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

#378 Feature/sg 281 add kd notebook

Merged
Ghost merged 1 commits into Deci-AI:master from deci-ai:feature/SG-281-add_kd_notebook
Conversation 0 Files changed 1 Commits 1
File Compare
super-gradients
/
docs
/
super_gradients.training.datasets.html
View left file View right file
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
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
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376 
  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.       <link rel="stylesheet" href="_static/graphviz.css" type="text/css" />
    11. 

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

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

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

  14.   
    15. 

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

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

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

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

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

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

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

  22. </head>
    23. 

  23. 

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

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

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

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

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

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

  30.           </a>
    31. 

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

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

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

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

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

  36.   </form>
    37. 

  37. </div>
    38. 

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

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

  40. <ul>
    41. 

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

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

  43. </ul>
    44. 

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

  45. <ul>
    46. 

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

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

  48. </ul>
    49. 

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

  50. <ul>
    51. 

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

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

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

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

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

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

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

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

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

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

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

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

  63. </ul>
    64. 

  64. 

  65.         </div>
    66. 

  66.       </div>
    67. 

  67.     </nav>
    68. 

  68. 

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

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

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

  72.       </nav>
    73. 

  73. 

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

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

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

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

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

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

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

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

  82.       </li>
    83. 

  83.   </ul>
    84. 

  84.   <hr/>
    85. 

  85. </div>
    86. 

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

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

  88.              
    89. 

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

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

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

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

  93. <div class="toctree-wrapper compound">
    94. 

  94. <ul>
    95. 

  95. <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>
    96. 

  96. <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>
    97. 

  97. </ul>
    98. 

  98. </li>
    99. 

  99. <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>
    100. 

  100. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#submodules">Submodules</a></li>
    101. 

  101. <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>
    102. 

  102. <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>
    103. 

  103. </ul>
    104. 

  104. </li>
    105. 

  105. <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>
    106. 

  106. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.detection_datasets.html#submodules">Submodules</a></li>
    107. 

  107. <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>
    108. 

  108. <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>
    109. 

  109. <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>
    110. 

  110. <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>
    111. 

  111. </ul>
    112. 

  112. </li>
    113. 

  113. <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>
    114. 

  114. <li class="toctree-l2"><a class="reference internal" href="super_gradients.training.datasets.segmentation_datasets.html#submodules">Submodules</a></li>
    115. 

  115. <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>
    116. 

  116. <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>
    117. 

  117. <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>
    118. 

  118. <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>
    119. 

  119. <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>
    120. 

  120. <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>
    121. 

  121. </ul>
    122. 

  122. </li>
    123. 

  123. </ul>
    124. 

  124. </div>
    125. 

  125. </section>
    126. 

  126. <section id="submodules">
    127. 

  127. <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline"></a></h2>
    128. 

  128. </section>
    129. 

  129. <section id="module-super_gradients.training.datasets.all_datasets">
    130. 

  130. <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>
    131. 

  131. <dl class="py exception">
    132. 

  132. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.DataSetDoesNotExistException">
    133. 

  133. <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>
    134. 

  134. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">Exception</span></code></p>
    135. 

  135. <p>The requested dataset does not exist, or is not implemented.</p>
    136. 

  136. </dd></dl>
    137. 

  137. 

  138. <dl class="py class">
    139. 

  139. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets">
    140. 

  140. <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>
    141. 

  141. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    142. 

  142. <p>Holds all of the different library dataset dictionaries, by DL Task mapping</p>
    143. 

  143. <blockquote>
    144. 

  144. <div><dl class="simple">
    145. 

  145. <dt>Attributes:</dt><dd><p>CLASSIFICATION          Dictionary of Classification Data sets
    146. 

  146. OBJECT_DETECTION        Dictionary of Object Detection Data sets
    147. 

  147. SEMANTIC_SEGMENTATION   Dictionary of Semantic Segmentation Data sets</p>
    148. 

  148. </dd>
    149. 

  149. </dl>
    150. 

  150. </div></blockquote>
    151. 

  151. <dl class="py attribute">
    152. 

  152. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.CLASSIFICATION">
    153. 

  153. <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>
    154. 

  154. <dd></dd></dl>
    155. 

  155. 

  156. <dl class="py attribute">
    157. 

  157. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.OBJECT_DETECTION">
    158. 

  158. <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></em><a class="headerlink" href="#super_gradients.training.datasets.all_datasets.SgLibraryDatasets.OBJECT_DETECTION" title="Permalink to this definition"></a></dt>
    159. 

  159. <dd></dd></dl>
    160. 

  160. 

  161. <dl class="py attribute">
    162. 

  162. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.SEMANTIC_SEGMENTATION">
    163. 

  163. <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>
    164. 

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

  165. 

  166. <dl class="py method">
    167. 

  167. <dt class="sig sig-object py" id="super_gradients.training.datasets.all_datasets.SgLibraryDatasets.get_all_available_datasets">
    168. 

  168. <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>
    169. 

  169. <dd><p>Gets all the available datasets.</p>
    170. 

  170. </dd></dl>
    171. 

  171. 

  172. <dl class="py method">
    173. 

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

  174. <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>
    175. 

  175. <dd><p>Get’s a dataset with a given name for a given deep learning task.
    176. 

  176. examp:
    177. 

  177. &gt;&gt;&gt; SgLibraryDatasets.get_dataset(dl_task=’classification’, dataset_name=’cifar_100’)
    178. 

  178. &gt;&gt;&gt; &lt;Cifar100DatasetInterface instance&gt;</p>
    179. 

  179. </dd></dl>
    180. 

  180. 

  181. </dd></dl>
    182. 

  182. 

  183. </section>
    184. 

  184. <section id="module-super_gradients.training.datasets.auto_augment">
    185. 

  185. <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>
    186. 

  186. <p>RandAugment
    187. 

  187. RandAugment is a variant of AutoAugment which randomly selects transformations</p>
    188. 

  188. <blockquote>
    189. 

  189. <div><p>from AutoAugment to be applied on an image.</p>
    190. 

  190. </div></blockquote>
    191. 

  191. <dl class="simple">
    192. 

  192. <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>
    193. 

  193. </dd>
    194. 

  194. <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>
    195. 

  195. </dd>
    196. 

  196. </dl>
    197. 

  197. <dl class="py function">
    198. 

  198. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.shear_x">
    199. 

  199. <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>
    200. 

  200. <dd></dd></dl>
    201. 

  201. 

  202. <dl class="py function">
    203. 

  203. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.shear_y">
    204. 

  204. <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>
    205. 

  205. <dd></dd></dl>
    206. 

  206. 

  207. <dl class="py function">
    208. 

  208. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_x_rel">
    209. 

  209. <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>
    210. 

  210. <dd></dd></dl>
    211. 

  211. 

  212. <dl class="py function">
    213. 

  213. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_y_rel">
    214. 

  214. <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>
    215. 

  215. <dd></dd></dl>
    216. 

  216. 

  217. <dl class="py function">
    218. 

  218. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.translate_x_abs">
    219. 

  219. <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>
    220. 

  220. <dd></dd></dl>
    221. 

  221. 

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

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

  224. <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>
    225. 

  225. <dd></dd></dl>
    226. 

  226. 

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

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

  229. <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>
    230. 

  230. <dd></dd></dl>
    231. 

  231. 

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

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

  234. <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>
    235. 

  235. <dd></dd></dl>
    236. 

  236. 

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

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

  239. <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>
    240. 

  240. <dd></dd></dl>
    241. 

  241. 

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

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

  244. <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>
    245. 

  245. <dd></dd></dl>
    246. 

  246. 

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

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

  249. <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>
    250. 

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

  251. 

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

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

  254. <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>
    255. 

  255. <dd></dd></dl>
    256. 

  256. 

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

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

  259. <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>
    260. 

  260. <dd></dd></dl>
    261. 

  261. 

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

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

  264. <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>
    265. 

  265. <dd></dd></dl>
    266. 

  266. 

  267. <dl class="py function">
    268. 

  268. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.color">
    269. 

  269. <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>
    270. 

  270. <dd></dd></dl>
    271. 

  271. 

  272. <dl class="py function">
    273. 

  273. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.brightness">
    274. 

  274. <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>
    275. 

  275. <dd></dd></dl>
    276. 

  276. 

  277. <dl class="py function">
    278. 

  278. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.sharpness">
    279. 

  279. <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>
    280. 

  280. <dd></dd></dl>
    281. 

  281. 

  282. <dl class="py class">
    283. 

  283. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.AugmentOp">
    284. 

  284. <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>
    285. 

  285. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    286. 

  286. <p>single auto augment operations</p>
    287. 

  287. </dd></dl>
    288. 

  288. 

  289. <dl class="py function">
    290. 

  290. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.rand_augment_ops">
    291. 

  291. <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>
    292. 

  292. <dd></dd></dl>
    293. 

  293. 

  294. <dl class="py class">
    295. 

  295. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.RandAugment">
    296. 

  296. <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>
    297. 

  297. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    298. 

  298. <p>Random auto augment class, will select auto augment transforms according to probability weights for each op</p>
    299. 

  299. </dd></dl>
    300. 

  300. 

  301. <dl class="py function">
    302. 

  302. <dt class="sig sig-object py" id="super_gradients.training.datasets.auto_augment.rand_augment_transform">
    303. 

  303. <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>
    304. 

  304. <dd><p>Create a RandAugment transform</p>
    305. 

  305. <dl class="field-list simple">
    306. 

  306. <dt class="field-odd">Parameters</dt>
    307. 

  307. <dd class="field-odd"><p><strong>config_str</strong> – String defining configuration of random augmentation. Consists of multiple sections separated by</p>
    308. 

  308. </dd>
    309. 

  309. </dl>
    310. 

  310. <p>dashes (‘-‘). The first section defines the specific variant of rand augment (currently only ‘rand’). The remaining
    311. 

  311. sections, not order sepecific determine</p>
    312. 

  312. <blockquote>
    313. 

  313. <div><p>‘m’ - integer magnitude of rand augment
    314. 

  314. ‘n’ - integer num layers (number of transform ops selected per image)
    315. 

  315. ‘w’ - integer probabiliy weight index (index of a set of weights to influence choice of op)
    316. 

  316. ‘mstd’ -  float std deviation of magnitude noise applied
    317. 

  317. ‘inc’ - integer (bool), use augmentations that increase in severity with magnitude (default: 0)</p>
    318. 

  318. </div></blockquote>
    319. 

  319. <p>Ex ‘rand-m9-n3-mstd0.5’ results in RandAugment with magnitude 9, num_layers 3, magnitude_std 0.5
    320. 

  320. ‘rand-mstd1-w0’ results in magnitude_std 1.0, weights 0, default magnitude of 10 and num_layers 2</p>
    321. 

  321. <dl class="field-list simple">
    322. 

  322. <dt class="field-odd">Parameters</dt>
    323. 

  323. <dd class="field-odd"><p><strong>hparams</strong> – Other hparams (kwargs) for the RandAugmentation scheme</p>
    324. 

  324. </dd>
    325. 

  325. <dt class="field-even">Returns</dt>
    326. 

  326. <dd class="field-even"><p>A PyTorch compatible Transform</p>
    327. 

  327. </dd>
    328. 

  328. </dl>
    329. 

  329. </dd></dl>
    330. 

  330. 

  331. </section>
    332. 

  332. <section id="module-super_gradients.training.datasets.data_augmentation">
    333. 

  333. <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>
    334. 

  334. <dl class="py class">
    335. 

  335. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation">
    336. 

  336. <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>
    337. 

  337. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    338. 

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

  339. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.to_tensor">
    340. 

  340. <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>
    341. 

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

  342. 

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

  344. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.normalize">
    345. 

  345. <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>
    346. 

  346. <dd></dd></dl>
    347. 

  347. 

  348. <dl class="py method">
    349. 

  349. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.DataAugmentation.cutout">
    350. 

  350. <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>
    351. 

  351. <dd></dd></dl>
    352. 

  352. 

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

  354. 

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

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

  357. <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">0.8140],</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>
    358. 

  358. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    359. 

  359. <p>Lighting noise(AlexNet - style PCA - based noise)
    360. 

  360. Taken from fastai Imagenet training -
    361. 

  361. <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>
    362. 

  362. To use:</p>
    363. 

  363. <blockquote>
    364. 

  364. <div><ul class="simple">
    365. 

  365. <li><p>training_params = {“imagenet_pca_aug”: 0.1}</p></li>
    366. 

  366. <li><p>Default training_params arg is 0.0 (“don’t use”)</p></li>
    367. 

  367. <li><p>0.1 is that default in the original paper</p></li>
    368. 

  368. </ul>
    369. 

  369. </div></blockquote>
    370. 

  370. </dd></dl>
    371. 

  371. 

  372. <dl class="py class">
    373. 

  373. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.RandomErase">
    374. 

  374. <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>
    375. 

  375. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchvision.transforms.transforms.RandomErasing</span></code></p>
    376. 

  376. <p>A simple class that translates the parameters supported in SuperGradient’s code base</p>
    377. 

  377. <dl class="py attribute">
    378. 

  378. <dt class="sig sig-object py" id="super_gradients.training.datasets.data_augmentation.RandomErase.training">
    379. 

  379. <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>
    380. 

  380. <dd></dd></dl>
    381. 

  381. 

  382. </dd></dl>
    383. 

  383. 

  384. </section>
    385. 

  385. <section id="module-super_gradients.training.datasets.datasets_conf">
    386. 

  386. <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>
    387. 

  387. </section>
    388. 

  388. <section id="module-super_gradients.training.datasets.datasets_utils">
    389. 

  389. <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>
    390. 

  390. <dl class="py function">
    391. 

  391. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.get_mean_and_std_torch">
    392. 

  392. <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>
    393. 

  393. <dd><p>A function for getting the mean and std of large datasets using pytorch dataloader and gpu functionality.</p>
    394. 

  394. <dl class="field-list simple">
    395. 

  395. <dt class="field-odd">Parameters</dt>
    396. 

  396. <dd class="field-odd"><ul class="simple">
    397. 

  397. <li><p><strong>data_dir</strong> – String, path to none-library dataset folder. For example “/data/Imagenette” or “/data/TinyImagenet”</p></li>
    398. 

  398. <li><p><strong>dataloader</strong> – a torch DataLoader, as it would feed the data into the trainer (including transforms etc).</p></li>
    399. 

  399. <li><p><strong>RandomResizeSize</strong> – Int, the size of the RandomResizeCrop as it appears in the DataInterface (for example, for Imagenet,</p></li>
    400. 

  400. </ul>
    401. 

  401. </dd>
    402. 

  402. </dl>
    403. 

  403. <p>this value should be 224).
    404. 

  404. :return: 2 lists,mean and std, each one of len 3 (1 for each channel)</p>
    405. 

  405. </dd></dl>
    406. 

  406. 

  407. <dl class="py function">
    408. 

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

  409. <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>
    410. 

  410. <dd><p>Compute the mean and std value of dataset.</p>
    411. 

  411. </dd></dl>
    412. 

  412. 

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

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

  415. <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>
    416. 

  416. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">abc.ABC</span></code></p>
    417. 

  417. <p>A collate function (for torch DataLoader)</p>
    418. 

  418. </dd></dl>
    419. 

  419. 

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

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

  422. <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>
    423. 

  423. <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>
    424. 

  424. <p>A function (for torch DataLoader) which executes a sequence of sub collate functions</p>
    425. 

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

  426. 

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

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

  429. <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>
    430. 

  430. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    431. 

  431. </dd></dl>
    432. 

  432. 

  433. <dl class="py class">
    434. 

  434. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.MultiScaleCollateFunction">
    435. 

  435. <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>
    436. 

  436. <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>
    437. 

  437. <p>a collate function to implement multi-scale data augmentation
    438. 

  438. according to <a class="reference external" href="https://arxiv.org/pdf/1612.08242.pdf">https://arxiv.org/pdf/1612.08242.pdf</a></p>
    439. 

  439. </dd></dl>
    440. 

  440. 

  441. <dl class="py class">
    442. 

  442. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.AbstractPrePredictionCallback">
    443. 

  443. <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">AbstractPrePredictionCallback</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#AbstractPrePredictionCallback"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.AbstractPrePredictionCallback" title="Permalink to this definition"></a></dt>
    444. 

  444. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">abc.ABC</span></code></p>
    445. 

  445. <dl class="simple">
    446. 

  446. <dt>Abstract class for forward pass preprocessing function, to be used by passing its inheritors through training_params</dt><dd><p>pre_prediction_callback keyword arg.</p>
    447. 

  447. </dd>
    448. 

  448. </dl>
    449. 

  449. <p>Should implement __call__ and return images, targets after applying the desired preprocessing.</p>
    450. 

  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.MultiscalePrePredictionCallback">
    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">MultiscalePrePredictionCallback</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">multiscale_range</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">5</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#MultiscalePrePredictionCallback"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback" title="Permalink to this definition"></a></dt>
    455. 

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

  456. <p>Mutiscale pre-prediction callback pass function.</p>
    457. 

  457. <p>When passed through train_params images, targets will be applied by the below transform to support multi scaling
    458. 

  458. on the fly.</p>
    459. 

  459. <dl class="simple">
    460. 

  460. <dt>After each self.frequency forward passes, change size randomly from</dt><dd><p>(input_size-self.multiscale_range*self.image_size_steps, input_size-(self.multiscale_range-1)*self.image_size_steps,
    461. 

  461. …input_size+self.multiscale_range*self.image_size_steps)</p>
    462. 

  462. </dd>
    463. 

  463. </dl>
    464. 

  464. <dl class="py attribute">
    465. 

  465. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback.multiscale_range">
    466. 

  466. <span class="sig-name descname"><span class="pre">multiscale_range</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback.multiscale_range" title="Permalink to this definition"></a></dt>
    467. 

  467. <dd><p>(int) Range of values for resize sizes as discussed above (default=5)</p>
    468. 

  468. </dd></dl>
    469. 

  469. 

  470. <dl class="py attribute">
    471. 

  471. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback.image_size_steps">
    472. 

  472. <span class="sig-name descname"><span class="pre">image_size_steps</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback.image_size_steps" title="Permalink to this definition"></a></dt>
    473. 

  473. <dd><p>(int) Image step sizes as discussed abov (default=32)</p>
    474. 

  474. </dd></dl>
    475. 

  475. 

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

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

  478. <span class="sig-name descname"><span class="pre">change_frequency</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback.change_frequency" title="Permalink to this definition"></a></dt>
    479. 

  479. <dd><p>(int) The frequency to apply change in input size.</p>
    480. 

  480. </dd></dl>
    481. 

  481. 

  482. </dd></dl>
    483. 

  483. 

  484. <dl class="py class">
    485. 

  485. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback">
    486. 

  486. <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">DetectionMultiscalePrePredictionCallback</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">multiscale_range</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">5</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#DetectionMultiscalePrePredictionCallback"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback" title="Permalink to this definition"></a></dt>
    487. 

  487. <dd><p>Bases: <a class="reference internal" href="#super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback" title="super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.datasets_utils.MultiscalePrePredictionCallback</span></code></a></p>
    488. 

  488. <p>Mutiscalepre-prediction callback for object detection.</p>
    489. 

  489. <p>When passed through train_params images, targets will be applied by the below transform to support multi scaling
    490. 

  490. on the fly.</p>
    491. 

  491. <dl class="simple">
    492. 

  492. <dt>After each self.frequency forward passes, change size randomly from</dt><dd><p>(input_size-self.multiscale_range*self.image_size_steps, input_size-(self.multiscale_range-1)*self.image_size_steps,
    493. 

  493. …input_size+self.multiscale_range*self.image_size_steps) and apply the same rescaling to the box coordinates.</p>
    494. 

  494. </dd>
    495. 

  495. </dl>
    496. 

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

  497. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.multiscale_range">
    498. 

  498. <span class="sig-name descname"><span class="pre">multiscale_range</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.multiscale_range" title="Permalink to this definition"></a></dt>
    499. 

  499. <dd><p>(int) Range of values for resize sizes as discussed above (default=5)</p>
    500. 

  500. </dd></dl>
    501. 

  501. 

  502. <dl class="py attribute">
    503. 

  503. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.image_size_steps">
    504. 

  504. <span class="sig-name descname"><span class="pre">image_size_steps</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.image_size_steps" title="Permalink to this definition"></a></dt>
    505. 

  505. <dd><p>(int) Image step sizes as discussed abov (default=32)</p>
    506. 

  506. </dd></dl>
    507. 

  507. 

  508. <dl class="py attribute">
    509. 

  509. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.change_frequency">
    510. 

  510. <span class="sig-name descname"><span class="pre">change_frequency</span></span><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.DetectionMultiscalePrePredictionCallback.change_frequency" title="Permalink to this definition"></a></dt>
    511. 

  511. <dd><p>(int) The frequency to apply change in input size.</p>
    512. 

  512. </dd></dl>
    513. 

  513. 

  514. </dd></dl>
    515. 

  515. 

  516. <dl class="py class">
    517. 

  517. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation">
    518. 

  518. <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>
    519. 

  519. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchvision.transforms.transforms.RandomResizedCrop</span></code></p>
    520. 

  520. <p>Crop the given PIL Image to random size and aspect ratio with explicitly chosen or random interpolation.</p>
    521. 

  521. <p>A crop of random size (default: of 0.08 to 1.0) of the original size and a random
    522. 

  522. aspect ratio (default: of 3/4 to 4/3) of the original aspect ratio is made. This crop
    523. 

  523. is finally resized to given size.
    524. 

  524. This is popularly used to train the Inception networks.</p>
    525. 

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

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

  527. <dd class="field-odd"><ul class="simple">
    528. 

  528. <li><p><strong>size</strong> – expected output size of each edge</p></li>
    529. 

  529. <li><p><strong>scale</strong> – range of size of the origin size cropped</p></li>
    530. 

  530. <li><p><strong>ratio</strong> – range of aspect ratio of the origin aspect ratio cropped</p></li>
    531. 

  531. <li><p><strong>interpolation</strong> – Default: PIL.Image.BILINEAR</p></li>
    532. 

  532. </ul>
    533. 

  533. </dd>
    534. 

  534. </dl>
    535. 

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

  536. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.forward">
    537. 

  537. <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>
    538. 

  538. <dd><dl class="field-list simple">
    539. 

  539. <dt class="field-odd">Parameters</dt>
    540. 

  540. <dd class="field-odd"><p><strong>img</strong> (<em>PIL Image</em>) – Image to be cropped and resized.</p>
    541. 

  541. </dd>
    542. 

  542. <dt class="field-even">Returns</dt>
    543. 

  543. <dd class="field-even"><p>Randomly cropped and resized image.</p>
    544. 

  544. </dd>
    545. 

  545. <dt class="field-odd">Return type</dt>
    546. 

  546. <dd class="field-odd"><p>PIL Image</p>
    547. 

  547. </dd>
    548. 

  548. </dl>
    549. 

  549. </dd></dl>
    550. 

  550. 

  551. <dl class="py attribute">
    552. 

  552. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.RandomResizedCropAndInterpolation.training">
    553. 

  553. <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>
    554. 

  554. <dd></dd></dl>
    555. 

  555. 

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

  557. 

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

  559. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger">
    560. 

  560. <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">sg_logger</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">super_gradients.common.sg_loggers.abstract_sg_logger.AbstractSGLogger</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>
    561. 

  561. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    562. 

  562. <dl class="py attribute">
    563. 

  563. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.logger">
    564. 

  564. <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>
    565. 

  565. <dd></dd></dl>
    566. 

  566. 

  567. <dl class="py attribute">
    568. 

  568. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.DEFAULT_SUMMARY_PARAMS">
    569. 

  569. <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>
    570. 

  570. <dd></dd></dl>
    571. 

  571. 

  572. <dl class="py method">
    573. 

  573. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.DatasetStatisticsTensorboardLogger.analyze">
    574. 

  574. <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">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">all_classes</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></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>
    575. 

  575. <dd><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>data_loader</strong> – the dataset data loader</p></li>
    579. 

  579. <li><p><strong>dataset_params</strong> – the dataset parameters</p></li>
    580. 

  580. <li><p><strong>title</strong> – the title for this dataset (i.e. Coco 2017 test set)</p></li>
    581. 

  581. <li><p><strong>anchors</strong> – the list of anchors used by the model. applicable only for detection datasets</p></li>
    582. 

  582. <li><p><strong>all_classes</strong> – the list of all classes names</p></li>
    583. 

  583. </ul>
    584. 

  584. </dd>
    585. 

  585. </dl>
    586. 

  586. </dd></dl>
    587. 

  587. 

  588. </dd></dl>
    589. 

  589. 

  590. <dl class="py function">
    591. 

  591. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.get_color_augmentation">
    592. 

  592. <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>
    593. 

  593. <dd><p>Returns color augmentation class. As these augmentation cannot work on top one another, only one is returned
    594. 

  594. according to rand_augment_config_string</p>
    595. 

  595. <dl class="field-list simple">
    596. 

  596. <dt class="field-odd">Parameters</dt>
    597. 

  597. <dd class="field-odd"><ul class="simple">
    598. 

  598. <li><p><strong>rand_augment_config_string</strong> – string which defines the auto augment configurations.
    599. 

  599. If none, color jitter will be returned. For possibile values see auto_augment.py</p></li>
    600. 

  600. <li><p><strong>color_jitter</strong> – tuple for color jitter value.</p></li>
    601. 

  601. <li><p><strong>crop_size</strong> – relevant only for auto augment</p></li>
    602. 

  602. <li><p><strong>img_mean</strong> – relevant only for auto augment</p></li>
    603. 

  603. </ul>
    604. 

  604. </dd>
    605. 

  605. <dt class="field-even">Returns</dt>
    606. 

  606. <dd class="field-even"><p>RandAugment transform or ColorJitter</p>
    607. 

  607. </dd>
    608. 

  608. </dl>
    609. 

  609. </dd></dl>
    610. 

  610. 

  611. <dl class="py function">
    612. 

  612. <dt class="sig sig-object py" id="super_gradients.training.datasets.datasets_utils.worker_init_reset_seed">
    613. 

  613. <span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.datasets_utils.</span></span><span class="sig-name descname"><span class="pre">worker_init_reset_seed</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">worker_id</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/datasets_utils.html#worker_init_reset_seed"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.datasets_utils.worker_init_reset_seed" title="Permalink to this definition"></a></dt>
    614. 

  614. <dd><p>Make sure each process has different random seed, especially for ‘fork’ method.
    615. 

  615. Check <a class="reference external" href="https://github.com/pytorch/pytorch/issues/63311">https://github.com/pytorch/pytorch/issues/63311</a> for more details.</p>
    616. 

  616. <dl class="field-list simple">
    617. 

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

  618. <dd class="field-odd"><p><strong>worker_id</strong> – placeholder (needs to be passed to DataLoader init).</p>
    619. 

  619. </dd>
    620. 

  620. </dl>
    621. 

  621. </dd></dl>
    622. 

  622. 

  623. </section>
    624. 

  624. <section id="module-super_gradients.training.datasets.mixup">
    625. 

  625. <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>
    626. 

  626. <p>Mixup and Cutmix</p>
    627. 

  627. <p>Papers:
    628. 

  628. mixup: Beyond Empirical Risk Minimization (<a class="reference external" href="https://arxiv.org/abs/1710.09412">https://arxiv.org/abs/1710.09412</a>)</p>
    629. 

  629. <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>
    630. 

  630. <p>Code Reference:
    631. 

  631. CutMix: <a class="reference external" href="https://github.com/clovaai/CutMix-PyTorch">https://github.com/clovaai/CutMix-PyTorch</a>
    632. 

  632. 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>
    633. 

  633. <dl class="py function">
    634. 

  634. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.one_hot">
    635. 

  635. <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>
    636. 

  636. <dd></dd></dl>
    637. 

  637. 

  638. <dl class="py function">
    639. 

  639. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.mixup_target">
    640. 

  640. <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>
    641. 

  641. <dd><p>generate a smooth target (label) two-hot tensor to support the mixed images with different labels
    642. 

  642. :param target: the targets tensor
    643. 

  643. :param num_classes: number of classes (to set the final tensor size)
    644. 

  644. :param lam: percentage of label a range [0, 1] in the mixing
    645. 

  645. :param smoothing: the smoothing multiplier
    646. 

  646. :param device: usable device [‘cuda’, ‘cpu’]
    647. 

  647. :return:</p>
    648. 

  648. </dd></dl>
    649. 

  649. 

  650. <dl class="py function">
    651. 

  651. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.rand_bbox">
    652. 

  652. <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>
    653. 

  653. <dd><p>Standard CutMix bounding-box
    654. 

  654. Generates a random square bbox based on lambda value. This impl includes
    655. 

  655. support for enforcing a border margin as percent of bbox dimensions.</p>
    656. 

  656. <dl class="field-list simple">
    657. 

  657. <dt class="field-odd">Parameters</dt>
    658. 

  658. <dd class="field-odd"><ul class="simple">
    659. 

  659. <li><p><strong>img_shape</strong> – Image shape as tuple</p></li>
    660. 

  660. <li><p><strong>lam</strong> – Cutmix lambda value</p></li>
    661. 

  661. <li><p><strong>margin</strong> – Percentage of bbox dimension to enforce as margin (reduce amount of box outside image)</p></li>
    662. 

  662. <li><p><strong>count</strong> – Number of bbox to generate</p></li>
    663. 

  663. </ul>
    664. 

  664. </dd>
    665. 

  665. </dl>
    666. 

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

  667. 

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

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

  670. <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>
    671. 

  671. <dd><p>Min-Max CutMix bounding-box
    672. 

  672. Inspired by Darknet cutmix impl, generates a random rectangular bbox
    673. 

  673. based on min/max percent values applied to each dimension of the input image.</p>
    674. 

  674. <p>Typical defaults for minmax are usually in the  .2-.3 for min and .8-.9 range for max.</p>
    675. 

  675. <dl class="field-list simple">
    676. 

  676. <dt class="field-odd">Parameters</dt>
    677. 

  677. <dd class="field-odd"><ul class="simple">
    678. 

  678. <li><p><strong>img_shape</strong> – Image shape as tuple</p></li>
    679. 

  679. <li><p><strong>minmax</strong> – Min and max bbox ratios (as percent of image size)</p></li>
    680. 

  680. <li><p><strong>count</strong> – Number of bbox to generate</p></li>
    681. 

  681. </ul>
    682. 

  682. </dd>
    683. 

  683. </dl>
    684. 

  684. </dd></dl>
    685. 

  685. 

  686. <dl class="py function">
    687. 

  687. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.cutmix_bbox_and_lam">
    688. 

  688. <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>
    689. 

  689. <dd><p>Generate bbox and apply lambda correction.</p>
    690. 

  690. </dd></dl>
    691. 

  691. 

  692. <dl class="py class">
    693. 

  693. <dt class="sig sig-object py" id="super_gradients.training.datasets.mixup.CollateMixup">
    694. 

  694. <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>
    695. 

  695. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    696. 

  696. <p>Collate with Mixup/Cutmix that applies different params to each element or whole batch
    697. 

  697. A Mixup impl that’s performed while collating the batches.</p>
    698. 

  698. </dd></dl>
    699. 

  699. 

  700. </section>
    701. 

  701. <section id="module-super_gradients.training.datasets.sg_dataset">
    702. 

  702. <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>
    703. 

  703. <dl class="py class">
    704. 

  704. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset">
    705. 

  705. <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>
    706. 

  706. <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>
    707. 

  707. <dl class="py method">
    708. 

  708. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.numpy_loader_func">
    709. 

  709. <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>
    710. 

  710. <dd><dl class="simple">
    711. 

  711. <dt>_numpy_loader_func - Uses numpy load func</dt><dd><dl class="field-list simple">
    712. 

  712. <dt class="field-odd">param path</dt>
    713. 

  713. <dd class="field-odd"><p></p></dd>
    714. 

  714. <dt class="field-even">return</dt>
    715. 

  715. <dd class="field-even"><p></p></dd>
    716. 

  716. </dl>
    717. 

  717. </dd>
    718. 

  718. </dl>
    719. 

  719. </dd></dl>
    720. 

  720. 

  721. <dl class="py method">
    722. 

  722. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.BaseSgVisionDataset.text_file_loader_func">
    723. 

  723. <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>
    724. 

  724. <dd><blockquote>
    725. 

  725. <div><dl class="simple">
    726. 

  726. <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">
    727. 

  727. <dt class="field-odd">param text_file_path</dt>
    728. 

  728. <dd class="field-odd"><p>Input text file</p>
    729. 

  729. </dd>
    730. 

  730. <dt class="field-even">param inline_splitter</dt>
    731. 

  731. <dd class="field-even"><p>The char to use in order to separate between different VALUES of the SAME vector
    732. 

  732. please notice that DIFFERENT VECTORS SHOULD BE IN SEPARATE LINES (’</p>
    733. 

  733. </dd>
    734. 

  734. </dl>
    735. 

  735. </dd>
    736. 

  736. </dl>
    737. 

  737. </div></blockquote>
    738. 

  738. <dl class="simple">
    739. 

  739. <dt>‘) SEPARATED</dt><dd><dl class="field-list simple">
    740. 

  740. <dt class="field-odd">return</dt>
    741. 

  741. <dd class="field-odd"><p>a list of tuples, where each tuple is a vector of target values</p>
    742. 

  742. </dd>
    743. 

  743. </dl>
    744. 

  744. </dd>
    745. 

  745. </dl>
    746. 

  746. </dd></dl>
    747. 

  747. 

  748. </dd></dl>
    749. 

  749. 

  750. <dl class="py class">
    751. 

  751. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.DirectoryDataSet">
    752. 

  752. <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>
    753. 

  753. <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>
    754. 

  754. <dl class="simple">
    755. 

  755. <dt>DirectoryDataSet - A PyTorch Vision Data Set extension that receives a root Dir and two separate sub directories:</dt><dd><ul class="simple">
    756. 

  756. <li><p>Sub-Directory for Samples</p></li>
    757. 

  757. <li><p>Sub-Directory for Targets</p></li>
    758. 

  758. </ul>
    759. 

  759. </dd>
    760. 

  760. </dl>
    761. 

  761. </dd></dl>
    762. 

  762. 

  763. <dl class="py class">
    764. 

  764. <dt class="sig sig-object py" id="super_gradients.training.datasets.sg_dataset.ListDataset">
    765. 

  765. <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>
    766. 

  766. <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>
    767. 

  767. <dl>
    768. 

  768. <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
    769. 

  769. path but with a different extension, i.e:</p>
    770. 

  770. <blockquote>
    771. 

  771. <div><dl class="simple">
    772. 

  772. <dt>for the samples paths:  (That appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.png
    773. 

  773. /root/dataset/class_y/sample123.png</p>
    774. 

  774. </dd>
    775. 

  775. <dt>the matching labels paths:  (That DO NOT appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.ext
    776. 

  776. /root/dataset/class_y/sample123.ext</p>
    777. 

  777. </dd>
    778. 

  778. </dl>
    779. 

  779. </div></blockquote>
    780. 

  780. </dd>
    781. 

  781. </dl>
    782. 

  782. </dd></dl>
    783. 

  783. 

  784. </section>
    785. 

  785. <section id="module-super_gradients.training.datasets">
    786. 

  786. <span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-super_gradients.training.datasets" title="Permalink to this headline"></a></h2>
    787. 

  787. <dl class="py class">
    788. 

  788. <dt class="sig sig-object py" id="super_gradients.training.datasets.DataAugmentation">
    789. 

  789. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</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.DataAugmentation" title="Permalink to this definition"></a></dt>
    790. 

  790. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    791. 

  791. <dl class="py method">
    792. 

  792. <dt class="sig sig-object py" id="super_gradients.training.datasets.DataAugmentation.to_tensor">
    793. 

  793. <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.DataAugmentation.to_tensor" title="Permalink to this definition"></a></dt>
    794. 

  794. <dd></dd></dl>
    795. 

  795. 

  796. <dl class="py method">
    797. 

  797. <dt class="sig sig-object py" id="super_gradients.training.datasets.DataAugmentation.normalize">
    798. 

  798. <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.DataAugmentation.normalize" title="Permalink to this definition"></a></dt>
    799. 

  799. <dd></dd></dl>
    800. 

  800. 

  801. <dl class="py method">
    802. 

  802. <dt class="sig sig-object py" id="super_gradients.training.datasets.DataAugmentation.cutout">
    803. 

  803. <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.DataAugmentation.cutout" title="Permalink to this definition"></a></dt>
    804. 

  804. <dd></dd></dl>
    805. 

  805. 

  806. </dd></dl>
    807. 

  807. 

  808. <dl class="py class">
    809. 

  809. <dt class="sig sig-object py" id="super_gradients.training.datasets.ListDataset">
    810. 

  810. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</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.ListDataset" title="Permalink to this definition"></a></dt>
    811. 

  811. <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>
    812. 

  812. <dl>
    813. 

  813. <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
    814. 

  814. path but with a different extension, i.e:</p>
    815. 

  815. <blockquote>
    816. 

  816. <div><dl class="simple">
    817. 

  817. <dt>for the samples paths:  (That appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.png
    818. 

  818. /root/dataset/class_y/sample123.png</p>
    819. 

  819. </dd>
    820. 

  820. <dt>the matching labels paths:  (That DO NOT appear in the list file)</dt><dd><p>/root/dataset/class_x/sample1.ext
    821. 

  821. /root/dataset/class_y/sample123.ext</p>
    822. 

  822. </dd>
    823. 

  823. </dl>
    824. 

  824. </div></blockquote>
    825. 

  825. </dd>
    826. 

  826. </dl>
    827. 

  827. </dd></dl>
    828. 

  828. 

  829. <dl class="py class">
    830. 

  830. <dt class="sig sig-object py" id="super_gradients.training.datasets.DirectoryDataSet">
    831. 

  831. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</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.DirectoryDataSet" title="Permalink to this definition"></a></dt>
    832. 

  832. <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>
    833. 

  833. <dl class="simple">
    834. 

  834. <dt>DirectoryDataSet - A PyTorch Vision Data Set extension that receives a root Dir and two separate sub directories:</dt><dd><ul class="simple">
    835. 

  835. <li><p>Sub-Directory for Samples</p></li>
    836. 

  836. <li><p>Sub-Directory for Targets</p></li>
    837. 

  837. </ul>
    838. 

  838. </dd>
    839. 

  839. </dl>
    840. 

  840. </dd></dl>
    841. 

  841. 

  842. <dl class="py class">
    843. 

  843. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationDataSet">
    844. 

  844. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">SegmentationDataSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">root</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">list_file</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">samples_sub_directory</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">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">targets_sub_directory</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">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_size</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">608</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</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">512</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">augment</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset_hyper_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">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_labels</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_images</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_loader</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Callable</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">target_loader</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Callable</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">collate_fn</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Callable</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">target_extension</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">'.png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_mask_transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torchvision.transforms.transforms.Compose</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_mask_transforms_aug</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torchvision.transforms.transforms.Compose</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/segmentation_datasets/segmentation_dataset.html#SegmentationDataSet"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationDataSet" title="Permalink to this definition"></a></dt>
    845. 

  845. <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>
    846. 

  846. <dl class="py method">
    847. 

  847. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationDataSet.sample_loader">
    848. 

  848. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">sample_loader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_path</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">&lt;module</span> <span class="pre">‘PIL.Image’</span> <span class="pre">from</span> <span class="pre">‘/Users/shaniperl/opt/anaconda3/lib/python3.9/site-packages/PIL/Image.py’&gt;</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/segmentation_dataset.html#SegmentationDataSet.sample_loader"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationDataSet.sample_loader" title="Permalink to this definition"></a></dt>
    849. 

  849. <dd><dl class="simple">
    850. 

  850. <dt>sample_loader - Loads a dataset image from path using PIL</dt><dd><dl class="field-list simple">
    851. 

  851. <dt class="field-odd">param sample_path</dt>
    852. 

  852. <dd class="field-odd"><p>The path to the sample image</p>
    853. 

  853. </dd>
    854. 

  854. <dt class="field-even">return</dt>
    855. 

  855. <dd class="field-even"><p>The loaded Image</p>
    856. 

  856. </dd>
    857. 

  857. </dl>
    858. 

  858. </dd>
    859. 

  859. </dl>
    860. 

  860. </dd></dl>
    861. 

  861. 

  862. <dl class="py method">
    863. 

  863. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationDataSet.sample_transform">
    864. 

  864. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">sample_transform</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/segmentation_dataset.html#SegmentationDataSet.sample_transform"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationDataSet.sample_transform" title="Permalink to this definition"></a></dt>
    865. 

  865. <dd><p>sample_transform - Transforms the sample image</p>
    866. 

  866. <blockquote>
    867. 

  867. <div><dl class="field-list simple">
    868. 

  868. <dt class="field-odd">param image</dt>
    869. 

  869. <dd class="field-odd"><p>The input image to transform</p>
    870. 

  870. </dd>
    871. 

  871. <dt class="field-even">return</dt>
    872. 

  872. <dd class="field-even"><p>The transformed image</p>
    873. 

  873. </dd>
    874. 

  874. </dl>
    875. 

  875. </div></blockquote>
    876. 

  876. </dd></dl>
    877. 

  877. 

  878. <dl class="py method">
    879. 

  879. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationDataSet.target_loader">
    880. 

  880. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">target_loader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">target_path</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">&lt;module</span> <span class="pre">‘PIL.Image’</span> <span class="pre">from</span> <span class="pre">‘/Users/shaniperl/opt/anaconda3/lib/python3.9/site-packages/PIL/Image.py’&gt;</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/segmentation_dataset.html#SegmentationDataSet.target_loader"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationDataSet.target_loader" title="Permalink to this definition"></a></dt>
    881. 

  881. <dd><dl class="field-list simple">
    882. 

  882. <dt class="field-odd">Parameters</dt>
    883. 

  883. <dd class="field-odd"><p><strong>target_path</strong> – The path to the sample image</p>
    884. 

  884. </dd>
    885. 

  885. <dt class="field-even">Returns</dt>
    886. 

  886. <dd class="field-even"><p>The loaded Image</p>
    887. 

  887. </dd>
    888. 

  888. </dl>
    889. 

  889. </dd></dl>
    890. 

  890. 

  891. <dl class="py method">
    892. 

  892. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationDataSet.target_transform">
    893. 

  893. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">target_transform</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">target</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/segmentation_dataset.html#SegmentationDataSet.target_transform"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationDataSet.target_transform" title="Permalink to this definition"></a></dt>
    894. 

  894. <dd><p>target_transform - Transforms the sample image</p>
    895. 

  895. <blockquote>
    896. 

  896. <div><dl class="field-list simple">
    897. 

  897. <dt class="field-odd">param target</dt>
    898. 

  898. <dd class="field-odd"><p>The target mask to transform</p>
    899. 

  899. </dd>
    900. 

  900. <dt class="field-even">return</dt>
    901. 

  901. <dd class="field-even"><p>The transformed target mask</p>
    902. 

  902. </dd>
    903. 

  903. </dl>
    904. 

  904. </div></blockquote>
    905. 

  905. </dd></dl>
    906. 

  906. 

  907. </dd></dl>
    908. 

  908. 

  909. <dl class="py class">
    910. 

  910. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalVOC2012SegmentationDataSet">
    911. 

  911. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">PascalVOC2012SegmentationDataSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_suffix</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">target_suffix</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="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/pascal_voc_segmentation.html#PascalVOC2012SegmentationDataSet"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalVOC2012SegmentationDataSet" title="Permalink to this definition"></a></dt>
    912. 

  912. <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>
    913. 

  913. <p>PascalVOC2012SegmentationDataSet - Segmentation Data Set Class for Pascal VOC 2012 Data Set</p>
    914. 

  914. <dl class="py method">
    915. 

  915. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalVOC2012SegmentationDataSet.decode_segmentation_mask">
    916. 

  916. <span class="sig-name descname"><span class="pre">decode_segmentation_mask</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">label_mask</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/pascal_voc_segmentation.html#PascalVOC2012SegmentationDataSet.decode_segmentation_mask"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalVOC2012SegmentationDataSet.decode_segmentation_mask" title="Permalink to this definition"></a></dt>
    917. 

  917. <dd><dl class="simple">
    918. 

  918. <dt>decode_segmentation_mask - Decodes the colors for the Segmentation Mask</dt><dd><dl class="field-list simple">
    919. 

  919. <dt class="field-odd">param</dt>
    920. 

  920. <dd class="field-odd"><p>label_mask:  an (M,N) array of integer values denoting
    921. 

  921. the class label at each spatial location.</p>
    922. 

  922. </dd>
    923. 

  923. </dl>
    924. 

  924. </dd>
    925. 

  925. </dl>
    926. 

  926. <dl class="field-list simple">
    927. 

  927. <dt class="field-odd">Returns</dt>
    928. 

  928. <dd class="field-odd"><p></p>
    929. 

  929. </dd>
    930. 

  930. </dl>
    931. 

  931. </dd></dl>
    932. 

  932. 

  933. </dd></dl>
    934. 

  934. 

  935. <dl class="py class">
    936. 

  936. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalAUG2012SegmentationDataSet">
    937. 

  937. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">PascalAUG2012SegmentationDataSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/pascal_aug_segmentation.html#PascalAUG2012SegmentationDataSet"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalAUG2012SegmentationDataSet" title="Permalink to this definition"></a></dt>
    938. 

  938. <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>
    939. 

  939. <p>PascalAUG2012SegmentationDataSet - Segmentation Data Set Class for Pascal AUG 2012 Data Set</p>
    940. 

  940. <dl class="py method">
    941. 

  941. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalAUG2012SegmentationDataSet.target_loader">
    942. 

  942. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">target_loader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">target_path</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">&lt;module</span> <span class="pre">‘PIL.Image’</span> <span class="pre">from</span> <span class="pre">‘/Users/shaniperl/opt/anaconda3/lib/python3.9/site-packages/PIL/Image.py’&gt;</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/pascal_aug_segmentation.html#PascalAUG2012SegmentationDataSet.target_loader"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalAUG2012SegmentationDataSet.target_loader" title="Permalink to this definition"></a></dt>
    943. 

  943. <dd><dl class="field-list simple">
    944. 

  944. <dt class="field-odd">Parameters</dt>
    945. 

  945. <dd class="field-odd"><p><strong>target_path</strong> – The path to the target data</p>
    946. 

  946. </dd>
    947. 

  947. <dt class="field-even">Returns</dt>
    948. 

  948. <dd class="field-even"><p>The loaded target</p>
    949. 

  949. </dd>
    950. 

  950. </dl>
    951. 

  951. </dd></dl>
    952. 

  952. 

  953. </dd></dl>
    954. 

  954. 

  955. <dl class="py class">
    956. 

  956. <dt class="sig sig-object py" id="super_gradients.training.datasets.CoCoSegmentationDataSet">
    957. 

  957. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">CoCoSegmentationDataSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_classes_inclusion_tuples_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span> <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="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/coco_segmentation.html#CoCoSegmentationDataSet"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.CoCoSegmentationDataSet" title="Permalink to this definition"></a></dt>
    958. 

  958. <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>
    959. 

  959. <p>CoCoSegmentationDataSet - Segmentation Data Set Class for COCO 2017 Segmentation Data Set</p>
    960. 

  960. <dl class="py method">
    961. 

  961. <dt class="sig sig-object py" id="super_gradients.training.datasets.CoCoSegmentationDataSet.target_loader">
    962. 

  962. <span class="sig-name descname"><span class="pre">target_loader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mask_metadata_tuple</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">&lt;module</span> <span class="pre">‘PIL.Image’</span> <span class="pre">from</span> <span class="pre">‘/Users/shaniperl/opt/anaconda3/lib/python3.9/site-packages/PIL/Image.py’&gt;</span><a class="reference internal" href="_modules/super_gradients/training/datasets/segmentation_datasets/coco_segmentation.html#CoCoSegmentationDataSet.target_loader"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.CoCoSegmentationDataSet.target_loader" title="Permalink to this definition"></a></dt>
    963. 

  963. <dd><dl class="field-list simple">
    964. 

  964. <dt class="field-odd">Parameters</dt>
    965. 

  965. <dd class="field-odd"><p><strong>mask_metadata_tuple</strong> – A tuple of (coco_image_id, original_image_height, original_image_width)</p>
    966. 

  966. </dd>
    967. 

  967. <dt class="field-even">Returns</dt>
    968. 

  968. <dd class="field-even"><p>The mask image created from the array</p>
    969. 

  969. </dd>
    970. 

  970. </dl>
    971. 

  971. </dd></dl>
    972. 

  972. 

  973. </dd></dl>
    974. 

  974. 

  975. <dl class="py class">
    976. 

  976. <dt class="sig sig-object py" id="super_gradients.training.datasets.TestDatasetInterface">
    977. 

  977. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">TestDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">trainset</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</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/dataset_interfaces/dataset_interface.html#TestDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.TestDatasetInterface" title="Permalink to this definition"></a></dt>
    978. 

  978. <dd><p>Bases: <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"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></code></a></p>
    979. 

  979. <dl class="py method">
    980. 

  980. <dt class="sig sig-object py" id="super_gradients.training.datasets.TestDatasetInterface.get_data_loaders">
    981. 

  981. <span class="sig-name descname"><span class="pre">get_data_loaders</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size_factor</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">num_workers</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">train_batch_size</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">val_batch_size</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">distributed_sampler</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#TestDatasetInterface.get_data_loaders"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.TestDatasetInterface.get_data_loaders" title="Permalink to this definition"></a></dt>
    982. 

  982. <dd><p>Get self.train_loader, self.val_loader, self.test_loader, self.classes.</p>
    983. 

  983. <p>If the data loaders haven’t been initialized yet, build them first.</p>
    984. 

  984. <dl class="field-list simple">
    985. 

  985. <dt class="field-odd">Parameters</dt>
    986. 

  986. <dd class="field-odd"><p><strong>kwargs</strong> – kwargs are passed to build_data_loaders.</p>
    987. 

  987. </dd>
    988. 

  988. </dl>
    989. 

  989. </dd></dl>
    990. 

  990. 

  991. </dd></dl>
    992. 

  992. 

  993. <dl class="py class">
    994. 

  994. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface">
    995. 

  995. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">DatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">train_loader</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">val_loader</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">test_loader</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">classes</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/dataset_interfaces/dataset_interface.html#DatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface" title="Permalink to this definition"></a></dt>
    996. 

  996. <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
    997. 

  997. <p>DatasetInterface - This class manages all of the “communiation” the Model has with the Data Sets</p>
    998. 

  998. <dl class="py method">
    999. 

  999. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.download_from_cloud">
    1000. 

  1000. <span class="sig-name descname"><span class="pre">download_from_cloud</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#DatasetInterface.download_from_cloud"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.download_from_cloud" title="Permalink to this definition"></a></dt>
    1001. 

  1001. <dd></dd></dl>
    1002. 

  1002. 

  1003. <dl class="py method">
    1004. 

  1004. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.build_data_loaders">
    1005. 

  1005. <span class="sig-name descname"><span class="pre">build_data_loaders</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size_factor</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">num_workers</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">train_batch_size</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">val_batch_size</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">test_batch_size</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">distributed_sampler</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#DatasetInterface.build_data_loaders"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.build_data_loaders" title="Permalink to this definition"></a></dt>
    1006. 

  1006. <dd><p>define train, val (and optionally test) loaders. The method deals separately with distributed training and standard
    1007. 

  1007. (non distributed, or parallel training). In the case of distributed training we need to rely on distributed
    1008. 

  1008. samplers.
    1009. 

  1009. :param batch_size_factor: int - factor to multiply the batch size (usually for multi gpu)
    1010. 

  1010. :param num_workers: int - number of workers (parallel processes) for dataloaders
    1011. 

  1011. :param train_batch_size: int - batch size for train loader, if None will be taken from dataset_params
    1012. 

  1012. :param val_batch_size: int - batch size for val loader, if None will be taken from dataset_params
    1013. 

  1013. :param distributed_sampler: boolean flag for distributed training mode
    1014. 

  1014. :return: train_loader, val_loader, classes: list of classes</p>
    1015. 

  1015. </dd></dl>
    1016. 

  1016. 

  1017. <dl class="py method">
    1018. 

  1018. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.get_data_loaders">
    1019. 

  1019. <span class="sig-name descname"><span class="pre">get_data_loaders</span></span><span class="sig-paren">(</span><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/dataset_interfaces/dataset_interface.html#DatasetInterface.get_data_loaders"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.get_data_loaders" title="Permalink to this definition"></a></dt>
    1020. 

  1020. <dd><p>Get self.train_loader, self.val_loader, self.test_loader, self.classes.</p>
    1021. 

  1021. <p>If the data loaders haven’t been initialized yet, build them first.</p>
    1022. 

  1022. <dl class="field-list simple">
    1023. 

  1023. <dt class="field-odd">Parameters</dt>
    1024. 

  1024. <dd class="field-odd"><p><strong>kwargs</strong> – kwargs are passed to build_data_loaders.</p>
    1025. 

  1025. </dd>
    1026. 

  1026. </dl>
    1027. 

  1027. </dd></dl>
    1028. 

  1028. 

  1029. <dl class="py method">
    1030. 

  1030. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.get_val_sample">
    1031. 

  1031. <span class="sig-name descname"><span class="pre">get_val_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_samples</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#DatasetInterface.get_val_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.get_val_sample" title="Permalink to this definition"></a></dt>
    1032. 

  1032. <dd></dd></dl>
    1033. 

  1033. 

  1034. <dl class="py method">
    1035. 

  1035. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.get_dataset_params">
    1036. 

  1036. <span class="sig-name descname"><span class="pre">get_dataset_params</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#DatasetInterface.get_dataset_params"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.get_dataset_params" title="Permalink to this definition"></a></dt>
    1037. 

  1037. <dd></dd></dl>
    1038. 

  1038. 

  1039. <dl class="py method">
    1040. 

  1040. <dt class="sig sig-object py" id="super_gradients.training.datasets.DatasetInterface.print_dataset_details">
    1041. 

  1041. <span class="sig-name descname"><span class="pre">print_dataset_details</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#DatasetInterface.print_dataset_details"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DatasetInterface.print_dataset_details" title="Permalink to this definition"></a></dt>
    1042. 

  1042. <dd></dd></dl>
    1043. 

  1043. 

  1044. </dd></dl>
    1045. 

  1045. 

  1046. <dl class="py class">
    1047. 

  1047. <dt class="sig sig-object py" id="super_gradients.training.datasets.Cifar10DatasetInterface">
    1048. 

  1048. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">Cifar10DatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#Cifar10DatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.Cifar10DatasetInterface" title="Permalink to this definition"></a></dt>
    1049. 

  1049. <dd><p>Bases: <a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.LibraryDatasetInterface" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.LibraryDatasetInterface"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.LibraryDatasetInterface</span></code></a></p>
    1050. 

  1050. </dd></dl>
    1051. 

  1051. 

  1052. <dl class="py class">
    1053. 

  1053. <dt class="sig sig-object py" id="super_gradients.training.datasets.CoCoSegmentationDatasetInterface">
    1054. 

  1054. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">CoCoSegmentationDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_labels</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_images</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset_classes_inclusion_tuples_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span> <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/dataset_interfaces/dataset_interface.html#CoCoSegmentationDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.CoCoSegmentationDatasetInterface" title="Permalink to this definition"></a></dt>
    1055. 

  1055. <dd><p>Bases: <a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCoDataSetInterfaceBase" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCoDataSetInterfaceBase"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.CoCoDataSetInterfaceBase</span></code></a></p>
    1056. 

  1056. </dd></dl>
    1057. 

  1057. 

  1058. <dl class="py class">
    1059. 

  1059. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalVOC2012SegmentationDataSetInterface">
    1060. 

  1060. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">PascalVOC2012SegmentationDataSetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_labels</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_images</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#PascalVOC2012SegmentationDataSetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalVOC2012SegmentationDataSetInterface" title="Permalink to this definition"></a></dt>
    1061. 

  1061. <dd><p>Bases: <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"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></code></a></p>
    1062. 

  1062. </dd></dl>
    1063. 

  1063. 

  1064. <dl class="py class">
    1065. 

  1065. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalAUG2012SegmentationDataSetInterface">
    1066. 

  1066. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">PascalAUG2012SegmentationDataSetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_labels</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_images</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#PascalAUG2012SegmentationDataSetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalAUG2012SegmentationDataSetInterface" title="Permalink to this definition"></a></dt>
    1067. 

  1067. <dd><p>Bases: <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"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></code></a></p>
    1068. 

  1068. </dd></dl>
    1069. 

  1069. 

  1070. <dl class="py class">
    1071. 

  1071. <dt class="sig sig-object py" id="super_gradients.training.datasets.TestYoloDetectionDatasetInterface">
    1072. 

  1072. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">TestYoloDetectionDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_dims</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(3,</span> <span class="pre">32,</span> <span class="pre">32)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#TestYoloDetectionDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.TestYoloDetectionDatasetInterface" title="Permalink to this definition"></a></dt>
    1073. 

  1073. <dd><p>Bases: <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"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></code></a></p>
    1074. 

  1074. <p>note: the output size is (batch_size, 6) in the test while in real training
    1075. 

  1075. the size of axis 0 can vary (the number of bounding boxes)</p>
    1076. 

  1076. </dd></dl>
    1077. 

  1077. 

  1078. <dl class="py class">
    1079. 

  1079. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionTestDatasetInterface">
    1080. 

  1080. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">DetectionTestDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">320</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</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">classes</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/dataset_interfaces/dataset_interface.html#DetectionTestDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionTestDatasetInterface" title="Permalink to this definition"></a></dt>
    1081. 

  1081. <dd><p>Bases: <a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface</span></code></a></p>
    1082. 

  1082. </dd></dl>
    1083. 

  1083. 

  1084. <dl class="py class">
    1085. 

  1085. <dt class="sig sig-object py" id="super_gradients.training.datasets.ClassificationTestDatasetInterface">
    1086. 

  1086. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">ClassificationTestDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_size</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">batch_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</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/dataset_interfaces/dataset_interface.html#ClassificationTestDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.ClassificationTestDatasetInterface" title="Permalink to this definition"></a></dt>
    1087. 

  1087. <dd><p>Bases: <a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface</span></code></a></p>
    1088. 

  1088. </dd></dl>
    1089. 

  1089. 

  1090. <dl class="py class">
    1091. 

  1091. <dt class="sig sig-object py" id="super_gradients.training.datasets.SegmentationTestDatasetInterface">
    1092. 

  1092. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">SegmentationTestDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">512</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#SegmentationTestDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.SegmentationTestDatasetInterface" title="Permalink to this definition"></a></dt>
    1093. 

  1093. <dd><p>Bases: <a class="reference internal" href="super_gradients.training.datasets.dataset_interfaces.html#super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface" title="super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.TestDatasetInterface</span></code></a></p>
    1094. 

  1094. </dd></dl>
    1095. 

  1095. 

  1096. <dl class="py class">
    1097. 

  1097. <dt class="sig sig-object py" id="super_gradients.training.datasets.ImageNetDatasetInterface">
    1098. 

  1098. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">ImageNetDatasetInterface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset_params</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'/data/Imagenet'</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/dataset_interfaces/dataset_interface.html#ImageNetDatasetInterface"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.ImageNetDatasetInterface" title="Permalink to this definition"></a></dt>
    1099. 

  1099. <dd><p>Bases: <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"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.datasets.dataset_interfaces.dataset_interface.DatasetInterface</span></code></a></p>
    1100. 

  1100. </dd></dl>
    1101. 

  1101. 

  1102. <dl class="py class">
    1103. 

  1103. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset">
    1104. 

  1104. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">DetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_dim</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">original_target_format</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="super_gradients.training.utils.html#super_gradients.training.utils.detection_utils.DetectionTargetsFormat" title="super_gradients.training.utils.detection_utils.DetectionTargetsFormat"><span class="pre">super_gradients.training.utils.detection_utils.DetectionTargetsFormat</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_num_samples</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">super_gradients.training.transforms.transforms.DetectionTransform</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_classes_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_inclusion_list</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_empty_annotations</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_fields</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset" title="Permalink to this definition"></a></dt>
    1105. 

  1105. <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>
    1106. 

  1106. <p>Detection dataset.</p>
    1107. 

  1107. <p>This is a boilerplate class to facilitate the implementation of datasets.</p>
    1108. 

  1108. <dl>
    1109. 

  1109. <dt>HOW TO CREATE A DATASET THAT INHERITS FROM DetectionDataSet ?</dt><dd><ul class="simple">
    1110. 

  1110. <li><p>Inherit from DetectionDataSet</p></li>
    1111. 

  1111. <li><p>implement the method self._load_annotation to return at least the fields “target” and “img_path”</p></li>
    1112. 

  1112. <li><dl class="simple">
    1113. 

  1113. <dt>Call super().__init__ with the required params.</dt><dd><dl class="simple">
    1114. 

  1114. <dt>//!super().__init__ will call self._load_annotation, so make sure that every required</dt><dd><p>attributes are set up before calling super().__init__ (ideally just call it last)</p>
    1115. 

  1115. </dd>
    1116. 

  1116. </dl>
    1117. 

  1117. </dd>
    1118. 

  1118. </dl>
    1119. 

  1119. </li>
    1120. 

  1120. </ul>
    1121. 

  1121. </dd>
    1122. 

  1122. <dt>WORKFLOW:</dt><dd><ul class="simple">
    1123. 

  1123. <li><dl class="simple">
    1124. 

  1124. <dt>On instantiation:</dt><dd><ul>
    1125. 

  1125. <li><p>All annotations are cached. If class_inclusion_list was specified, there is also subclassing at this step.</p></li>
    1126. 

  1126. <li><p>If cache is True, the images are also cached</p></li>
    1127. 

  1127. </ul>
    1128. 

  1128. </dd>
    1129. 

  1129. </dl>
    1130. 

  1130. </li>
    1131. 

  1131. <li><dl class="simple">
    1132. 

  1132. <dt>On call (__getitem__) for a specific image index:</dt><dd><ul>
    1133. 

  1133. <li><p>The image and annotations are grouped together in a dict called SAMPLE</p></li>
    1134. 

  1134. <li><p>the sample is processed according to th transform</p></li>
    1135. 

  1135. <li><p>Only the specified fields are returned by __getitem__</p></li>
    1136. 

  1136. </ul>
    1137. 

  1137. </dd>
    1138. 

  1138. </dl>
    1139. 

  1139. </li>
    1140. 

  1140. </ul>
    1141. 

  1141. </dd>
    1142. 

  1142. <dt>TERMINOLOGY</dt><dd><ul>
    1143. 

  1143. <li><p>TARGET:       Groundtruth, made of bboxes. The format can vary from one dataset to another</p></li>
    1144. 

  1144. <li><dl class="simple">
    1145. 

  1145. <dt>ANNOTATION:   Combination of targets (groundtruth) and metadata of the image, but without the image itself.</dt><dd><p>&gt; Has to include the fields “target” and “img_path”
    1146. 

  1146. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    1147. 

  1147. </dd>
    1148. 

  1148. </dl>
    1149. 

  1149. </li>
    1150. 

  1150. <li><dl class="simple">
    1151. 

  1151. <dt>SAMPLE:       Outout of the dataset:</dt><dd><p>&gt; Has to include the fields “target” and “image”
    1152. 

  1152. &gt; Can include other fields like “crowd_target”, “image_info”, “segmentation”, …</p>
    1153. 

  1153. </dd>
    1154. 

  1154. </dl>
    1155. 

  1155. </li>
    1156. 

  1156. <li><p>INDEX:        Refers to the index in the dataset.</p></li>
    1157. 

  1157. <li><dl>
    1158. 

  1158. <dt>SAMPLE ID:    Refers to the id of sample before droping any annotaion.</dt><dd><p>Let’s imagine a situation where the downloaded data is made of 120 images but 20 were drop
    1159. 

  1159. because they had no annotation. In that case:</p>
    1160. 

  1160. <blockquote>
    1161. 

  1161. <div><p>&gt; We have 120 samples so sample_id will be between 0 and 119
    1162. 

  1162. &gt; But only 100 will be indexed so index will be between 0 and 99
    1163. 

  1163. &gt; Therefore, we also have len(self) = 100</p>
    1164. 

  1164. </div></blockquote>
    1165. 

  1165. </dd>
    1166. 

  1166. </dl>
    1167. 

  1167. </li>
    1168. 

  1168. </ul>
    1169. 

  1169. </dd>
    1170. 

  1170. </dl>
    1171. 

  1171. <dl class="py method">
    1172. 

  1172. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.get_random_item">
    1173. 

  1173. <span class="sig-name descname"><span class="pre">get_random_item</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_item"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.get_random_item" title="Permalink to this definition"></a></dt>
    1174. 

  1174. <dd></dd></dl>
    1175. 

  1175. 

  1176. <dl class="py method">
    1177. 

  1177. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.get_sample">
    1178. 

  1178. <span class="sig-name descname"><span class="pre">get_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.get_sample" title="Permalink to this definition"></a></dt>
    1179. 

  1179. <dd><p>Get raw sample, before any transform (beside subclassing).
    1180. 

  1180. :param index:   Image index
    1181. 

  1181. :return:        Sample, i.e. a dictionary including at least “image” and “target”</p>
    1182. 

  1182. </dd></dl>
    1183. 

  1183. 

  1184. <dl class="py method">
    1185. 

  1185. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.get_resized_image">
    1186. 

  1186. <span class="sig-name descname"><span class="pre">get_resized_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">numpy.ndarray</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_resized_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.get_resized_image" title="Permalink to this definition"></a></dt>
    1187. 

  1187. <dd><p>Get the resized image at a specific sample_id, either from cache or by loading from disk, based on self.cached_imgs
    1188. 

  1188. :param index:  Image index
    1189. 

  1189. :return:       Resized image</p>
    1190. 

  1190. </dd></dl>
    1191. 

  1191. 

  1192. <dl class="py method">
    1193. 

  1193. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.apply_transforms">
    1194. 

  1194. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    1195. 

  1195. <dd><p>Applies self.transforms sequentially to sample</p>
    1196. 

  1196. <dl class="simple">
    1197. 

  1197. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_annotations” will load
    1198. 

  1198. only additional samples with objects in them.</p>
    1199. 

  1199. </dd>
    1200. 

  1200. </dl>
    1201. 

  1201. <dl class="field-list simple">
    1202. 

  1202. <dt class="field-odd">Parameters</dt>
    1203. 

  1203. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.get_sample)</p>
    1204. 

  1204. </dd>
    1205. 

  1205. <dt class="field-even">Returns</dt>
    1206. 

  1206. <dd class="field-even"><p>Transformed sample</p>
    1207. 

  1207. </dd>
    1208. 

  1208. </dl>
    1209. 

  1209. </dd></dl>
    1210. 

  1210. 

  1211. <dl class="py method">
    1212. 

  1212. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.get_random_samples">
    1213. 

  1213. <span class="sig-name descname"><span class="pre">get_random_samples</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">count</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> &#x2192; <span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_samples"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.get_random_samples" title="Permalink to this definition"></a></dt>
    1214. 

  1214. <dd><p>Load random samples.</p>
    1215. 

  1215. <dl class="field-list simple">
    1216. 

  1216. <dt class="field-odd">Parameters</dt>
    1217. 

  1217. <dd class="field-odd"><ul class="simple">
    1218. 

  1218. <li><p><strong>count</strong> – The number of samples wanted</p></li>
    1219. 

  1219. <li><p><strong>non_empty_annotations_only</strong> – If true, only return samples with at least 1 annotation</p></li>
    1220. 

  1220. </ul>
    1221. 

  1221. </dd>
    1222. 

  1222. <dt class="field-even">Returns</dt>
    1223. 

  1223. <dd class="field-even"><p>A list of samples satisfying input params</p>
    1224. 

  1224. </dd>
    1225. 

  1225. </dl>
    1226. 

  1226. </dd></dl>
    1227. 

  1227. 

  1228. <dl class="py method">
    1229. 

  1229. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.get_random_sample">
    1230. 

  1230. <span class="sig-name descname"><span class="pre">get_random_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">non_empty_annotations_only</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.get_random_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.get_random_sample" title="Permalink to this definition"></a></dt>
    1231. 

  1231. <dd></dd></dl>
    1232. 

  1232. 

  1233. <dl class="py property">
    1234. 

  1234. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.output_target_format">
    1235. 

  1235. <em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">output_target_format</span></span><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.output_target_format" title="Permalink to this definition"></a></dt>
    1236. 

  1236. <dd></dd></dl>
    1237. 

  1237. 

  1238. <dl class="py method">
    1239. 

  1239. <dt class="sig sig-object py" id="super_gradients.training.datasets.DetectionDataset.plot">
    1240. 

  1240. <span class="sig-name descname"><span class="pre">plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_samples_per_plot</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_plots</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_transformed_data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/detection_dataset.html#DetectionDataset.plot"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.DetectionDataset.plot" title="Permalink to this definition"></a></dt>
    1241. 

  1241. <dd><p>Combine samples of images with bbox into plots and display the result.</p>
    1242. 

  1242. <dl class="field-list simple">
    1243. 

  1243. <dt class="field-odd">Parameters</dt>
    1244. 

  1244. <dd class="field-odd"><ul class="simple">
    1245. 

  1245. <li><p><strong>max_samples_per_plot</strong> – Maximum number of images to be displayed per plot</p></li>
    1246. 

  1246. <li><p><strong>n_plots</strong> – Number of plots to display (each plot being a combination of img with bbox)</p></li>
    1247. 

  1247. <li><p><strong>plot_transformed_data</strong> – If True, the plot will be over samples after applying transforms (i.e. on __getitem__).
    1248. 

  1248. If False, the plot will be over the raw samples (i.e. on get_sample)</p></li>
    1249. 

  1249. </ul>
    1250. 

  1250. </dd>
    1251. 

  1251. <dt class="field-even">Returns</dt>
    1252. 

  1252. <dd class="field-even"><p></p>
    1253. 

  1253. </dd>
    1254. 

  1254. </dl>
    1255. 

  1255. </dd></dl>
    1256. 

  1256. 

  1257. </dd></dl>
    1258. 

  1258. 

  1259. <dl class="py class">
    1260. 

  1260. <dt class="sig sig-object py" id="super_gradients.training.datasets.COCODetectionDataset">
    1261. 

  1261. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">COCODetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_size</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">tuple</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">json_file</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'instances_train2017.json'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'images/train2017'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache_dir_path</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tight_box_rotation</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transforms</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">list</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_crowd</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.COCODetectionDataset" title="Permalink to this definition"></a></dt>
    1262. 

  1262. <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>
    1263. 

  1263. <p>Detection dataset COCO implementation</p>
    1264. 

  1264. <dl class="py method">
    1265. 

  1265. <dt class="sig sig-object py" id="super_gradients.training.datasets.COCODetectionDataset.load_resized_img">
    1266. 

  1266. <span class="sig-name descname"><span class="pre">load_resized_img</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_resized_img"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.COCODetectionDataset.load_resized_img" title="Permalink to this definition"></a></dt>
    1267. 

  1267. <dd><p>Loads image at index, and resizes it to self.input_dim</p>
    1268. 

  1268. <dl class="field-list simple">
    1269. 

  1269. <dt class="field-odd">Parameters</dt>
    1270. 

  1270. <dd class="field-odd"><p><strong>index</strong> – index to load the image from</p>
    1271. 

  1271. </dd>
    1272. 

  1272. <dt class="field-even">Returns</dt>
    1273. 

  1273. <dd class="field-even"><p>resized_img</p>
    1274. 

  1274. </dd>
    1275. 

  1275. </dl>
    1276. 

  1276. </dd></dl>
    1277. 

  1277. 

  1278. <dl class="py method">
    1279. 

  1279. <dt class="sig sig-object py" id="super_gradients.training.datasets.COCODetectionDataset.load_sample">
    1280. 

  1280. <span class="sig-name descname"><span class="pre">load_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_sample"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.COCODetectionDataset.load_sample" title="Permalink to this definition"></a></dt>
    1281. 

  1281. <dd><dl>
    1282. 

  1282. <dt>Loads sample at self.ids[index] as dictionary that holds:</dt><dd><p>“image”: Image resized to self.input_dim
    1283. 

  1283. “target”: Detection ground truth, np.array shaped (num_targets, 5), format is [class,x1,y1,x2,y2] with</p>
    1284. 

  1284. <blockquote>
    1285. 

  1285. <div><p>image coordinates.</p>
    1286. 

  1286. </div></blockquote>
    1287. 

  1287. <p>“target_seg”: Segmentation map convex hull derived detection target.
    1288. 

  1288. “info”: Original shape (height,width).
    1289. 

  1289. “id”: COCO image id</p>
    1290. 

  1290. </dd>
    1291. 

  1291. </dl>
    1292. 

  1292. <dl class="field-list simple">
    1293. 

  1293. <dt class="field-odd">Parameters</dt>
    1294. 

  1294. <dd class="field-odd"><p><strong>index</strong> – Sample index</p>
    1295. 

  1295. </dd>
    1296. 

  1296. <dt class="field-even">Returns</dt>
    1297. 

  1297. <dd class="field-even"><p>sample as described above</p>
    1298. 

  1298. </dd>
    1299. 

  1299. </dl>
    1300. 

  1300. </dd></dl>
    1301. 

  1301. 

  1302. <dl class="py method">
    1303. 

  1303. <dt class="sig sig-object py" id="super_gradients.training.datasets.COCODetectionDataset.load_image">
    1304. 

  1304. <span class="sig-name descname"><span class="pre">load_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.load_image"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.COCODetectionDataset.load_image" title="Permalink to this definition"></a></dt>
    1305. 

  1305. <dd><p>Loads image at index with its original resolution
    1306. 

  1306. :param index: index in self.annotations
    1307. 

  1307. :return: image (np.array)</p>
    1308. 

  1308. </dd></dl>
    1309. 

  1309. 

  1310. <dl class="py method">
    1311. 

  1311. <dt class="sig sig-object py" id="super_gradients.training.datasets.COCODetectionDataset.apply_transforms">
    1312. 

  1312. <span class="sig-name descname"><span class="pre">apply_transforms</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/coco_detection.html#COCODetectionDataset.apply_transforms"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.COCODetectionDataset.apply_transforms" title="Permalink to this definition"></a></dt>
    1313. 

  1313. <dd><p>Applies self.transforms sequentially to sample</p>
    1314. 

  1314. <dl class="simple">
    1315. 

  1315. <dt>If a transforms has the attribute ‘additional_samples_count’, additional samples will be loaded and stored in</dt><dd><p>sample[“additional_samples”] prior to applying it. Combining with the attribute “non_empty_targets” will load
    1316. 

  1316. only additional samples with objects in them.</p>
    1317. 

  1317. </dd>
    1318. 

  1318. </dl>
    1319. 

  1319. <dl class="field-list simple">
    1320. 

  1320. <dt class="field-odd">Parameters</dt>
    1321. 

  1321. <dd class="field-odd"><p><strong>sample</strong> – Sample to apply the transforms on to (loaded with self.load_sample)</p>
    1322. 

  1322. </dd>
    1323. 

  1323. <dt class="field-even">Returns</dt>
    1324. 

  1324. <dd class="field-even"><p>Transformed sample</p>
    1325. 

  1325. </dd>
    1326. 

  1326. </dl>
    1327. 

  1327. </dd></dl>
    1328. 

  1328. 

  1329. </dd></dl>
    1330. 

  1330. 

  1331. <dl class="py class">
    1332. 

  1332. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalVOCDetectionDataset">
    1333. 

  1333. <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.datasets.</span></span><span class="sig-name descname"><span class="pre">PascalVOCDetectionDataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">images_sub_directory</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalVOCDetectionDataset" title="Permalink to this definition"></a></dt>
    1334. 

  1334. <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>
    1335. 

  1335. <p>Dataset for Pascal VOC object detection</p>
    1336. 

  1336. <dl class="py method">
    1337. 

  1337. <dt class="sig sig-object py" id="super_gradients.training.datasets.PascalVOCDetectionDataset.download">
    1338. 

  1338. <em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">download</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_dir</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/datasets/detection_datasets/pascal_voc_detection.html#PascalVOCDetectionDataset.download"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.datasets.PascalVOCDetectionDataset.download" title="Permalink to this definition"></a></dt>
    1339. 

  1339. <dd><p>Download Pascal dataset in XYXY_LABEL format.</p>
    1340. 

  1340. <p>Data extracted form <a class="reference external" href="http://host.robots.ox.ac.uk/pascal/VOC/">http://host.robots.ox.ac.uk/pascal/VOC/</a></p>
    1341. 

  1341. </dd></dl>
    1342. 

  1342. 

  1343. </dd></dl>
    1344. 

  1344. 

  1345. </section>
    1346. 

  1346. </section>
    1347. 

  1347. 

  1348. 

  1349.            </div>
    1350. 

  1350.           </div>
    1351. 

  1351.           <footer>
    1352. 

  1352. 

  1353.   <hr/>
    1354. 

  1354. 

  1355.   <div role="contentinfo">
    1356. 

  1356.     <p>&#169; Copyright 2021, SuperGradients team.</p>
    1357. 

  1357.   </div>
    1358. 

  1358. 

  1359.   Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    1360. 

  1360.     <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    1361. 

  1361.     provided by <a href="https://readthedocs.org">Read the Docs</a>.
    1362. 

  1362.    
    1363. 

  1363. 

  1364. </footer>
    1365. 

  1365.         </div>
    1366. 

  1366.       </div>
    1367. 

  1367.     </section>
    1368. 

  1368.   </div>
    1369. 

  1369.   <script>
    1370. 

  1370.       jQuery(function () {
    1371. 

  1371.           SphinxRtdTheme.Navigation.enable(true);
    1372. 

  1372.       });
    1373. 

  1373.   </script> 
    1374. 

  1374. 

  1375. </body>
    1376. 

  1376. </html>
    1377.
Discard
Prev
Next
Tip!

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