-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathextending.html
1291 lines (1081 loc) · 112 KB
/
extending.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
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
<!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 name="robots" content="noindex">
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Extending PyTorch — PyTorch 1.7.0 documentation</title>
<link rel="canonical" href="https://pytorch.org/docs/stable/notes/extending.html"/>
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<!-- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" /> -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/jit.css" type="text/css" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/katex-math.css" type="text/css" />
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
<link rel="next" title="Frequently Asked Questions" href="faq.html" />
<link rel="prev" title="Distributed Data Parallel" href="ddp.html" />
<script src="../_static/js/modernizr.min.js"></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">
<!-- 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">
</head>
<div class="container-fluid header-holder tutorials-header" id="header-holder">
<div class="container">
<div class="header-container">
<a class="header-logo" href="https://pytorch.org/" aria-label="PyTorch"></a>
<div class="main-menu">
<ul>
<li>
<a href="https://pytorch.org/get-started">Get Started</a>
</li>
<li>
<div class="ecosystem-dropdown">
<a id="dropdownMenuButton" data-toggle="ecosystem-dropdown">
Ecosystem
</a>
<div class="ecosystem-dropdown-menu">
<a class="nav-dropdown-item" href="https://pytorch.org/hub"">
<span class=dropdown-title>Models (Beta)</span>
<p>Discover, publish, and reuse pre-trained models</p>
</a>
<a class="nav-dropdown-item" href="https://pytorch.org/ecosystem">
<span class=dropdown-title>Tools & Libraries</span>
<p>Explore the ecosystem of tools and libraries</p>
</a>
</div>
</div>
</li>
<li>
<a href="https://pytorch.org/mobile">Mobile</a>
</li>
<li>
<a href="https://pytorch.org/blog/">Blog</a>
</li>
<li>
<a href="https://pytorch.org/tutorials">Tutorials</a>
</li>
<li class="active">
<a href="https://pytorch.org/docs/stable/index.html">Docs</a>
</li>
<li>
<div class="resources-dropdown">
<a id="resourcesDropdownButton" data-toggle="resources-dropdown">
Resources
</a>
<div class="resources-dropdown-menu">
<a class="nav-dropdown-item" href="https://pytorch.org/resources"">
<span class=dropdown-title>Developer Resources</span>
<p>Find resources and get questions answered</p>
</a>
<a class="nav-dropdown-item" href="https://pytorch.org/features">
<span class=dropdown-title>About</span>
<p>Learn about PyTorch’s features and capabilities</p>
</a>
</div>
</div>
</li>
<li>
<a href="https://github.com/pytorch/pytorch">Github</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>
<a href="#" class="toggle-table-of-contents" data-behavior="toggle-table-of-contents"></a>
</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">
<a href='http://pytorch.org/docs/versions.html'>1.7.0 ▼</a>
</div>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
<input type="text" name="q" placeholder="Search Docs" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div>
<p class="caption"><span class="caption-text">Notes</span></p>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="amp_examples.html">Automatic Mixed Precision examples</a></li>
<li class="toctree-l1"><a class="reference internal" href="autograd.html">Autograd mechanics</a></li>
<li class="toctree-l1"><a class="reference internal" href="broadcasting.html">Broadcasting semantics</a></li>
<li class="toctree-l1"><a class="reference internal" href="cpu_threading_torchscript_inference.html">CPU threading and TorchScript inference</a></li>
<li class="toctree-l1"><a class="reference internal" href="cuda.html">CUDA semantics</a></li>
<li class="toctree-l1"><a class="reference internal" href="ddp.html">Distributed Data Parallel</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">Extending PyTorch</a></li>
<li class="toctree-l1"><a class="reference internal" href="faq.html">Frequently Asked Questions</a></li>
<li class="toctree-l1"><a class="reference internal" href="large_scale_deployments.html">Features for large-scale deployments</a></li>
<li class="toctree-l1"><a class="reference internal" href="multiprocessing.html">Multiprocessing best practices</a></li>
<li class="toctree-l1"><a class="reference internal" href="randomness.html">Reproducibility</a></li>
<li class="toctree-l1"><a class="reference internal" href="serialization.html">Serialization semantics</a></li>
<li class="toctree-l1"><a class="reference internal" href="windows.html">Windows FAQ</a></li>
</ul>
<p class="caption"><span class="caption-text">Language Bindings</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../cpp_index.html">C++</a></li>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/javadoc/">Javadoc</a></li>
</ul>
<p class="caption"><span class="caption-text">Python API</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../torch.html">torch</a></li>
<li class="toctree-l1"><a class="reference internal" href="../nn.html">torch.nn</a></li>
<li class="toctree-l1"><a class="reference internal" href="../nn.functional.html">torch.nn.functional</a></li>
<li class="toctree-l1"><a class="reference internal" href="../tensors.html">torch.Tensor</a></li>
<li class="toctree-l1"><a class="reference internal" href="../tensor_attributes.html">Tensor Attributes</a></li>
<li class="toctree-l1"><a class="reference internal" href="../tensor_view.html">Tensor Views</a></li>
<li class="toctree-l1"><a class="reference internal" href="../autograd.html">torch.autograd</a></li>
<li class="toctree-l1"><a class="reference internal" href="../cuda.html">torch.cuda</a></li>
<li class="toctree-l1"><a class="reference internal" href="../amp.html">torch.cuda.amp</a></li>
<li class="toctree-l1"><a class="reference internal" href="../backends.html">torch.backends</a></li>
<li class="toctree-l1"><a class="reference internal" href="../distributed.html">torch.distributed</a></li>
<li class="toctree-l1"><a class="reference internal" href="../distributions.html">torch.distributions</a></li>
<li class="toctree-l1"><a class="reference internal" href="../fft.html">torch.fft</a></li>
<li class="toctree-l1"><a class="reference internal" href="../futures.html">torch.futures</a></li>
<li class="toctree-l1"><a class="reference internal" href="../hub.html">torch.hub</a></li>
<li class="toctree-l1"><a class="reference internal" href="../jit.html">torch.jit</a></li>
<li class="toctree-l1"><a class="reference internal" href="../linalg.html">torch.linalg</a></li>
<li class="toctree-l1"><a class="reference internal" href="../nn.init.html">torch.nn.init</a></li>
<li class="toctree-l1"><a class="reference internal" href="../onnx.html">torch.onnx</a></li>
<li class="toctree-l1"><a class="reference internal" href="../optim.html">torch.optim</a></li>
<li class="toctree-l1"><a class="reference internal" href="../complex_numbers.html">Complex Numbers</a></li>
<li class="toctree-l1"><a class="reference internal" href="../quantization.html">Quantization</a></li>
<li class="toctree-l1"><a class="reference internal" href="../rpc.html">Distributed RPC Framework</a></li>
<li class="toctree-l1"><a class="reference internal" href="../random.html">torch.random</a></li>
<li class="toctree-l1"><a class="reference internal" href="../sparse.html">torch.sparse</a></li>
<li class="toctree-l1"><a class="reference internal" href="../storage.html">torch.Storage</a></li>
<li class="toctree-l1"><a class="reference internal" href="../bottleneck.html">torch.utils.bottleneck</a></li>
<li class="toctree-l1"><a class="reference internal" href="../checkpoint.html">torch.utils.checkpoint</a></li>
<li class="toctree-l1"><a class="reference internal" href="../cpp_extension.html">torch.utils.cpp_extension</a></li>
<li class="toctree-l1"><a class="reference internal" href="../data.html">torch.utils.data</a></li>
<li class="toctree-l1"><a class="reference internal" href="../dlpack.html">torch.utils.dlpack</a></li>
<li class="toctree-l1"><a class="reference internal" href="../mobile_optimizer.html">torch.utils.mobile_optimizer</a></li>
<li class="toctree-l1"><a class="reference internal" href="../model_zoo.html">torch.utils.model_zoo</a></li>
<li class="toctree-l1"><a class="reference internal" href="../tensorboard.html">torch.utils.tensorboard</a></li>
<li class="toctree-l1"><a class="reference internal" href="../type_info.html">Type Info</a></li>
<li class="toctree-l1"><a class="reference internal" href="../named_tensor.html">Named Tensors</a></li>
<li class="toctree-l1"><a class="reference internal" href="../name_inference.html">Named Tensors operator coverage</a></li>
<li class="toctree-l1"><a class="reference internal" href="../__config__.html">torch.__config__</a></li>
</ul>
<p class="caption"><span class="caption-text">Libraries</span></p>
<ul>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/audio">torchaudio</a></li>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/text">torchtext</a></li>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/vision">torchvision</a></li>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/elastic/">TorchElastic</a></li>
<li class="toctree-l1"><a class="reference external" href="https://pytorch.org/serve">TorchServe</a></li>
<li class="toctree-l1"><a class="reference external" href="http://pytorch.org/xla/">PyTorch on XLA Devices</a></li>
</ul>
<p class="caption"><span class="caption-text">Community</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../community/contribution_guide.html">PyTorch Contribution Guide</a></li>
<li class="toctree-l1"><a class="reference internal" href="../community/governance.html">PyTorch Governance</a></li>
<li class="toctree-l1"><a class="reference internal" href="../community/persons_of_interest.html">PyTorch Governance | Persons of Interest</a></li>
</ul>
</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>Extending PyTorch</li>
<li class="pytorch-breadcrumbs-aside">
<a href="../_sources/notes/extending.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">
<div class="section" id="extending-pytorch">
<h1>Extending PyTorch<a class="headerlink" href="#extending-pytorch" title="Permalink to this headline">¶</a></h1>
<p>In this note we’ll cover ways of extending <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch.nn</span></code>,
<a class="reference internal" href="../autograd.html#module-torch.autograd" title="torch.autograd"><code class="xref py py-mod docutils literal notranslate"><span class="pre">torch.autograd</span></code></a>, <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code>, and writing custom C extensions utilizing our C
libraries.</p>
<div class="section" id="extending-torch-autograd">
<span id="extending-autograd"></span><h2>Extending <a class="reference internal" href="../autograd.html#module-torch.autograd" title="torch.autograd"><code class="xref py py-mod docutils literal notranslate"><span class="pre">torch.autograd</span></code></a><a class="headerlink" href="#extending-torch-autograd" title="Permalink to this headline">¶</a></h2>
<p>Adding operations to <a class="reference internal" href="../autograd.html#module-torch.autograd" title="torch.autograd"><code class="xref py py-mod docutils literal notranslate"><span class="pre">autograd</span></code></a> requires implementing a new
<a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a> subclass for each operation. Recall that <a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a> s
are what <a class="reference internal" href="../autograd.html#module-torch.autograd" title="torch.autograd"><code class="xref py py-mod docutils literal notranslate"><span class="pre">autograd</span></code></a> uses to compute the results and gradients, and
encode the operation history. Every new function requires you to implement 2 methods:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.Function.forward" title="torch.autograd.Function.forward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">forward()</span></code></a> - the code that performs the operation. It can take
as many arguments as you want, with some of them being optional, if you
specify the default values. All kinds of Python objects are accepted here.
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> arguments that track history (i.e., with
<code class="docutils literal notranslate"><span class="pre">requires_grad=True</span></code>) will be converted to ones that don’t track history
before the call, and their use will be registered in the graph. Note that this
logic won’t traverse lists/dicts/any other data structures and will only
consider <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> s that are direct arguments to the call. You can
return either a single <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> output, or a <a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#tuple" title="(in Python v3.9)"><code class="xref py py-class docutils literal notranslate"><span class="pre">tuple</span></code></a> of
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> s if there are multiple outputs. Also, please refer to the
docs of <a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a> to find descriptions of useful methods that can be
called only from <a class="reference internal" href="../autograd.html#torch.autograd.Function.forward" title="torch.autograd.Function.forward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">forward()</span></code></a>.</p></li>
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.Function.backward" title="torch.autograd.Function.backward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">backward()</span></code></a> - gradient formula. It will be given
as many <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> arguments as there were outputs, with each of them
representing gradient w.r.t. that output. It should return as many
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> s as there were inputs, with each of them containing the
gradient w.r.t. its corresponding input. If your inputs didn’t require
gradient (<code class="xref py py-attr docutils literal notranslate"><span class="pre">needs_input_grad</span></code> is a tuple of booleans indicating
whether each input needs gradient computation), or were non-<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>
objects, you can return <code class="xref py py-class docutils literal notranslate"><span class="pre">None</span></code>. Also, if you have optional
arguments to <a class="reference internal" href="../autograd.html#torch.autograd.Function.forward" title="torch.autograd.Function.forward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">forward()</span></code></a> you can return more gradients than there
were inputs, as long as they’re all <a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.9)"><code class="docutils literal notranslate"><span class="pre">None</span></code></a>.</p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It’s the user’s responsibility to use the special functions in the forward’s <cite>ctx</cite>
properly in order to ensure that the new <a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a> works properly with
the autograd engine.</p>
<ul class="simple">
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.function._ContextMethodMixin.save_for_backward" title="torch.autograd.function._ContextMethodMixin.save_for_backward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">save_for_backward()</span></code></a> must be
used when saving input or output of the forward to be used later in the backward.</p></li>
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.function._ContextMethodMixin.mark_dirty" title="torch.autograd.function._ContextMethodMixin.mark_dirty"><code class="xref py py-meth docutils literal notranslate"><span class="pre">mark_dirty()</span></code></a> must be used to
mark any input that is modified inplace by the forward function.</p></li>
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.function._ContextMethodMixin.mark_non_differentiable" title="torch.autograd.function._ContextMethodMixin.mark_non_differentiable"><code class="xref py py-meth docutils literal notranslate"><span class="pre">mark_non_differentiable()</span></code></a> must
be used to tell the engine if an output is not differentiable.</p></li>
<li><p><a class="reference internal" href="../autograd.html#torch.autograd.function._ContextMethodMixin.set_materialize_grads" title="torch.autograd.function._ContextMethodMixin.set_materialize_grads"><code class="xref py py-meth docutils literal notranslate"><span class="pre">set_materialize_grads()</span></code></a> can be
used to tell the autograd engine to optimize gradient computations in the cases where
the output does not depend on the input by not materializing grad tensors given to backward
function. That is, if set to False, None object in python or “undefined tensor” (tensor x for
which x.defined() is False) in C++ will not be converted to a tensor filled with zeros prior
to calling backward. However, supporting this optimization means your custom autograd function
has to handle gradients that are represented in this way and is thus opt-in. Default value is True.</p></li>
</ul>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>By default, all the output Tensors that are of differentiable type will be set to
require gradient and have all autograd metadata set for them. If you don’t want
them to require gradients, you can use the <cite>mark_non_differentiable</cite> method mentioned
above. For output Tensors that are not of differentiable type (integer types for example),
they won’t be marked as requiring gradients.</p>
</div>
<p>Below you can find code for a <code class="docutils literal notranslate"><span class="pre">Linear</span></code> function from <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch.nn</span></code>, with
additional comments:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># Inherit from Function</span>
<span class="k">class</span> <span class="nc">LinearFunction</span><span class="p">(</span><span class="n">Function</span><span class="p">):</span>
<span class="c1"># Note that both forward and backward are @staticmethods</span>
<span class="nd">@staticmethod</span>
<span class="c1"># bias is an optional argument</span>
<span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="nb">input</span><span class="p">,</span> <span class="n">weight</span><span class="p">,</span> <span class="n">bias</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">ctx</span><span class="o">.</span><span class="n">save_for_backward</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="n">weight</span><span class="p">,</span> <span class="n">bias</span><span class="p">)</span>
<span class="n">output</span> <span class="o">=</span> <span class="nb">input</span><span class="o">.</span><span class="n">mm</span><span class="p">(</span><span class="n">weight</span><span class="o">.</span><span class="n">t</span><span class="p">())</span>
<span class="k">if</span> <span class="n">bias</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">output</span> <span class="o">+=</span> <span class="n">bias</span><span class="o">.</span><span class="n">unsqueeze</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span><span class="o">.</span><span class="n">expand_as</span><span class="p">(</span><span class="n">output</span><span class="p">)</span>
<span class="k">return</span> <span class="n">output</span>
<span class="c1"># This function has only a single output, so it gets only one gradient</span>
<span class="nd">@staticmethod</span>
<span class="k">def</span> <span class="nf">backward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="n">grad_output</span><span class="p">):</span>
<span class="c1"># This is a pattern that is very convenient - at the top of backward</span>
<span class="c1"># unpack saved_tensors and initialize all gradients w.r.t. inputs to</span>
<span class="c1"># None. Thanks to the fact that additional trailing Nones are</span>
<span class="c1"># ignored, the return statement is simple even when the function has</span>
<span class="c1"># optional inputs.</span>
<span class="nb">input</span><span class="p">,</span> <span class="n">weight</span><span class="p">,</span> <span class="n">bias</span> <span class="o">=</span> <span class="n">ctx</span><span class="o">.</span><span class="n">saved_tensors</span>
<span class="n">grad_input</span> <span class="o">=</span> <span class="n">grad_weight</span> <span class="o">=</span> <span class="n">grad_bias</span> <span class="o">=</span> <span class="kc">None</span>
<span class="c1"># These needs_input_grad checks are optional and there only to</span>
<span class="c1"># improve efficiency. If you want to make your code simpler, you can</span>
<span class="c1"># skip them. Returning gradients for inputs that don't require it is</span>
<span class="c1"># not an error.</span>
<span class="k">if</span> <span class="n">ctx</span><span class="o">.</span><span class="n">needs_input_grad</span><span class="p">[</span><span class="mi">0</span><span class="p">]:</span>
<span class="n">grad_input</span> <span class="o">=</span> <span class="n">grad_output</span><span class="o">.</span><span class="n">mm</span><span class="p">(</span><span class="n">weight</span><span class="p">)</span>
<span class="k">if</span> <span class="n">ctx</span><span class="o">.</span><span class="n">needs_input_grad</span><span class="p">[</span><span class="mi">1</span><span class="p">]:</span>
<span class="n">grad_weight</span> <span class="o">=</span> <span class="n">grad_output</span><span class="o">.</span><span class="n">t</span><span class="p">()</span><span class="o">.</span><span class="n">mm</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
<span class="k">if</span> <span class="n">bias</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="n">ctx</span><span class="o">.</span><span class="n">needs_input_grad</span><span class="p">[</span><span class="mi">2</span><span class="p">]:</span>
<span class="n">grad_bias</span> <span class="o">=</span> <span class="n">grad_output</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
<span class="k">return</span> <span class="n">grad_input</span><span class="p">,</span> <span class="n">grad_weight</span><span class="p">,</span> <span class="n">grad_bias</span>
</pre></div>
</div>
<p>Now, to make it easier to use these custom ops, we recommend aliasing their
<code class="docutils literal notranslate"><span class="pre">apply</span></code> method:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">linear</span> <span class="o">=</span> <span class="n">LinearFunction</span><span class="o">.</span><span class="n">apply</span>
</pre></div>
</div>
<p>Here, we give an additional example of a function that is parametrized by
non-Tensor arguments:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">MulConstant</span><span class="p">(</span><span class="n">Function</span><span class="p">):</span>
<span class="nd">@staticmethod</span>
<span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="n">tensor</span><span class="p">,</span> <span class="n">constant</span><span class="p">):</span>
<span class="c1"># ctx is a context object that can be used to stash information</span>
<span class="c1"># for backward computation</span>
<span class="n">ctx</span><span class="o">.</span><span class="n">constant</span> <span class="o">=</span> <span class="n">constant</span>
<span class="k">return</span> <span class="n">tensor</span> <span class="o">*</span> <span class="n">constant</span>
<span class="nd">@staticmethod</span>
<span class="k">def</span> <span class="nf">backward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="n">grad_output</span><span class="p">):</span>
<span class="c1"># We return as many input gradients as there were arguments.</span>
<span class="c1"># Gradients of non-Tensor arguments to forward must be None.</span>
<span class="k">return</span> <span class="n">grad_output</span> <span class="o">*</span> <span class="n">ctx</span><span class="o">.</span><span class="n">constant</span><span class="p">,</span> <span class="kc">None</span>
</pre></div>
</div>
<p>And here, we optimize the above example by calling set_materialize_grads(False):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">MulConstant</span><span class="p">(</span><span class="n">Function</span><span class="p">):</span>
<span class="nd">@staticmethod</span>
<span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="n">tensor</span><span class="p">,</span> <span class="n">constant</span><span class="p">):</span>
<span class="n">ctx</span><span class="o">.</span><span class="n">set_materialize_grads</span><span class="p">(</span><span class="kc">False</span><span class="p">)</span>
<span class="n">ctx</span><span class="o">.</span><span class="n">constant</span> <span class="o">=</span> <span class="n">constant</span>
<span class="k">return</span> <span class="n">tensor</span> <span class="o">*</span> <span class="n">constant</span>
<span class="nd">@staticmethod</span>
<span class="k">def</span> <span class="nf">backward</span><span class="p">(</span><span class="n">ctx</span><span class="p">,</span> <span class="n">grad_output</span><span class="p">):</span>
<span class="c1"># Here we must handle None grad_output tensor. In this case we</span>
<span class="c1"># can skip unnecessary computations and just return None.</span>
<span class="k">if</span> <span class="n">grad_output</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="k">return</span> <span class="kc">None</span><span class="p">,</span> <span class="kc">None</span>
<span class="c1"># We return as many input gradients as there were arguments.</span>
<span class="c1"># Gradients of non-Tensor arguments to forward must be None.</span>
<span class="k">return</span> <span class="n">grad_output</span> <span class="o">*</span> <span class="n">ctx</span><span class="o">.</span><span class="n">constant</span><span class="p">,</span> <span class="kc">None</span>
</pre></div>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Inputs to <code class="docutils literal notranslate"><span class="pre">backward</span></code>, i.e., <code class="xref py py-attr docutils literal notranslate"><span class="pre">grad_output</span></code>, can also be Tensors that
track history. So if <code class="docutils literal notranslate"><span class="pre">backward</span></code> is implemented with differentiable
operations, (e.g., invocation of another custom
<code class="xref py py-class docutils literal notranslate"><span class="pre">function</span></code>), higher order derivatives will work.
In this case, the Tensors saved with <code class="docutils literal notranslate"><span class="pre">save_for_backward</span></code> can also be used
in the backward and have gradients flowing back but Tensors saved in the <code class="docutils literal notranslate"><span class="pre">ctx</span></code>
won’t have gradients flowing back for them.
If you need gradients to flow back for a Tensor saved in the <code class="docutils literal notranslate"><span class="pre">ctx</span></code>, you should
make it an output of the custom <code class="docutils literal notranslate"><span class="pre">Function</span></code> and save it with <code class="docutils literal notranslate"><span class="pre">save_for_backward</span></code>.</p>
</div>
<p>You probably want to check if the backward method you implemented actually
computes the derivatives of your function. It is possible by comparing with
numerical approximations using small finite differences:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">torch.autograd</span> <span class="kn">import</span> <span class="n">gradcheck</span>
<span class="c1"># gradcheck takes a tuple of tensors as input, check if your gradient</span>
<span class="c1"># evaluated with these tensors are close enough to numerical</span>
<span class="c1"># approximations and returns True if they all verify this condition.</span>
<span class="nb">input</span> <span class="o">=</span> <span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="mi">20</span><span class="p">,</span><span class="mi">20</span><span class="p">,</span><span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">double</span><span class="p">,</span><span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">),</span> <span class="n">torch</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="mi">30</span><span class="p">,</span><span class="mi">20</span><span class="p">,</span><span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">double</span><span class="p">,</span><span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">))</span>
<span class="n">test</span> <span class="o">=</span> <span class="n">gradcheck</span><span class="p">(</span><span class="n">linear</span><span class="p">,</span> <span class="nb">input</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="mf">1e-6</span><span class="p">,</span> <span class="n">atol</span><span class="o">=</span><span class="mf">1e-4</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">test</span><span class="p">)</span>
</pre></div>
</div>
<p>See <a class="reference internal" href="../autograd.html#grad-check"><span class="std std-ref">Numerical gradient checking</span></a> for more details on finite-difference gradient comparisons.
If your function is used in higher order derivatives (differentiating the backward pass) you
can use the <code class="docutils literal notranslate"><span class="pre">gradgradcheck</span></code> function from the same package to check higher order derivatives.</p>
</div>
<div class="section" id="extending-torch-nn">
<h2>Extending <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch.nn</span></code><a class="headerlink" href="#extending-torch-nn" title="Permalink to this headline">¶</a></h2>
<p><code class="xref py py-mod docutils literal notranslate"><span class="pre">nn</span></code> exports two kinds of interfaces - modules and their functional
versions. You can extend it in both ways, but we recommend using modules for
all kinds of layers, that hold any parameters or buffers, and recommend using
a functional form parameter-less operations like activation functions, pooling,
etc.</p>
<p>Adding a functional version of an operation is already fully covered in the
section above.</p>
<div class="section" id="adding-a-module">
<h3>Adding a <a class="reference internal" href="../generated/torch.nn.Module.html#torch.nn.Module" title="torch.nn.Module"><code class="xref py py-class docutils literal notranslate"><span class="pre">Module</span></code></a><a class="headerlink" href="#adding-a-module" title="Permalink to this headline">¶</a></h3>
<p>Since <code class="xref py py-mod docutils literal notranslate"><span class="pre">nn</span></code> heavily utilizes <a class="reference internal" href="../autograd.html#module-torch.autograd" title="torch.autograd"><code class="xref py py-mod docutils literal notranslate"><span class="pre">autograd</span></code></a>, adding a new
<a class="reference internal" href="../generated/torch.nn.Module.html#torch.nn.Module" title="torch.nn.Module"><code class="xref py py-class docutils literal notranslate"><span class="pre">Module</span></code></a> requires implementing a <a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a>
that performs the operation and can compute the gradient. From now on let’s
assume that we want to implement a <code class="docutils literal notranslate"><span class="pre">Linear</span></code> module and we have the function
implemented as in the listing above. There’s very little code required to
add this. Now, there are two functions that need to be implemented:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">__init__</span></code> (<em>optional</em>) - takes in arguments such as kernel sizes, numbers
of features, etc. and initializes parameters and buffers.</p></li>
<li><p><a class="reference internal" href="../generated/torch.nn.Module.html#torch.nn.Module.forward" title="torch.nn.Module.forward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">forward()</span></code></a> - instantiates a <a class="reference internal" href="../autograd.html#torch.autograd.Function" title="torch.autograd.Function"><code class="xref py py-class docutils literal notranslate"><span class="pre">Function</span></code></a> and
uses it to perform the operation. It’s very similar to a functional wrapper
shown above.</p></li>
</ul>
<p>This is how a <code class="docutils literal notranslate"><span class="pre">Linear</span></code> module can be implemented:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">Linear</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">input_features</span><span class="p">,</span> <span class="n">output_features</span><span class="p">,</span> <span class="n">bias</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
<span class="nb">super</span><span class="p">(</span><span class="n">Linear</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="bp">self</span><span class="o">.</span><span class="n">input_features</span> <span class="o">=</span> <span class="n">input_features</span>
<span class="bp">self</span><span class="o">.</span><span class="n">output_features</span> <span class="o">=</span> <span class="n">output_features</span>
<span class="c1"># nn.Parameter is a special kind of Tensor, that will get</span>
<span class="c1"># automatically registered as Module's parameter once it's assigned</span>
<span class="c1"># as an attribute. Parameters and buffers need to be registered, or</span>
<span class="c1"># they won't appear in .parameters() (doesn't apply to buffers), and</span>
<span class="c1"># won't be converted when e.g. .cuda() is called. You can use</span>
<span class="c1"># .register_buffer() to register buffers.</span>
<span class="c1"># nn.Parameters require gradients by default.</span>
<span class="bp">self</span><span class="o">.</span><span class="n">weight</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">(</span><span class="n">output_features</span><span class="p">,</span> <span class="n">input_features</span><span class="p">))</span>
<span class="k">if</span> <span class="n">bias</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">bias</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">(</span><span class="n">output_features</span><span class="p">))</span>
<span class="k">else</span><span class="p">:</span>
<span class="c1"># You should always register all possible parameters, but the</span>
<span class="c1"># optional ones can be None if you want.</span>
<span class="bp">self</span><span class="o">.</span><span class="n">register_parameter</span><span class="p">(</span><span class="s1">'bias'</span><span class="p">,</span> <span class="kc">None</span><span class="p">)</span>
<span class="c1"># Not a very smart way to initialize weights</span>
<span class="bp">self</span><span class="o">.</span><span class="n">weight</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">uniform_</span><span class="p">(</span><span class="o">-</span><span class="mf">0.1</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">)</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">bias</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">bias</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">uniform_</span><span class="p">(</span><span class="o">-</span><span class="mf">0.1</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="nb">input</span><span class="p">):</span>
<span class="c1"># See the autograd section for explanation of what happens here.</span>
<span class="k">return</span> <span class="n">LinearFunction</span><span class="o">.</span><span class="n">apply</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">weight</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">bias</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">extra_repr</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="c1"># (Optional)Set the extra information about this module. You can test</span>
<span class="c1"># it by printing an object of this class.</span>
<span class="k">return</span> <span class="s1">'input_features=</span><span class="si">{}</span><span class="s1">, output_features=</span><span class="si">{}</span><span class="s1">, bias=</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
<span class="bp">self</span><span class="o">.</span><span class="n">input_features</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">output_features</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">bias</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span>
<span class="p">)</span>
</pre></div>
</div>
</div>
</div>
<div class="section" id="extending-torch">
<h2>Extending <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code><a class="headerlink" href="#extending-torch" title="Permalink to this headline">¶</a></h2>
<p>You can create custom types that emulate <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> by defining a custom
class with methods that match <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>. But what if you want to be able
to pass these types to functions like <a class="reference internal" href="../generated/torch.add.html#torch.add" title="torch.add"><code class="xref py py-func docutils literal notranslate"><span class="pre">torch.add()</span></code></a> in the top-level
<code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> namespace that accept <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> operands?</p>
<p>If your custom python type defines a method named <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>, PyTorch
will invoke your <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation when an instance of your
custom class is passed to a function in the <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> namespace. This makes
it possible to define custom implementations for any of the functions in the
<code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> namespace which your <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation can call,
allowing your users to make use of your custom type with existing PyTorch
workflows that they have already written for <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>. This works with
“duck” types that are unrelated to <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> as well as user-defined
subclasses of <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>.</p>
<div class="section" id="extending-torch-with-a-tensor-like-type">
<h3>Extending <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> with a <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>-like type<a class="headerlink" href="#extending-torch-with-a-tensor-like-type" title="Permalink to this headline">¶</a></h3>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This functionality is inspired by the NumPy <code class="docutils literal notranslate"><span class="pre">__array_function__</span></code>
protocol. See <a class="reference external" href="https://docs.scipy.org/doc/numpy/user/basics.dispatch.html#basics-dispatch">the NumPy documentation</a>
and <a class="reference external" href="https://numpy.org/neps/nep-0018-array-function-protocol.html">NEP-0018</a> for
more details.</p>
</div>
<p>To make this concrete, let’s begin with a simple example that illustrates the
API dispatch mechanism. We’ll create a custom type that represents a 2D scalar
tensor, parametrized by the order <code class="docutils literal notranslate"><span class="pre">N</span></code> and value along the diagonal entries,
<code class="docutils literal notranslate"><span class="pre">value</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">ScalarTensor</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">N</span><span class="p">,</span> <span class="n">value</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_N</span> <span class="o">=</span> <span class="n">N</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_value</span> <span class="o">=</span> <span class="n">value</span>
<span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"DiagonalTensor(N=</span><span class="si">{}</span><span class="s2">, value=</span><span class="si">{}</span><span class="s2">)"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_N</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">_value</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">tensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_value</span> <span class="o">*</span> <span class="n">torch</span><span class="o">.</span><span class="n">eye</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_N</span><span class="p">)</span>
</pre></div>
</div>
<p>This first iteration of the design isn’t very useful. The main functionality of
<code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> is to provide a more compact string representation of a scalar
tensor than in the base tensor class:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">d</span> <span class="o">=</span> <span class="n">ScalarTensor</span><span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">d</span>
<span class="go">ScalarTensor(N=5, value=2)</span>
<span class="gp">>>> </span><span class="n">d</span><span class="o">.</span><span class="n">tensor</span><span class="p">()</span>
<span class="go">tensor([[2., 0., 0., 0., 0.],</span>
<span class="go"> [0., 2., 0., 0., 0.],</span>
<span class="go"> [0., 0., 2., 0., 0.],</span>
<span class="go"> [0., 0., 0., 2., 0.],</span>
<span class="go"> [0., 0., 0., 0., 2.]])</span>
</pre></div>
</div>
<p>If we try to use this object with the <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API, we will run
into issues:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">import</span> <span class="nn">torch</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">d</span><span class="p">)</span>
<span class="go">TypeError: mean(): argument 'input' (position 1) must be Tensor, not ScalarTensor</span>
</pre></div>
</div>
<p>Adding a <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation to <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> makes it
possible for the above operation to succeed. Let’s re-do our implementation,
this time adding a <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">HANDLED_FUNCTIONS</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">class</span> <span class="nc">ScalarTensor</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">N</span><span class="p">,</span> <span class="n">value</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_N</span> <span class="o">=</span> <span class="n">N</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_value</span> <span class="o">=</span> <span class="n">value</span>
<span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"DiagonalTensor(N=</span><span class="si">{}</span><span class="s2">, value=</span><span class="si">{}</span><span class="s2">)"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_N</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">_value</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">tensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_value</span> <span class="o">*</span> <span class="n">torch</span><span class="o">.</span><span class="n">eye</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_N</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">__torch_function__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">func</span><span class="p">,</span> <span class="n">types</span><span class="p">,</span> <span class="n">args</span><span class="o">=</span><span class="p">(),</span> <span class="n">kwargs</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">if</span> <span class="n">kwargs</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">kwargs</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">if</span> <span class="n">func</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">HANDLED_FUNCTIONS</span> <span class="ow">or</span> <span class="ow">not</span> <span class="nb">all</span><span class="p">(</span>
<span class="nb">issubclass</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">ScalarTensor</span><span class="p">))</span>
<span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="n">types</span>
<span class="p">):</span>
<span class="k">return</span> <span class="bp">NotImplemented</span>
<span class="k">return</span> <span class="n">HANDLED_FUNCTIONS</span><span class="p">[</span><span class="n">func</span><span class="p">](</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
</pre></div>
</div>
<p>The <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> method takes four arguments: <code class="docutils literal notranslate"><span class="pre">func</span></code>, a reference
to the torch API function that is being overridden, <code class="docutils literal notranslate"><span class="pre">types</span></code>, the list of
types of Tensor-likes that implement <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>, <code class="docutils literal notranslate"><span class="pre">args</span></code>, the
tuple of arguments passed to the function, and <code class="docutils literal notranslate"><span class="pre">kwargs</span></code>, the dict of keyword
arguments passed to the function. It uses a global dispatch stable named
<code class="docutils literal notranslate"><span class="pre">HANDLED_FUNCTIONS</span></code> to store custom implementations. The keys of this
dictionary are functions in the <code class="docutils literal notranslate"><span class="pre">torch</span></code> namespace and the values are
implementations for <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Using a global dispatch table is not a mandated part of the
<code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> API, it is just a useful design pattern for
structuring your override implementations.</p>
</div>
<p>This class definition isn’t quite enough to make <code class="docutils literal notranslate"><span class="pre">torch.mean</span></code> do the right
thing when we pass it a <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> – we also need to define an
implementation for <code class="docutils literal notranslate"><span class="pre">torch.mean</span></code> for <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> operands and add the
implementation to the <code class="docutils literal notranslate"><span class="pre">HANDLED_FUNCTIONS</span></code> dispatch table dictionary. One way
of doing this is to define a decorator:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">functools</span>
<span class="k">def</span> <span class="nf">implements</span><span class="p">(</span><span class="n">torch_function</span><span class="p">):</span>
<span class="sd">"""Register a torch function override for ScalarTensor"""</span>
<span class="nd">@functools</span><span class="o">.</span><span class="n">wraps</span><span class="p">(</span><span class="n">torch_function</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">decorator</span><span class="p">(</span><span class="n">func</span><span class="p">):</span>
<span class="n">HANDLED_FUNCTIONS</span><span class="p">[</span><span class="n">torch_function</span><span class="p">]</span> <span class="o">=</span> <span class="n">func</span>
<span class="k">return</span> <span class="n">func</span>
<span class="k">return</span> <span class="n">decorator</span>
</pre></div>
</div>
<p>which can be applied to the implementation of our override:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@implements</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">mean</span><span class="p">(</span><span class="nb">input</span><span class="p">):</span>
<span class="k">return</span> <span class="nb">float</span><span class="p">(</span><span class="nb">input</span><span class="o">.</span><span class="n">_value</span><span class="p">)</span> <span class="o">/</span> <span class="nb">input</span><span class="o">.</span><span class="n">_N</span>
</pre></div>
</div>
<p>With this change we can now use <code class="docutils literal notranslate"><span class="pre">torch.mean</span></code> with <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">d</span> <span class="o">=</span> <span class="n">ScalarTensor</span><span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">d</span><span class="p">)</span>
<span class="go">0.4</span>
</pre></div>
</div>
<p>Of course <code class="docutils literal notranslate"><span class="pre">torch.mean</span></code> is an example of the simplest kind of function to
override since it only takes one operand. We can use the same machinery to
override a function that takes more than one operand, any one of which might be
a tensor or tensor-like that defines <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>, for example for
<a class="reference internal" href="../generated/torch.add.html#torch.add" title="torch.add"><code class="xref py py-func docutils literal notranslate"><span class="pre">torch.add()</span></code></a>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">ensure_tensor</span><span class="p">(</span><span class="n">data</span><span class="p">):</span>
<span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">ScalarTensor</span><span class="p">):</span>
<span class="k">return</span> <span class="n">data</span><span class="o">.</span><span class="n">tensor</span><span class="p">()</span>
<span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">as_tensor</span><span class="p">(</span><span class="n">data</span><span class="p">)</span>
<span class="nd">@implements</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">add</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="n">other</span><span class="p">):</span>
<span class="k">try</span><span class="p">:</span>
<span class="k">if</span> <span class="nb">input</span><span class="o">.</span><span class="n">_N</span> <span class="o">==</span> <span class="n">other</span><span class="o">.</span><span class="n">_N</span><span class="p">:</span>
<span class="k">return</span> <span class="n">ScalarTensor</span><span class="p">(</span><span class="nb">input</span><span class="o">.</span><span class="n">_N</span><span class="p">,</span> <span class="nb">input</span><span class="o">.</span><span class="n">_value</span> <span class="o">+</span> <span class="n">other</span><span class="o">.</span><span class="n">_value</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Shape mismatch!"</span><span class="p">)</span>
<span class="k">except</span> <span class="ne">AttributeError</span><span class="p">:</span>
<span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">ensure_tensor</span><span class="p">(</span><span class="nb">input</span><span class="p">),</span> <span class="n">ensure_tensor</span><span class="p">(</span><span class="n">other</span><span class="p">))</span>
</pre></div>
</div>
<p>This version has a fast path for when both operands are <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code>
instances and also a slower path which degrades to converting the data to
tensors when either operand is not a <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code>. That makes the override
function correctly when either operand is a <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> or a regular
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">s</span> <span class="o">=</span> <span class="n">ScalarTensor</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">s</span><span class="p">,</span> <span class="n">s</span><span class="p">)</span>
<span class="go">DiagonalTensor(N=2, value=4)</span>
<span class="gp">>>> </span><span class="n">t</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">([[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">,],</span> <span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">]])</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">s</span><span class="p">,</span> <span class="n">t</span><span class="p">)</span>
<span class="go">tensor([[3., 1.],</span>
<span class="go"> [1., 3.]])</span>
</pre></div>
</div>
<p>Note that our implementation of <code class="docutils literal notranslate"><span class="pre">add</span></code> does not take <code class="docutils literal notranslate"><span class="pre">alpha</span></code> or <code class="docutils literal notranslate"><span class="pre">out</span></code> as
keyword arguments like <a class="reference internal" href="../generated/torch.add.html#torch.add" title="torch.add"><code class="xref py py-func docutils literal notranslate"><span class="pre">torch.add()</span></code></a> does:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">s</span><span class="p">,</span> <span class="n">s</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="go">TypeError: add() got an unexpected keyword argument 'alpha'</span>
</pre></div>
</div>
<p>For speed and flexibility the <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> dispatch mechanism does not
check that the signature of an override function matches the signature of the
function being overrided in the <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API. For some applications ignoring
optional arguments would be fine but to ensure full compatibility with
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>, user implementations of torch API functions should take care to
exactly emulate the API of the function that is being overrided.</p>
<p>Functions in the <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API that do not have explicit overrides will
return <code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code> from <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>. If all operands with
<code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> defined on them return <code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code>, PyTorch will
raise a <code class="docutils literal notranslate"><span class="pre">TypeError</span></code>. This means that most of the time operations that do not
have explicit overrides for a type will raise a <code class="docutils literal notranslate"><span class="pre">TypeError</span></code> when an instance
of such a type is passed:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">mul</span><span class="p">(</span><span class="n">s</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span>
<span class="go">TypeError: no implementation found for 'torch.mul' on types that</span>
<span class="go">implement __torch_function__: [ScalarTensor]</span>
</pre></div>
</div>
<p>In practice this means that if you would like to implement your overrides using
a <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation along these lines, you will need to
explicitly implement the full <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API or the entire subset of the API
that you care about for your use case. This may be a tall order as the full
<code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API is quite extensive.</p>
<p>Another option is to not return <code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code> for operations that are not
handled but to instead pass a <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> to the original <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code>
function when no override is available. For example, if we change our
implementation of <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> for <code class="docutils literal notranslate"><span class="pre">ScalarTensor</span></code> to the one below:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">__torch_function__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">func</span><span class="p">,</span> <span class="n">types</span><span class="p">,</span> <span class="n">args</span><span class="o">=</span><span class="p">(),</span> <span class="n">kwargs</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">if</span> <span class="n">kwargs</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">kwargs</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">if</span> <span class="n">func</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">HANDLED_FUNCTIONS</span> <span class="ow">or</span> <span class="ow">not</span> <span class="nb">all</span><span class="p">(</span>
<span class="nb">issubclass</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">ScalarTensor</span><span class="p">))</span>
<span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="n">types</span>
<span class="p">):</span>
<span class="n">args</span> <span class="o">=</span> <span class="p">[</span><span class="n">a</span><span class="o">.</span><span class="n">tensor</span><span class="p">()</span> <span class="k">if</span> <span class="nb">hasattr</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="s1">'tensor'</span><span class="p">)</span> <span class="k">else</span> <span class="n">a</span> <span class="k">for</span> <span class="n">a</span> <span class="ow">in</span> <span class="n">args</span><span class="p">]</span>
<span class="k">return</span> <span class="n">func</span><span class="p">(</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">return</span> <span class="n">HANDLED_FUNCTIONS</span><span class="p">[</span><span class="n">func</span><span class="p">](</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
</pre></div>
</div>
<p>Then <a class="reference internal" href="../generated/torch.mul.html#torch.mul" title="torch.mul"><code class="xref py py-func docutils literal notranslate"><span class="pre">torch.mul()</span></code></a> will work correctly, although the return type will always
be a <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> rather than a <code class="xref py py-class docutils literal notranslate"><span class="pre">ScalarTensor</span></code>, even if both operands
are <code class="xref py py-class docutils literal notranslate"><span class="pre">ScalarTensor</span></code> instances:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">s</span> <span class="o">=</span> <span class="n">ScalarTensor</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">mul</span><span class="p">(</span><span class="n">s</span><span class="p">,</span> <span class="n">s</span><span class="p">)</span>
<span class="go">tensor([[4., 0.],</span>
<span class="go"> [0., 4.]])</span>
</pre></div>
</div>
<p>Also see the <code class="docutils literal notranslate"><span class="pre">MetadataTensor</span></code> example below for another variation on this
pattern but instead always returns a <code class="docutils literal notranslate"><span class="pre">MetadataTensor</span></code> to propagate metadata
through operations in the <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API.</p>
<p>The <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> protocol is designed for full coverage of the API,
partial coverage may lead to undesirable results, in particular, certain
functions raising a <code class="docutils literal notranslate"><span class="pre">TypeError</span></code>. This is especially true for subclasses,
where all three of <cite>torch.add</cite>, <cite>torch.Tensor.__add__</cite> and <cite>torch.Tensor.add</cite>
must be covered, even if they return exactly the same result. Failing to do
this may also lead to infinite recursion. If one requires the implementation
of a function from <code class="docutils literal notranslate"><span class="pre">torch.Tensor</span></code> subclasses, they must use
<code class="docutils literal notranslate"><span class="pre">super().__torch_function__</span></code> inside their implementation.</p>
</div>
<div class="section" id="subclassing-torch-tensor">
<h3>Subclassing <code class="docutils literal notranslate"><span class="pre">torch.Tensor</span></code><a class="headerlink" href="#subclassing-torch-tensor" title="Permalink to this headline">¶</a></h3>
<p>As of version 1.7.0, methods and functions applied on <code class="docutils literal notranslate"><span class="pre">torch.Tensor</span></code> subclasses
will return subclass instances instead of <code class="docutils literal notranslate"><span class="pre">torch.Tensor</span></code> instances:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="k">class</span> <span class="nc">SubTensor</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">):</span>
<span class="gp">... </span> <span class="k">pass</span>
<span class="gp">>>> </span><span class="nb">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">SubTensor</span><span class="p">([</span><span class="mi">0</span><span class="p">]),</span> <span class="n">SubTensor</span><span class="p">([</span><span class="mi">1</span><span class="p">])))</span><span class="o">.</span><span class="vm">__name__</span>
<span class="go">'SubTensor'</span>
<span class="gp">>>> </span><span class="nb">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">SubTensor</span><span class="p">([</span><span class="mi">0</span><span class="p">]),</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">([</span><span class="mi">1</span><span class="p">])))</span><span class="o">.</span><span class="vm">__name__</span>
<span class="go">'SubTensor'</span>
</pre></div>
</div>
<p>If multiple subclasses exist, the lowest one in the hierarchy will be chosen by
default. If there is no unique way to determine such a case, then a
<code class="docutils literal notranslate"><span class="pre">TypeError</span></code> is raised:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="nb">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">SubTensor2</span><span class="p">([</span><span class="mi">0</span><span class="p">]),</span> <span class="n">SubTensor</span><span class="p">([</span><span class="mi">1</span><span class="p">])))</span><span class="o">.</span><span class="vm">__name__</span>
<span class="go">'SubTensor2'</span>
<span class="gp">>>> </span><span class="nb">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">SubTensor2</span><span class="p">([</span><span class="mi">0</span><span class="p">]),</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">([</span><span class="mi">1</span><span class="p">])))</span><span class="o">.</span><span class="vm">__name__</span>
<span class="go">'SubTensor2'</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">SubTensor</span><span class="p">([</span><span class="mi">0</span><span class="p">]),</span> <span class="n">OtherSubTensor</span><span class="p">([</span><span class="mi">1</span><span class="p">]))</span>
<span class="gt">Traceback (most recent call last):</span>
File <span class="nb">"<stdin>"</span>, line <span class="m">1</span>, in <span class="n"><module></span>
<span class="gr">TypeError</span>: <span class="n">no implementation found for 'torch.add' on types that implement __torch_function__: [SubTensor, OtherSubTensor]</span>
</pre></div>
</div>
<p>If one wishes to have a global override for all tensor methods, one can use
<code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>. Here is an example that logs all function/method
calls:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">LoggingTensor</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">):</span>
<span class="nd">@classmethod</span>
<span class="k">def</span> <span class="nf">__torch_function__</span><span class="p">(</span><span class="bp">cls</span><span class="p">,</span> <span class="n">func</span><span class="p">,</span> <span class="n">types</span><span class="p">,</span> <span class="n">args</span><span class="o">=</span><span class="p">(),</span> <span class="n">kwargs</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">logging</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="sa">f</span><span class="s2">"func: </span><span class="si">{</span><span class="n">func</span><span class="o">.</span><span class="vm">__name__</span><span class="si">}</span><span class="s2">, args: </span><span class="si">{</span><span class="n">args</span><span class="si">!r}</span><span class="s2">, kwargs: </span><span class="si">{</span><span class="n">kwargs</span><span class="si">!r}</span><span class="s2">"</span><span class="p">)</span>
<span class="k">if</span> <span class="n">kwargs</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">kwargs</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">return</span> <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__torch_function__</span><span class="p">(</span><span class="n">func</span><span class="p">,</span> <span class="n">types</span><span class="p">,</span> <span class="n">args</span><span class="p">,</span> <span class="n">kwargs</span><span class="p">)</span>
</pre></div>
</div>
<p>However, if one instead wishes to override a method on the Tensor subclass,
there one can do so either by directly overriding the method (by defining
it for a subclass), or by using <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> and matching with
<code class="docutils literal notranslate"><span class="pre">func</span></code>.</p>
<p>One should be careful within <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> for subclasses to always
call <code class="docutils literal notranslate"><span class="pre">super().__torch_function__(func,</span> <span class="pre">...)</span></code> instead of <code class="docutils literal notranslate"><span class="pre">func</span></code> directly,
as was the case before version 1.7.0. Failing to do this may cause <code class="docutils literal notranslate"><span class="pre">func</span></code>
to recurse back into <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> and therefore cause infinite
recursion.</p>
</div>
<div class="section" id="extending-torch-with-a-tensor-wrapper-type">
<h3>Extending <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> with a <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> wrapper type<a class="headerlink" href="#extending-torch-with-a-tensor-wrapper-type" title="Permalink to this headline">¶</a></h3>
<p>Another useful case is a type that wraps a <code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code>, either as an
attribute or via subclassing. Below we implement a special case of this sort of
type, a <code class="docutils literal notranslate"><span class="pre">MetadataTensor</span></code> that attaches a dictionary of metadata to a
<code class="xref py py-class docutils literal notranslate"><span class="pre">Tensor</span></code> that is propagated through <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> operations. Since this
is a generic sort of wrapping for the full <code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API, we do not need to
individually implement each override so we can make the <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>
implementation more permissive about what operations are allowed:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">class</span> <span class="nc">MetadataTensor</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">metadata</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_t</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">as_tensor</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_metadata</span> <span class="o">=</span> <span class="n">metadata</span>
<span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"Metadata:</span><span class="se">\n</span><span class="si">{}</span><span class="se">\n\n</span><span class="s2">data:</span><span class="se">\n</span><span class="si">{}</span><span class="s2">"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_metadata</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">_t</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">__torch_function__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">func</span><span class="p">,</span> <span class="n">types</span><span class="p">,</span> <span class="n">args</span><span class="o">=</span><span class="p">(),</span> <span class="n">kwargs</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">if</span> <span class="n">kwargs</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">kwargs</span> <span class="o">=</span> <span class="p">{}</span>
<span class="n">args</span> <span class="o">=</span> <span class="p">[</span><span class="n">a</span><span class="o">.</span><span class="n">_t</span> <span class="k">if</span> <span class="nb">hasattr</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="s1">'_t'</span><span class="p">)</span> <span class="k">else</span> <span class="n">a</span> <span class="k">for</span> <span class="n">a</span> <span class="ow">in</span> <span class="n">args</span><span class="p">]</span>
<span class="n">ret</span> <span class="o">=</span> <span class="n">func</span><span class="p">(</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">return</span> <span class="n">MetadataTensor</span><span class="p">(</span><span class="n">ret</span><span class="p">,</span> <span class="n">metadata</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">_metadata</span><span class="p">)</span>
</pre></div>
</div>
<p>This simple implementation won’t necessarily work with every function in the
<code class="xref py py-mod docutils literal notranslate"><span class="pre">torch</span></code> API but it is good enough to capture most common operations:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">metadata</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'owner'</span><span class="p">:</span> <span class="s1">'Ministry of Silly Walks'</span><span class="p">}</span>
<span class="gp">>>> </span><span class="n">m</span> <span class="o">=</span> <span class="n">MetadataTensor</span><span class="p">([[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">],</span> <span class="p">[</span><span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">]],</span> <span class="n">metadata</span><span class="o">=</span><span class="n">metadata</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">t</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">([[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">],</span> <span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">]]])</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="n">m</span><span class="p">)</span>
<span class="go">Metadata:</span>
<span class="go">{'owner': 'Ministry of Silly Walks'}</span>
<span class="go">data:</span>
<span class="go">tensor([[2, 4],</span>
<span class="go"> [4, 6]])</span>
<span class="gp">>>> </span><span class="n">torch</span><span class="o">.</span><span class="n">mul</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="n">m</span><span class="p">)</span>
<span class="go">Metadata:</span>
<span class="go">{'owner': 'Ministry of Silly Walks'}</span>
<span class="go">data:</span>
<span class="go">tensor([[1, 4],</span>
<span class="go"> [3, 8]])</span>
</pre></div>
</div>
</div>
<div class="section" id="operations-on-multiple-types-that-define-torch-function">
<h3>Operations on multiple types that define <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code><a class="headerlink" href="#operations-on-multiple-types-that-define-torch-function" title="Permalink to this headline">¶</a></h3>
<p>It is possible to use the torch API with multiple distinct types that each have
a <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementation, but special care must be taken. In such
a case the rules are:</p>
<ul class="simple">
<li><p>The dispatch operation gathers all distinct implementations of
<code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> for each operand and calls them in order: subclasses
before superclasses, and otherwise left to right in the operator expression.</p></li>
<li><p>If any value other than <code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code> is returned, that value is
returned as the result. Implementations can register that they do not
implement an operation by returning <code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code>.</p></li>
<li><p>If all of the <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> implementations return
<code class="docutils literal notranslate"><span class="pre">NotImplemented</span></code>, PyTorch raises a <code class="docutils literal notranslate"><span class="pre">TypeError</span></code>.</p></li>
</ul>
</div>
<div class="section" id="testing-coverage-of-overrides-for-the-pytorch-api">
<h3>Testing Coverage of Overrides for the PyTorch API<a class="headerlink" href="#testing-coverage-of-overrides-for-the-pytorch-api" title="Permalink to this headline">¶</a></h3>
<p>One troublesome aspect of implementing <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> is that if some
operations do and others do not have overrides, users will at best see an
inconsistent experience, or at worst will see errors raised at runtime when they
use a function that does not have an override. To ease this process, PyTorch
provides a developer-facing API for ensuring full support for
<code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code> overrides. This API is private and may be subject to
changes without warning in the future.</p>
<p>First, to get a listing of all overridable functions, use
<code class="docutils literal notranslate"><span class="pre">torch._overrides.get_overridable_functions</span></code>. This returns a dictionary whose
keys are namespaces in the <code class="docutils literal notranslate"><span class="pre">PyTorch</span></code> Python API and whose values are a list of
functions in that namespace that can be overriden. For example, let’s print the
names of the first 5 functions in <code class="docutils literal notranslate"><span class="pre">torch.nn.functional</span></code> that can be
overriden:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">torch._overrides</span> <span class="kn">import</span> <span class="n">get_overridable_functions</span>
<span class="gp">>>> </span><span class="n">func_dict</span> <span class="o">=</span> <span class="n">get_overridable_functions</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">nn_funcs</span> <span class="o">=</span> <span class="n">func_dict</span><span class="p">[</span><span class="n">torch</span><span class="o">.</span><span class="n">nn</span><span class="o">.</span><span class="n">functional</span><span class="p">]</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">([</span><span class="n">f</span><span class="o">.</span><span class="vm">__name__</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">nn_funcs</span><span class="p">[:</span><span class="mi">5</span><span class="p">])</span>
<span class="go">['adaptive_avg_pool1d', 'adaptive_avg_pool2d', 'adaptive_avg_pool3d',</span>
<span class="go"> 'adaptive_max_pool1d', 'adaptive_max_pool1d_with_indices']</span>
</pre></div>
</div>
<p>This listing of functions makes it possible to iterate over all overridable
functions, however in practice this is not enough to write tests for all of
these functions without laboriously and manually copying the signature of each
function for each test. To ease this process, the
<code class="docutils literal notranslate"><span class="pre">torch._overrides.get_testing_overrides</span></code> function returns a dictionary mapping
overridable functions in the <code class="docutils literal notranslate"><span class="pre">PyTorch</span></code> API to dummy lambda functions that have
the same signature as the original function but unconditionally return -1. These
functions are most useful to use with <code class="docutils literal notranslate"><span class="pre">inspect</span></code> to analyze the function
signature of the original <code class="docutils literal notranslate"><span class="pre">PyTorch</span></code> function:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">import</span> <span class="nn">inspect</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">torch._overrides</span> <span class="kn">import</span> <span class="n">get_testing_overrides</span>
<span class="gp">>>> </span><span class="n">override_dict</span> <span class="o">=</span> <span class="n">get_testing_overrides</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">dummy_add</span> <span class="o">=</span> <span class="n">override_dict</span><span class="p">[</span><span class="n">torch</span><span class="o">.</span><span class="n">add</span><span class="p">]</span>
<span class="gp">>>> </span><span class="n">inspect</span><span class="o">.</span><span class="n">signature</span><span class="p">(</span><span class="n">dummy_add</span><span class="p">)</span>
<span class="go"><Signature (input, other, out=None)></span>
</pre></div>
</div>
<p>Finally, <code class="docutils literal notranslate"><span class="pre">torch._overrides.get_ignored_functions</span></code> returns a tuple of functions
that explicitly cannot be overrided by <code class="docutils literal notranslate"><span class="pre">__torch_function__</span></code>. This list can be
useful to confirm that a function that isn’t present in the dictionary returned
by <code class="docutils literal notranslate"><span class="pre">get_overridable_functions</span></code> cannot be overriden.</p>
</div>
</div>
<div class="section" id="writing-custom-c-extensions">
<h2>Writing custom C++ extensions<a class="headerlink" href="#writing-custom-c-extensions" title="Permalink to this headline">¶</a></h2>
<p>See this
<a class="reference external" href="https://pytorch.org/tutorials/advanced/cpp_extension.html">PyTorch tutorial</a>
for a detailed explanation and examples.</p>
<p>Documentations are available at <a class="reference internal" href="../cpp_extension.html"><span class="doc">torch.utils.cpp_extension</span></a>.</p>
</div>
<div class="section" id="id1">
<h2>Writing custom C extensions<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h2>
<p>Example available at
<a class="reference external" href="https://github.com/pytorch/extension-ffi">this GitHub repository</a>.</p>
</div>
</div>
</article>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="faq.html" class="btn btn-neutral float-right" title="Frequently Asked Questions" accesskey="n" rel="next">Next <img src="../_static/images/chevron-right-orange.svg" class="next-page"></a>
<a href="ddp.html" class="btn btn-neutral" title="Distributed Data Parallel" accesskey="p" rel="prev"><img src="../_static/images/chevron-right-orange.svg" class="previous-page"> Previous</a>
</div>
<hr>