-
-
Notifications
You must be signed in to change notification settings - Fork 636
/
Copy pathengine.html
960 lines (766 loc) · 56.9 KB
/
engine.html
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
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="author" content="PyTorch-Ignite Contributors">
<meta name="creator" content="PyTorch-Ignite Contributors">
<meta name="publisher" content="PyTorch-Ignite Contributors">
<meta name="generator" content="Sphinx 5.3.0">
<meta name="description" content="High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.">
<meta name="keywords" content="pytorch, pytorch ignite, ignite, engine, events, handlers, metrics">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="theme-color" content="#ee4c2c">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:site" content="@pytorch_ignite">
<meta name="twitter:creator" content="@pytorch_ignite">
<meta name="twitter:description" content="High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.">
<meta name="twitter:title" content="ignite.engine — PyTorch-Ignite v0.5.1 Documentation">
<meta name="twitter:image" content="https://raw.githubusercontent.com/pytorch/ignite/master/assets/logo/ignite_logo.png">
<meta name="twitter:image:alt" content="PyTorch-Ignite logo">
<meta property="og:title" content="ignite.engine — PyTorch-Ignite v0.5.1 Documentation">
<meta property="og:type" content="website">
<meta property="og:url" content="https://pytorch.org/ignite/">
<meta property="og:site_name" content="PyTorch-Ignite">
<meta property="og:image" content="https://raw.githubusercontent.com/pytorch/ignite/master/assets/logo/ignite_logo.png">
<meta property="og:image:alt" content="PyTorch-Ignite logo">
<meta property="og:description" content="High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.">
<link rel="preconnect" href="https://BH4D9OD16A-dsn.algolia.net" crossorigin />
<title>ignite.engine — PyTorch-Ignite v0.5.1 Documentation</title>
<link rel="shortcut icon" type="image/svg+xml" href="_static/ignite_logomark.svg"/>
<link rel="canonical" href="https://pytorch.org/ignite/engine.html"/>
<link rel="preload" href="_static/css/theme.css" as="style" />
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
<!-- <link rel="stylesheet" href="_static/pygments.css" type="text/css" /> -->
<link rel="preload" href="_static/pygments.css" as="style" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<link rel="preload" href="_static/css/theme.css" as="style" />
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
<link rel="preload" href="_static/copybutton.css" as="style" />
<link rel="stylesheet" href="_static/copybutton.css" type="text/css" />
<link rel="preload" href="_static/togglebutton.css" as="style" />
<link rel="stylesheet" href="_static/togglebutton.css" type="text/css" />
<link rel="preload" href="_static/banner.css" as="style" />
<link rel="stylesheet" href="_static/banner.css" type="text/css" />
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" as="style" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" type="text/css" />
<link rel="preload" href="_static/katex-math.css" as="style" />
<link rel="stylesheet" href="_static/katex-math.css" type="text/css" />
<link rel="preload" href="_static/sphinx-design.5ea377869091fd0449014c60fc090103.min.css" as="style" />
<link rel="stylesheet" href="_static/sphinx-design.5ea377869091fd0449014c60fc090103.min.css" type="text/css" />
<link rel="preload" href="_static/css/ignite_theme.css" as="style" />
<link rel="stylesheet" href="_static/css/ignite_theme.css" type="text/css" />
<link rel="preload" href="https://cdn.jsdelivr.net/npm/@docsearch/css@3" as="style" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@docsearch/css@3" type="text/css" />
<link rel="index" title="Index" href="genindex.html" />
<link rel="search" title="Search" href="search.html" />
<link rel="next" title="Engine" href="generated/ignite.engine.engine.Engine.html" />
<link rel="prev" title="Ignite Your Networks!" href="index.html" />
<!-- katex links / this needs to be loaded before fonts.html -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" crossorigin="anonymous">
<script async src="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.js" crossorigin="anonymous"></script>
<!-- Preload the theme fonts -->
<link rel="preload" href="_static/fonts/FreightSans/freight-sans-book.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/FreightSans/freight-sans-medium.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/IBMPlexMono/IBMPlexMono-Medium.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/FreightSans/freight-sans-bold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/FreightSans/freight-sans-medium-italic.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/IBMPlexMono/IBMPlexMono-SemiBold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="_static/fonts/IBMPlexMono/IBMPlexMono-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<!-- Preload the katex fonts -->
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Math-Italic.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Main-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Main-Bold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Size1-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Size4-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Size2-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Size3-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/fonts/KaTeX_Caligraphic-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-5FELLLFHCP"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-5FELLLFHCP');
</script>
</head>
<div class="container-fluid header-holder tutorials-header" id="header-holder">
<div class="container">
<a class="header-logo" href="/ignite" aria-label="PyTorch-Ignite"></a>
<div class="header-container">
<div class="main-menu">
<ul>
<li>
<a href="https://pytorch-ignite.ai/tutorials/beginner/01-getting-started/">Quickstart</a>
</li>
<li>
<a href="https://pytorch-ignite.ai/concepts/">Concepts</a>
</li>
<!-- <li>
<a href="https://pytorch-ignite.ai/tutorials/beginner/01-getting-started/#complete-code">Examples</a>
</li> -->
<li>
<a href="https://pytorch-ignite.ai/how-to-guides/">FAQ</a>
</li>
<li>
<a href="https://github.com/pytorch/ignite" target="_blank" rel="noopener noreferrer">GitHub</a>
</li>
<li>
<a href="https://pytorch-ignite.ai/about/community/">About us</a>
</li>
<li>
<a href="https://pytorch-ignite.ai">⊳ pytorch-ignite.ai</a>
</li>
</ul>
</div>
<!-- <a class="main-menu-open-button" href="#" data-behavior="open-mobile-menu"></a> -->
</div>
</div>
</div>
<body class="pytorch-body">
<div class="table-of-contents-link-wrapper">
<span>Table of Contents</span>
<div class="toggle-table-of-contents" data-behavior="toggle-table-of-contents"></div>
</div>
<nav data-toggle="wy-nav-shift" class="pytorch-left-menu" id="pytorch-left-menu">
<div class="pytorch-side-scroll">
<div class="pytorch-menu pytorch-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<div class="pytorch-left-menu-search">
<div class="version">
v0.5.1
</div>
<div id="docsearch"></div>
</div>
<p class="caption" role="heading"><span class="caption-text">Package Reference</span></p>
<ul class="current">
<li class="toctree-l1 current"><a class="current reference internal" href="#">ignite.engine</a></li>
<li class="toctree-l1"><a class="reference internal" href="handlers.html">ignite.handlers</a></li>
<li class="toctree-l1"><a class="reference internal" href="metrics.html">ignite.metrics</a></li>
<li class="toctree-l1"><a class="reference internal" href="distributed.html">ignite.distributed</a></li>
<li class="toctree-l1"><a class="reference internal" href="exceptions.html">ignite.exceptions</a></li>
<li class="toctree-l1"><a class="reference internal" href="utils.html">ignite.utils</a></li>
</ul>
<p class="caption" role="heading"><span class="caption-text">Contrib Package Reference</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="contrib/engines.html">ignite.contrib.engines</a></li>
<li class="toctree-l1"><a class="reference internal" href="contrib/metrics.html">ignite.contrib.metrics</a></li>
<li class="toctree-l1"><a class="reference internal" href="contrib/handlers.html">ignite.contrib.handlers</a></li>
</ul>
</div>
</div>
<div class="rst-versions" data-toggle="rst-versions" role="note" aria-label="versions">
<span class="rst-current-version" data-toggle="rst-current-version">
<span class="fa fa-book"> Other Versions</span>
v: v0.5.1
<span class="fa fa-caret-down"></span>
</span>
<div class="rst-other-versions">
<dl>
<dt>Tags</dt>
<dd><a href="v0.3.0/engine.html">v0.3.0</a></dd>
<dd><a href="v0.4.0.post1/engine.html">v0.4.0.post1</a></dd>
<dd><a href="v0.4.1/engine.html">v0.4.1</a></dd>
<dd><a href="v0.4.10/engine.html">v0.4.10</a></dd>
<dd><a href="v0.4.11/engine.html">v0.4.11</a></dd>
<dd><a href="v0.4.12/engine.html">v0.4.12</a></dd>
<dd><a href="v0.4.13/engine.html">v0.4.13</a></dd>
<dd><a href="v0.4.2/engine.html">v0.4.2</a></dd>
<dd><a href="v0.4.3/engine.html">v0.4.3</a></dd>
<dd><a href="v0.4.4.post1/engine.html">v0.4.4.post1</a></dd>
<dd><a href="v0.4.5/engine.html">v0.4.5</a></dd>
<dd><a href="v0.4.6/engine.html">v0.4.6</a></dd>
<dd><a href="v0.4.7/engine.html">v0.4.7</a></dd>
<dd><a href="v0.4.8/engine.html">v0.4.8</a></dd>
<dd><a href="v0.4.9/engine.html">v0.4.9</a></dd>
<dd><a href="v0.4rc.0.post1/engine.html">v0.4rc.0.post1</a></dd>
<dd><a href="v0.5.0.post2/engine.html">v0.5.0.post2</a></dd>
<dd><a href="v0.5.1/engine.html">v0.5.1</a></dd>
</dl>
<dl>
<dt>Branches</dt>
<dd><a href="master/engine.html">master</a></dd>
</dl>
</div>
</div>
</nav>
<div class="pytorch-container">
<div class="pytorch-page-level-bar" id="pytorch-page-level-bar">
<div class="pytorch-breadcrumbs-wrapper">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="pytorch-breadcrumbs">
<li>
<a href="index.html">
Docs
</a> >
</li>
<li>ignite.engine</li>
<li class="pytorch-breadcrumbs-aside">
<a href="_sources/engine.rst.txt" rel="nofollow"><img src="_static/images/view-page-source-icon.svg"></a>
</li>
</ul>
</div>
</div>
<div class="pytorch-shortcuts-wrapper" id="pytorch-shortcuts-wrapper">
Shortcuts
</div>
</div>
<section data-toggle="wy-nav-shift" id="pytorch-content-wrap" class="pytorch-content-wrap">
<div class="pytorch-content-left">
<div class="rst-content">
<div role="main" class="main-content" itemscope="itemscope" itemtype="http://schema.org/Article">
<article itemprop="articleBody" id="pytorch-article" class="pytorch-article">
<section id="ignite-engine">
<h1>ignite.engine<a class="headerlink" href="#ignite-engine" title="Permalink to this heading">#</a></h1>
<p>Main module of the library containing:</p>
<section id="ignite-engine-engine">
<h2>ignite.engine.engine<a class="headerlink" href="#ignite-engine-engine" title="Permalink to this heading">#</a></h2>
<table class="autosummary longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.engine.Engine.html#ignite.engine.engine.Engine" title="ignite.engine.engine.Engine"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Engine</span></code></a></p></td>
<td><p>Runs a given <code class="docutils literal notranslate"><span class="pre">process_function</span></code> over each batch of a dataset, emitting events as it goes.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="ignite-engine-events">
<h2>ignite.engine.events<a class="headerlink" href="#ignite-engine-events" title="Permalink to this heading">#</a></h2>
<table class="autosummary longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.events.CallableEventWithFilter.html#ignite.engine.events.CallableEventWithFilter" title="ignite.engine.events.CallableEventWithFilter"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CallableEventWithFilter</span></code></a></p></td>
<td><p>Single Event containing a filter, specifying whether the event should be run at the current event (if the event type is correct)</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.events.EventEnum.html#ignite.engine.events.EventEnum" title="ignite.engine.events.EventEnum"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EventEnum</span></code></a></p></td>
<td><p>Base class for all <a class="reference internal" href="generated/ignite.engine.events.Events.html#ignite.engine.events.Events" title="ignite.engine.events.Events"><code class="xref py py-class docutils literal notranslate"><span class="pre">Events</span></code></a>.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.events.Events.html#ignite.engine.events.Events" title="ignite.engine.events.Events"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Events</span></code></a></p></td>
<td><p>Events that are fired by the <a class="reference internal" href="generated/ignite.engine.engine.Engine.html#ignite.engine.engine.Engine" title="ignite.engine.engine.Engine"><code class="xref py py-class docutils literal notranslate"><span class="pre">Engine</span></code></a> during execution.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.events.EventsList.html#ignite.engine.events.EventsList" title="ignite.engine.events.EventsList"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EventsList</span></code></a></p></td>
<td><p>Collection of events stacked by operator <cite>__or__</cite>.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.events.State.html#ignite.engine.events.State" title="ignite.engine.events.State"><code class="xref py py-obj docutils literal notranslate"><span class="pre">State</span></code></a></p></td>
<td><p>An object that is used to pass internal and user-defined state between event handlers.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.events.RemovableEventHandle.html#ignite.engine.events.RemovableEventHandle" title="ignite.engine.events.RemovableEventHandle"><code class="xref py py-obj docutils literal notranslate"><span class="pre">RemovableEventHandle</span></code></a></p></td>
<td><p>A weakref handle to remove a registered event.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="ignite-engine-deterministic">
<h2>ignite.engine.deterministic<a class="headerlink" href="#ignite-engine-deterministic" title="Permalink to this heading">#</a></h2>
<p>Helper methods for deterministic training</p>
<table class="autosummary longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.deterministic.DeterministicEngine.html#ignite.engine.deterministic.DeterministicEngine" title="ignite.engine.deterministic.DeterministicEngine"><code class="xref py py-obj docutils literal notranslate"><span class="pre">DeterministicEngine</span></code></a></p></td>
<td><p>Deterministic engine derived from <a class="reference internal" href="generated/ignite.engine.engine.Engine.html#ignite.engine.engine.Engine" title="ignite.engine.engine.Engine"><code class="xref py py-class docutils literal notranslate"><span class="pre">Engine</span></code></a>.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.deterministic.ReproducibleBatchSampler.html#ignite.engine.deterministic.ReproducibleBatchSampler" title="ignite.engine.deterministic.ReproducibleBatchSampler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ReproducibleBatchSampler</span></code></a></p></td>
<td><p>Reproducible batch sampler.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.deterministic.keep_random_state.html#ignite.engine.deterministic.keep_random_state" title="ignite.engine.deterministic.keep_random_state"><code class="xref py py-obj docutils literal notranslate"><span class="pre">keep_random_state</span></code></a></p></td>
<td><p>Helper decorator to keep random state of torch, numpy and random intact while executing a function.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.deterministic.update_dataloader.html#ignite.engine.deterministic.update_dataloader" title="ignite.engine.deterministic.update_dataloader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_dataloader</span></code></a></p></td>
<td><p>Helper function to replace current batch sampler of the dataloader by a new batch sampler.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="helper-methods-to-define-supervised-trainer-and-evaluator">
<h2>helper methods to define supervised trainer and evaluator<a class="headerlink" href="#helper-methods-to-define-supervised-trainer-and-evaluator" title="Permalink to this heading">#</a></h2>
<table class="autosummary longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.create_supervised_trainer.html#ignite.engine.create_supervised_trainer" title="ignite.engine.create_supervised_trainer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">create_supervised_trainer</span></code></a></p></td>
<td><p>Factory function for creating a trainer for supervised models.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.create_supervised_evaluator.html#ignite.engine.create_supervised_evaluator" title="ignite.engine.create_supervised_evaluator"><code class="xref py py-obj docutils literal notranslate"><span class="pre">create_supervised_evaluator</span></code></a></p></td>
<td><p>Factory function for creating an evaluator for supervised models.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_training_step.html#ignite.engine.supervised_training_step" title="ignite.engine.supervised_training_step"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_training_step</span></code></a></p></td>
<td><p>Factory function for supervised training.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_training_step_amp.html#ignite.engine.supervised_training_step_amp" title="ignite.engine.supervised_training_step_amp"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_training_step_amp</span></code></a></p></td>
<td><p>Factory function for supervised training using <code class="docutils literal notranslate"><span class="pre">torch.cuda.amp</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_training_step_apex.html#ignite.engine.supervised_training_step_apex" title="ignite.engine.supervised_training_step_apex"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_training_step_apex</span></code></a></p></td>
<td><p>Factory function for supervised training using apex.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_training_step_tpu.html#ignite.engine.supervised_training_step_tpu" title="ignite.engine.supervised_training_step_tpu"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_training_step_tpu</span></code></a></p></td>
<td><p>Factory function for supervised training using <code class="docutils literal notranslate"><span class="pre">torch_xla</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_evaluation_step.html#ignite.engine.supervised_evaluation_step" title="ignite.engine.supervised_evaluation_step"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_evaluation_step</span></code></a></p></td>
<td><p>Factory function for supervised evaluation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/ignite.engine.supervised_evaluation_step_amp.html#ignite.engine.supervised_evaluation_step_amp" title="ignite.engine.supervised_evaluation_step_amp"><code class="xref py py-obj docutils literal notranslate"><span class="pre">supervised_evaluation_step_amp</span></code></a></p></td>
<td><p>Factory function for supervised evaluation using <code class="docutils literal notranslate"><span class="pre">torch.cuda.amp</span></code>.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="resuming-the-training">
<h2>Resuming the training<a class="headerlink" href="#resuming-the-training" title="Permalink to this heading">#</a></h2>
<p>It is possible to resume the training from a checkpoint and approximately reproduce original run’s behaviour.
Using Ignite, this can be easily done using <a class="reference internal" href="generated/ignite.handlers.checkpoint.Checkpoint.html#ignite.handlers.checkpoint.Checkpoint" title="ignite.handlers.checkpoint.Checkpoint"><code class="xref py py-class docutils literal notranslate"><span class="pre">Checkpoint</span></code></a> handler. Engine provides two methods
to serialize and deserialize its internal state <a class="reference internal" href="generated/ignite.engine.engine.Engine.html#ignite.engine.engine.Engine.state_dict" title="ignite.engine.engine.Engine.state_dict"><code class="xref py py-meth docutils literal notranslate"><span class="pre">state_dict()</span></code></a> and
<a class="reference internal" href="generated/ignite.engine.engine.Engine.html#ignite.engine.engine.Engine.load_state_dict" title="ignite.engine.engine.Engine.load_state_dict"><code class="xref py py-meth docutils literal notranslate"><span class="pre">load_state_dict()</span></code></a>. In addition to serializing model, optimizer, lr scheduler, metrics, etc., user can
store the trainer and then resume the training. For example:</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span><span class="w"> </span><span class="nn">ignite.engine</span><span class="w"> </span><span class="kn">import</span> <span class="n">Engine</span><span class="p">,</span> <span class="n">Events</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">ignite.handlers</span><span class="w"> </span><span class="kn">import</span> <span class="n">Checkpoint</span><span class="p">,</span> <span class="n">DiskSaver</span>
<span class="n">trainer</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">model</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">optimizer</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">lr_scheduler</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">data_loader</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">metric</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">to_save</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'trainer'</span><span class="p">:</span> <span class="n">trainer</span><span class="p">,</span> <span class="s1">'model'</span><span class="p">:</span> <span class="n">model</span><span class="p">,</span> <span class="s1">'optimizer'</span><span class="p">:</span> <span class="n">optimizer</span><span class="p">,</span> <span class="s1">'lr_scheduler'</span><span class="p">:</span> <span class="n">lr_scheduler</span><span class="p">,</span> <span class="s1">'metric'</span><span class="p">:</span> <span class="n">metric</span><span class="p">}</span>
<span class="n">handler</span> <span class="o">=</span> <span class="n">Checkpoint</span><span class="p">(</span><span class="n">to_save</span><span class="p">,</span> <span class="n">DiskSaver</span><span class="p">(</span><span class="s1">'/tmp/training'</span><span class="p">,</span> <span class="n">create_dir</span><span class="o">=</span><span class="kc">True</span><span class="p">))</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">add_event_handler</span><span class="p">(</span><span class="n">Events</span><span class="o">.</span><span class="n">EPOCH_COMPLETED</span><span class="p">,</span> <span class="n">handler</span><span class="p">)</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">data_loader</span><span class="p">,</span> <span class="n">max_epochs</span><span class="o">=</span><span class="mi">100</span><span class="p">)</span>
</pre></div>
</div>
<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>ls<span class="w"> </span>/tmp/training
><span class="w"> </span><span class="s2">"checkpoint_50000.pt"</span>
</pre></div>
</div>
<p>We can then restore the training from the last checkpoint.</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span><span class="w"> </span><span class="nn">ignite.handlers</span><span class="w"> </span><span class="kn">import</span> <span class="n">Checkpoint</span>
<span class="n">trainer</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">model</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">optimizer</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">lr_scheduler</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">data_loader</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">metric</span> <span class="o">=</span> <span class="o">...</span>
<span class="n">to_load</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'trainer'</span><span class="p">:</span> <span class="n">trainer</span><span class="p">,</span> <span class="s1">'model'</span><span class="p">:</span> <span class="n">model</span><span class="p">,</span> <span class="s1">'optimizer'</span><span class="p">:</span> <span class="n">optimizer</span><span class="p">,</span> <span class="s1">'lr_scheduler'</span><span class="p">:</span> <span class="n">lr_scheduler</span><span class="p">,</span> <span class="s1">'metric'</span><span class="p">:</span> <span class="n">metric</span><span class="p">}</span>
<span class="n">checkpoint</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">checkpoint_file</span><span class="p">)</span>
<span class="n">Checkpoint</span><span class="o">.</span><span class="n">load_objects</span><span class="p">(</span><span class="n">to_load</span><span class="o">=</span><span class="n">to_load</span><span class="p">,</span> <span class="n">checkpoint</span><span class="o">=</span><span class="n">checkpoint</span><span class="p">)</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">train_loader</span><span class="p">,</span> <span class="n">max_epochs</span><span class="o">=</span><span class="mi">100</span><span class="p">)</span>
</pre></div>
</div>
<p>It is also possible to store checkpoints every N iterations and continue the training from one of these checkpoints, i.e
from iteration.</p>
<p>Complete examples that resumes the training from a checkpoint can be found here:</p>
<ul class="simple">
<li><p><a class="reference external" href="https://github.com/pytorch/ignite/tree/master/examples/mnist#training-save--resume">save/resume MNIST</a></p></li>
<li><p><a class="reference external" href="https://github.com/pytorch/ignite/tree/master/examples/cifar10#check-resume-training">save/resume Distributed CIFAR10</a></p></li>
</ul>
</section>
<section id="deterministic-training">
<h2>Deterministic training<a class="headerlink" href="#deterministic-training" title="Permalink to this heading">#</a></h2>
<p>In general, it is rather difficult task to achieve deterministic and reproducible trainings as it relies on multiple
aspects, e.g. data version, code version, software environment, hardware etc. According to <a class="reference external" href="https://pytorch.org/docs/stable/notes/randomness.html">PyTorch note on randomness</a>:
there are some steps to take in order to make computations deterministic on your specific problem on one specific
platform and PyTorch release:</p>
<ul class="simple">
<li><p>setup random state seed</p></li>
<li><p>set <a class="reference external" href="https://pytorch.org/docs/stable/notes/randomness.html#cuda-convolution-benchmarking">cudnn to deterministic</a> if applicable</p></li>
</ul>
<p>By default, these two options can be enough to run and rerun experiments in a deterministic way. Ignite’s engine does not impact this behaviour.</p>
<p>In this module we provide helper methods and classes to make additional “<a class="reference internal" href="#dataflow-synchronization"><span class="std std-ref">Dataflow synchronization</span></a>”
to ensure that model sees the same data for a given epoch:</p>
<ul class="simple">
<li><p><a class="reference internal" href="generated/ignite.engine.deterministic.DeterministicEngine.html#ignite.engine.deterministic.DeterministicEngine" title="ignite.engine.deterministic.DeterministicEngine"><code class="xref py py-class docutils literal notranslate"><span class="pre">DeterministicEngine</span></code></a></p></li>
<li><p><a class="reference internal" href="generated/ignite.engine.deterministic.ReproducibleBatchSampler.html#ignite.engine.deterministic.ReproducibleBatchSampler" title="ignite.engine.deterministic.ReproducibleBatchSampler"><code class="xref py py-class docutils literal notranslate"><span class="pre">ReproducibleBatchSampler</span></code></a></p></li>
</ul>
</section>
<section id="dataflow-synchronization">
<h2>Dataflow synchronization<a class="headerlink" href="#dataflow-synchronization" title="Permalink to this heading">#</a></h2>
<p>Ignite provides an option to control the dataflow by synchronizing random state on epochs. In this way, for a given
iteration/epoch the dataflow can be the same for a given seed. More precisely it is roughly looks like:</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="k">for</span> <span class="n">e</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">num_epochs</span><span class="p">):</span>
<span class="n">set_seed</span><span class="p">(</span><span class="n">seed</span> <span class="o">+</span> <span class="n">e</span><span class="p">)</span>
<span class="n">do_single_epoch_iterations</span><span class="p">(</span><span class="n">dataloader</span><span class="p">)</span>
</pre></div>
</div>
<p>In addition, if data provider is <code class="docutils literal notranslate"><span class="pre">torch.utils.data.DataLoader</span></code>, batch data indices can be made completely deterministic.
Here is a trivial example of usage:</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span><span class="w"> </span><span class="nn">torch</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">torch.utils.data</span><span class="w"> </span><span class="kn">import</span> <span class="n">DataLoader</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">ignite.engine</span><span class="w"> </span><span class="kn">import</span> <span class="n">DeterministicEngine</span><span class="p">,</span> <span class="n">Events</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">ignite.utils</span><span class="w"> </span><span class="kn">import</span> <span class="n">manual_seed</span>
<span class="k">def</span><span class="w"> </span><span class="nf">random_train_data_loader</span><span class="p">(</span><span class="n">size</span><span class="p">):</span>
<span class="n">data</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">size</span><span class="p">)</span>
<span class="k">return</span> <span class="n">DataLoader</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="mi">4</span><span class="p">,</span> <span class="n">shuffle</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">def</span><span class="w"> </span><span class="nf">print_train_data</span><span class="p">(</span><span class="n">engine</span><span class="p">,</span> <span class="n">batch</span><span class="p">):</span>
<span class="n">i</span> <span class="o">=</span> <span class="n">engine</span><span class="o">.</span><span class="n">state</span><span class="o">.</span><span class="n">iteration</span>
<span class="n">e</span> <span class="o">=</span> <span class="n">engine</span><span class="o">.</span><span class="n">state</span><span class="o">.</span><span class="n">epoch</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"train"</span><span class="p">,</span> <span class="n">e</span><span class="p">,</span> <span class="n">i</span><span class="p">,</span> <span class="n">batch</span><span class="o">.</span><span class="n">tolist</span><span class="p">())</span>
<span class="n">trainer</span> <span class="o">=</span> <span class="n">DeterministicEngine</span><span class="p">(</span><span class="n">print_train_data</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Original Run"</span><span class="p">)</span>
<span class="n">manual_seed</span><span class="p">(</span><span class="mi">56</span><span class="p">)</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">random_train_data_loader</span><span class="p">(</span><span class="mi">40</span><span class="p">),</span> <span class="n">max_epochs</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">epoch_length</span><span class="o">=</span><span class="mi">5</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Resumed Run"</span><span class="p">)</span>
<span class="c1"># Resume from 2nd epoch</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">load_state_dict</span><span class="p">({</span><span class="s2">"epoch"</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span> <span class="s2">"epoch_length"</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span> <span class="s2">"max_epochs"</span><span class="p">:</span> <span class="mi">2</span><span class="p">,</span> <span class="s2">"rng_states"</span><span class="p">:</span> <span class="kc">None</span><span class="p">})</span>
<span class="n">manual_seed</span><span class="p">(</span><span class="mi">56</span><span class="p">)</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">random_train_data_loader</span><span class="p">(</span><span class="mi">40</span><span class="p">))</span>
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>Original Run
train 1 1 [31, 13, 3, 4]
train 1 2 [23, 18, 6, 16]
train 1 3 [10, 8, 33, 36]
train 1 4 [1, 37, 19, 9]
train 1 5 [20, 30, 14, 26]
train 2 6 [29, 35, 38, 34]
train 2 7 [7, 22, 12, 17]
train 2 8 [25, 21, 24, 15]
train 2 9 [39, 5, 2, 28]
train 2 10 [27, 11, 32, 0]
Resumed Run
train 2 6 [29, 35, 38, 34]
train 2 7 [7, 22, 12, 17]
train 2 8 [25, 21, 24, 15]
train 2 9 [39, 5, 2, 28]
train 2 10 [27, 11, 32, 0]
</pre></div>
</div>
<p>We can see that the data samples are exactly the same between original and resumed runs.</p>
<p>Complete examples that simulates a crash on a defined iteration and resumes the training from a checkpoint can be found
here:</p>
<ul class="simple">
<li><p><a class="reference external" href="https://github.com/pytorch/ignite/tree/master/examples/mnist#training-save--resume">save/resume MNIST</a></p></li>
<li><p><a class="reference external" href="https://github.com/pytorch/ignite/tree/master/examples/cifar10#check-resume-training">save/resume Distributed CIFAR10</a></p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>In case when input data is <cite>torch.utils.data.DataLoader</cite>, previous batches are skipped and the first provided batch
corresponds to the batch after the checkpoint iteration. Internally, while resuming, previous datapoint indices are just
skipped without fetching the data.</p>
</div>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>However, while resuming from iteration, random data augmentations are not synchronized in the middle of the epoch and
thus batches remaining until the end of the epoch can be different of those from the initial run.</p>
</div>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>However, please, keep in mind that there can be an issue with dataflow synchronization on every epoch
if user’s handler synchronizes the random state, for example, by calling periodically <code class="docutils literal notranslate"><span class="pre">torch.manual_seed(seed)</span></code> during
the run. This can have an impact on the dataflow:</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="k">def</span><span class="w"> </span><span class="nf">random_train_data_generator</span><span class="p">():</span>
<span class="k">while</span> <span class="kc">True</span><span class="p">:</span>
<span class="k">yield</span> <span class="n">torch</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">100</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="p">))</span>
<span class="n">trainer</span> <span class="o">=</span> <span class="n">DeterministicEngine</span><span class="p">(</span><span class="n">print_train_data</span><span class="p">)</span>
<span class="nd">@trainer</span><span class="o">.</span><span class="n">on</span><span class="p">(</span><span class="n">Events</span><span class="o">.</span><span class="n">ITERATION_COMPLETED</span><span class="p">(</span><span class="n">every</span><span class="o">=</span><span class="mi">3</span><span class="p">))</span>
<span class="k">def</span><span class="w"> </span><span class="nf">user_handler</span><span class="p">():</span>
<span class="c1"># handler synchronizes the random state</span>
<span class="n">torch</span><span class="o">.</span><span class="n">manual_seed</span><span class="p">(</span><span class="mi">12</span><span class="p">)</span>
<span class="n">a</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
<span class="n">trainer</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">random_train_data_generator</span><span class="p">(),</span> <span class="n">max_epochs</span><span class="o">=</span><span class="mi">3</span><span class="p">,</span> <span class="n">epoch_length</span><span class="o">=</span><span class="mi">5</span><span class="p">);</span>
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>train 1 1 [32]
train 1 2 [29]
train 1 3 [40]
train 1 4 [3] <---
train 1 5 [22]
train 2 6 [77]
train 2 7 [3] <---
train 2 8 [22]
train 2 9 [77]
train 2 10 [3] <---
train 3 11 [22]
train 3 12 [77]
train 3 13 [3] <---
train 3 14 [22]
train 3 15 [77]
</pre></div>
</div>
<p>Initially, the function <code class="docutils literal notranslate"><span class="pre">random_train_data_generator()</span></code> generates randomly data batches using the random state set
up by <code class="docutils literal notranslate"><span class="pre">trainer</span></code>. This is intended behaviour until <code class="docutils literal notranslate"><span class="pre">user_handler()</span></code> is called.
After <code class="docutils literal notranslate"><span class="pre">user_handler()</span></code> execution, random state is altered and thus <code class="docutils literal notranslate"><span class="pre">random_train_data_generator()</span></code> will produce
random batches based on altered random state.</p>
<p>We provide helper decorator <a class="reference internal" href="generated/ignite.engine.deterministic.keep_random_state.html#ignite.engine.deterministic.keep_random_state" title="ignite.engine.deterministic.keep_random_state"><code class="xref py py-meth docutils literal notranslate"><span class="pre">keep_random_state()</span></code></a> to save and restore random states for
<cite>torch</cite>, <cite>numpy</cite> and <cite>random</cite>. Therefore, we can deal with described issue using this decorator:</p>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span><span class="w"> </span><span class="nn">ignite.engine.deterministic</span><span class="w"> </span><span class="kn">import</span> <span class="n">keep_random_state</span>
<span class="nd">@trainer</span><span class="o">.</span><span class="n">on</span><span class="p">(</span><span class="n">Events</span><span class="o">.</span><span class="n">ITERATION_COMPLETED</span><span class="p">(</span><span class="n">every</span><span class="o">=</span><span class="mi">3</span><span class="p">))</span>
<span class="nd">@keep_random_state</span>
<span class="k">def</span><span class="w"> </span><span class="nf">user_handler</span><span class="p">():</span>
<span class="c1"># handler synchronizes the random state</span>
<span class="n">torch</span><span class="o">.</span><span class="n">manual_seed</span><span class="p">(</span><span class="mi">12</span><span class="p">)</span>
<span class="n">a</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
</pre></div>
</div>
</div>
</section>
</section>
</article>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="generated/ignite.engine.engine.Engine.html" class="btn btn-neutral float-right" title="Engine" accesskey="n" rel="next">Next
<img src="_static/images/chevron-right-orange.svg" alt="right arrow" class="next-page">
</a>
<a href="index.html" class="btn btn-neutral" title="Ignite Your Networks!" accesskey="p" rel="prev">
<img src="_static/images/chevron-right-orange.svg" alt="left arrow" class="previous-page"> Previous
</a>
</div>
<hr>
<div role="contentinfo">
<table>
<tr>
<td>
<p>
© Copyright 2025, PyTorch-Ignite Contributors.
Last updated on 08/19/2023, 9:47:16 PM.
</p>
<div>
Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</div>
</td>
<td>
<div>
<a href="https://www.netlify.com" target="_blank" rel="noopener noreferrer">
<img
src="_static/img/netlify-light.svg"
alt="Deploys by Netlify"
width=114
height=51 />
</a>
</div>
</td>
</tr>
</table>
</div>
</footer>
</div>
</div>
<div class="pytorch-content-right" id="pytorch-content-right">
<div class="pytorch-right-menu" id="pytorch-right-menu">
<div class="pytorch-side-scroll" id="pytorch-side-scroll-right">
<ul>
<li><a class="reference internal" href="#">ignite.engine</a><ul>
<li><a class="reference internal" href="#ignite-engine-engine">ignite.engine.engine</a></li>
<li><a class="reference internal" href="#ignite-engine-events">ignite.engine.events</a></li>
<li><a class="reference internal" href="#ignite-engine-deterministic">ignite.engine.deterministic</a></li>
<li><a class="reference internal" href="#helper-methods-to-define-supervised-trainer-and-evaluator">helper methods to define supervised trainer and evaluator</a></li>
<li><a class="reference internal" href="#resuming-the-training">Resuming the training</a></li>
<li><a class="reference internal" href="#deterministic-training">Deterministic training</a></li>
<li><a class="reference internal" href="#dataflow-synchronization">Dataflow synchronization</a></li>
</ul>
</li>
</ul>
</div>
</div>
</div>
</section>
</div>
<script type="text/javascript" id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
<script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
<script src="_static/jquery.js"></script>
<script src="_static/underscore.js"></script>
<script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
<script src="_static/doctools.js"></script>
<script src="_static/sphinx_highlight.js"></script>
<script src="_static/clipboard.min.js"></script>
<script src="_static/copybutton.js"></script>
<script>let toggleHintShow = 'Show default setup';</script>
<script>let toggleHintHide = 'Hide default setup';</script>
<script>let toggleOpenOnPrint = 'true';</script>
<script src="_static/togglebutton.js"></script>
<script>var togglebuttonSelector = '.toggle, .admonition.dropdown';</script>
<script src="_static/design-tabs.js"></script>
<script type="text/javascript" src="_static/js/vendor/popper.min.js"></script>
<script type="text/javascript" src="_static/js/vendor/bootstrap.min.js"></script>
<script type="text/javascript" src="_static/js/theme.js"></script>
<script type="text/javascript">
jQuery(function () {
SphinxRtdTheme.Navigation.enable(true);
});
</script>
<!-- Begin Footer -->
<!-- commented out for Ignite -->
<div class="container-fluid docs-tutorials-resources" id="docs-tutorials-resources">
<!-- <div class="container">
<div class="row">
<div class="text-center col-md-4">
<h2>Docs</h2>
<p>Access comprehensive developer documentation for PyTorch</p>
<a class="with-right-arrow" href="https://pytorch.org/ignite/index.html">View Docs</a>
</div>
<div class="text-center col-md-4">
<h2>Tutorials</h2>
<p>Get in-depth tutorials for beginners and advanced developers</p>
<a class="with-right-arrow" href="">View Tutorials</a>
</div>
<div class="text-center col-md-4">
<h2>Resources</h2>
<p>Find development resources and get your questions answered</p>
<a class="with-right-arrow" href="">View Resources</a>
</div>
</div>
</div> -->
</div>
<!-- <footer class="site-footer">
<div class="container footer-container">
<div class="footer-logo-wrapper">
<a href="https://pytorch-ignite.ai" class="footer-logo"></a>
</div>
<div class="footer-links-wrapper">
<div class="footer-links-col">
<ul>
<li class="list-title"><a href="https://pytorch-ignite.ai">PyTorch</a></li>
<li><a href="">Get Started</a></li>
<li><a href="">Features</a></li>
<li><a href="">Ecosystem</a></li>
<li><a href="">Blog</a></li>
<li><a href="">Resources</a></li>
</ul>
</div>
<div class="footer-links-col">
<ul>
<li class="list-title"><a href="">Support</a></li>
<li><a href="">Tutorials</a></li>
<li><a href="https://pytorch.org/ignite/index.html">Docs</a></li>
<li><a href="" target="_blank">Discuss</a></li>
<li><a href="https://github.com/pytorch/ignite/issues" target="_blank">Github Issues</a></li>
<li><a href="" target="_blank">Slack</a></li>
<li><a href="https://github.com/pytorch/ignite/blob/master/CONTRIBUTING.md" target="_blank">Contributing</a></li>
</ul>
</div>
<div class="footer-links-col follow-us-col">
<ul>
<li class="list-title">Follow Us</li>
<li>
<div id="mc_embed_signup">
<form
action="https://twitter.us14.list-manage.com/subscribe/post?u=75419c71fe0a935e53dfa4a3f&id=91d0dccd39"
method="post"
id="mc-embedded-subscribe-form"
name="mc-embedded-subscribe-form"
class="email-subscribe-form validate"
target="_blank"
novalidate>
<div id="mc_embed_signup_scroll" class="email-subscribe-form-fields-wrapper">
<div class="mc-field-group">
<label for="mce-EMAIL" style="display:none;">Email Address</label>
<input type="email" value="" name="EMAIL" class="required email" id="mce-EMAIL" placeholder="Email Address">
</div>
<div id="mce-responses" class="clear">
<div class="response" id="mce-error-response" style="display:none"></div>
<div class="response" id="mce-success-response" style="display:none"></div>
</div> <!~~ real people should not fill this in and expect good things - do not remove this or risk form bot signups ~~>
<div style="position: absolute; left: -5000px;" aria-hidden="true"><input type="text" name="b_75419c71fe0a935e53dfa4a3f_91d0dccd39" tabindex="-1" value=""></div>
<div class="clear">
<input type="submit" value="" name="subscribe" id="mc-embedded-subscribe" class="button email-subscribe-button">
</div>
</div>
</form>
</div>
</li>
</ul>
<div class="footer-social-icons">
<a href="" target="_blank" class="facebook"></a>
<a href="" target="_blank" class="twitter"></a>
</div>
</div>
</div>
</div>
</footer> -->
<!-- end of commented out for Ignite -->
<!-- <div class="cookie-banner-wrapper">
<div class="container">
<p class="gdpr-notice">To analyze traffic and optimize your experience, we serve cookies on this site. By clicking or navigating, you agree to allow our usage of cookies. As the current maintainers of this site, Facebook’s Cookies Policy applies. Learn more, including about available controls: <a href="https://www.facebook.com/policies/cookies/">Cookies Policy</a>.</p>
<img class="close-button" src="_static/images/pytorch-x.svg">
</div>
</div> -->
<!-- End Footer -->
<!-- Begin Mobile Menu -->
<!--
<div class="mobile-main-menu">
<div class="container-fluid">
<div class="container">
<div class="mobile-main-menu-header-container">
<a class="header-logo" href="https://pytorch-ignite.ai" aria-label="PyTorch"></a>
<a class="main-menu-close-button" href="#" data-behavior="close-mobile-menu"></a>
</div>
</div>
</div>
<div class="mobile-main-menu-links-container">
<div class="main-menu">
<ul>
<li>
<a href="#">Get Started</a>
</li>
<li>
<a href="#">Features</a>
</li>
<li>
<a href="#">Ecosystem</a>
</li>
<li>
<a href="">Blog</a>
</li>
<li>
<a href="">Tutorials</a>
</li>
<li>
<a href="https://pytorch.org/ignite/index.html">Docs</a>
</li>
<li>
<a href="">Resources</a>
</li>
<li>
<a href="https://github.com/pytorch/ignite">Github</a>
</li>
</ul>
</div>
</div>
</div>
-->
<!-- End Mobile Menu -->
<script type="text/javascript" src="_static/js/vendor/anchor.min.js"></script>
<script type="text/javascript">
var collapsedSections = []
$(document).ready(function() {
mobileMenu.bind();
mobileTOC.bind();
pytorchAnchors.bind();
sideMenus.bind();
scrollToAnchor.bind();
highlightNavigation.bind();
// Add class to links that have code blocks, since we cannot create links in code blocks
$("article.pytorch-article a span.pre").each(function(e) {
$(this).closest("a").addClass("has-code");
});
})
</script>
<script src="https://cdn.jsdelivr.net/npm/@docsearch/js@3"></script>
<script type="text/javascript">
let VERSION
if ('v0.5.1'.startsWith('v')) {
VERSION = 'v0.5.1'
} else {
VERSION = 'master'
}
docsearch({
container: '#docsearch',
appId: '7EWYE1JCT3',
apiKey: '841e93e60c16975ba1bd8c7c716eed82',
indexName: 'pytorch-ignite',
placeholder: 'Search PyTorch-Ignite docs',
searchParameters: {
facetFilters: [`version:${VERSION}`, 'tags:API-reference'],
}
});
</script>
</body>
</html>