dt>
    611. 

  611. <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>
    612. 

  612. <dl class="py method">
    613. 

  613. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.numpy_loader_func">
    614. 

  614. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">numpy_loader_func</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/sg_dataset.html#BaseSgVisionDataset.numpy_loader_func"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.numpy_loader_func" title="Permalink to this definition"></a></dt>
    615. 

  615. <dd><dl class="simple">
    616. 

  616. <dt>_numpy_loader_func - Uses numpy load func</dt><dd><dl class="field-list simple">
    617. 

  617. <dt class="field-odd">param path</dt>
    618. 

  618. <dd class="field-odd"><p></p></dd>
    619. 

  619. <dt class="field-even">return</dt>
    620. 

  620. <dd class="field-even"><p></p></dd>
    621. 

  621. </dl>
    622. 

  622. </dd>
    623. 

  623. </dl>
    624. 

  624. </dd></dl>
    625. 

  625. 

  626. <dl class="py method">
    627. 

  627. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.text_file_loader_func">
    628. 

  628. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">text_file_loader_func</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">text_file_path</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">inline_splitter</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">'</span> <span class="pre">'</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">list</span><a class="reference internal" href="_modules/super_gradients/training/datasets/sg_dataset.html#BaseSgVisionDataset.text_file_loader_func"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.text_file_loader_func" title="Permalink to this definition"></a></dt>
    629. 

  629. <dd><blockquote>
    630. 

  630. <div><dl class="simple">
    631. 

  631. <dt>text_file_loader_func - Uses a line by line based code to get vectorized data from a text-based file</dt><dd><dl class="field-list simple">
    632. 

  632. <dt class="field-odd">param text_file_path</dt>
    633. 

  633. <dd class="field-odd"><p>Input text file</p>
    634. 

  634. </dd>
    635. 

  635. <dt class="field-even">param inline_splitter</dt>
    636. 

  636. <dd class="field-even"><p>The char to use in order to separate between different VALUES of the SAME vector
    637. 

  637. please notice that DIFFERENT VECTORS SHOULD BE IN SEPARATE LINES (’</p>
    638. 

  638. </dd>
    639. 

  639. </dl>
    640. 

  640. </dd>
    641. 

  641. </dl>
    642. 

  642. </div></blockquote>
    643. 

  643. <dl class="simple">
    644. 

  644. <dt>‘) SEPARATED</dt><dd><dl class="field-list simple">
    645. 

  645. <dt class="field-odd">return</dt>
    646. 

  646. <dd class="field-odd"><p>a list of tuples, where each tuple is a vector of target values</p>
    647. 

  647. </dd>
    648. 

  648. </dl>
    649. 

  649. </dd>
    650. 

  650. </dl>
    651. 

  651. </dd></dl>
    652. 

  652. 

  653. </dd></dl>
    654. 

  654. 

  655. <dl class="py class">
    656. 

  656. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.DirectoryDataSet">
    657. 

  657. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.sg_dataset.</span></span><span class="sig-name descname"><span class="pre">DirectoryDataSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">root:</span> <span class="pre">str</span></em>, <em class="sig-param"><span class="pre">samples_sub_directory:</span> <span class="pre">str</span></em>, <em class="sig-param"><span class="pre">targets_sub_directory:</span> <span class="pre">str</span></em>, <em class="sig-param"><span class="pre">target_extension:</span> <span class="pre">str</span></em>, <em class="sig-param"><span class="pre">sample_loader:</span> <span class="pre">Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">default_loader&gt;</span></em>, <em class="sig-param"><span class="pre">target_loader:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">collate_fn:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">sample_extensions:</span> <span class="pre">tuple</span> <span class="pre">=</span> <span class="pre">('.jpg'</span></em>, <em class="sig-param"><span class="pre">'.jpeg'</span></em>, <em class="sig-param"><span class="pre">'.png'</span></em>, <em class="sig-param"><span class="pre">'.ppm'</span></em>, <em class="sig-param"><span class="pre">'.bmp'</span></em>, <em class="sig-param"><span class="pre">'.pgm'</span></em>, <em class="sig-param"><span class="pre">'.tif'</span></em>, <em class="sig-param"><span class="pre">'.tiff'</span></em>, <em class="sig-param"><span class="pre">'.webp')</span></em>, <em class="sig-param"><span class="pre">sample_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">target_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/sg_dataset.html#DirectoryDataSet"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.sg_dataset.DirectoryDataSet" title="Permalink to this definition"></a></dt>
    658. 

  658. <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>
    659. 

  659. <dl class="simple">
    660. 

  660. <dt>DirectoryDataSet - A PyTorch Vision Data Set extension that receives a root Dir and two separate sub directories:</dt><dd><ul class="simple">
    661. 

  661. <li><p>Sub-Directory for Samples</p></li>
    662. 

  662. <li><p>Sub-Directory for Targets</p></li>
    663. 

  663. </ul>
    664. 

  664. </dd>
    665. 

  665. </dl>
    666. 

  666. </dd></dl>
    667. 

  667. 

  668. <dl class="py class">
    669. 

  669. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.ListDataset">
    670. 

  670. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.sg_dataset.</span></span><span class="sig-name descname"><span class="pre">ListDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">root</span></em>, <em class="sig-param"><span class="pre">file</span></em>, <em class="sig-param"><span class="pre">sample_loader:</span> <span class="pre">Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">default_loader&gt;</span></em>, <em class="sig-param"><span class="pre">target_loader:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">collate_fn:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">sample_extensions:</span> <span class="pre">tuple</span> <span class="pre">=</span> <span class="pre">('.jpg'</span></em>, <em class="sig-param"><span class="pre">'.jpeg'</span></em>, <em class="sig-param"><span class="pre">'.png'</span></em>, <em class="sig-param"><span class="pre">'.ppm'</span></em>, <em class="sig-param"><span class="pre">'.bmp'</span></em>, <em class="sig-param"><span class="pre">'.pgm'</span></em>, <em class="sig-param"><span class="pre">'.tif'</span></em>, <em class="sig-param"><span class="pre">'.tiff'</span></em>, <em class="sig-param"><span class="pre">'.webp')</span></em>, <em class="sig-param"><span class="pre">sample_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">target_transform:</span> <span class="pre">Optional[Callable]</span> <span class="pre">=</span> <span class="pre">None</span></em>, <em class="sig-param"><span class="pre">target_extension='.npy'</span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/sg_dataset.html#ListDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.sg_dataset.ListDataset" title="Permalink to this definition"></a></dt>
    671. 

  671. <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>
    672. 

  672. <dl>
    673. 

  673. <dt>ListDataset - A PyTorch Vision Data Set extension that receives a file with FULL PATH to each of the samples.</dt><dd><p>Then, the assumption is that for every sample, there is a * matching target * in the same
    674. 

  674. path but with a different extension, i.e:</p>
    675. 

  675. <blockquote>
    676. 

  676. <div><dl class="simple">
    677. 

  677. <dt>for the samples paths:  (That appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.png
    678. 

  678. /root/dataset/class_y/sample123.png</p>
    679. 

  679. </dd>
    680. 

  680. <dt>the matching labels paths:  (That DO NOT appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.ext
    681. 

  681. /root/dataset/class_y/sample123.ext</p>
    682. 

  682. </dd>
    683. 

  683. </dl>
    684. 

  684. </div></blockquote>
    685. 

  685. </dd>
    686. 

  686. </dl>
    687. 

  687. </dd></dl>
    688. 

  688. 

  689. </section>
    690. 

  690. <section id="module-super_gradients.training.datasets">
    691. 

  691. <span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-super_gradients.training.datasets" title="Permalink to this headline"></a></h2>
    692. 

  692. </section>
    693. 

  693. </section>
    694. 

  694. 

  695. 

  696.            </div>
    697. 

  697.           </div>
    698. 

  698.           <footer><div class="rst-footer-buttons" role="navigation" aria-label="Footer">
    699. 

  699.         <a href="super_gradients.training.html" class="btn btn-neutral float-left" title="super_gradients.training package" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left" aria-hidden="true"></span> Previous</a>
    700. 

  700.         <a href="super_gradients.training.datasets.classification_datasets.html" class="btn btn-neutral float-right" title="super_gradients.training.datasets.classification_datasets package" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right" aria-hidden="true"></span></a>
    701. 

  701.     </div>
    702. 

  702. 

  703.   <hr/>
    704. 

  704. 

  705.   <div role="contentinfo">
    706. 

  706.     <p>&#169; Copyright 2021, SuperGradients team.</p>
    707. 

  707.   </div>
    708. 

  708. 

  709.   Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    710. 

  710.     <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    711. 

  711.     provided by <a href="https://readthedocs.org">Read the Docs</a>.
    712. 

  712.    
    713. 

  713. 

  714. </footer>
    715. 

  715.         </div>
    716. 

  716.       </div>
    717. 

  717.     </section>
    718. 

  718.   </div>
    719. 

  719.   <script>
    720. 

  720.       jQuery(function () {
    721. 

  721.           SphinxRtdTheme.Navigation.enable(true);
    722. 

  722.       });
    723. 

  723.   </script> 
    724. 

  724. 

  725. </body>
    726. 

  726. </html>
    727.
Prev
Next
Tip!

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

Comments

Write Preview
Loading...
Integrate AWS S3
Use S3 remote
Select bucket
Access key
Finish

Use AWS S3 as storage!

Browsing data directories saved to S3 is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your S3 bucket

Access key (If needed)

Congratulations!

super-gradients is now integrated with AWS S3!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate Google Cloud Storage
Use Google Storage
Select bucket
Upload key
Finish

Use Google Cloud Storage!

Browsing data directories saved to Google Cloud Storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your Google Storage bucket

Service Account Key

Congratulations!

super-gradients is now integrated with Google Cloud Storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate Azure Cloud Storage
Use Azure Storage
Select bucket
Set key
Finish

Use Azure Cloud Storage!

Browsing data directories saved to Azure Cloud Storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your Azure Storage bucket

Access key (If needed)

Congratulations!

super-gradients is now integrated with Azure Cloud Storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate S3 compatible storage
Use S3 like remote
Select bucket
Access key
Finish

Use any S3 compatible storage!

Browsing data directories saved to S3 compatible storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your S3 bucket

Bucket name cannot be the same as the repository name. Please change one of them.

Access key (If needed)

Congratulations!

super-gradients is now integrated with your S3 compatible storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes