forked from pytorch/pytorch.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathnn
2547 lines (2207 loc) · 114 KB
/
nn
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
<h1 id="torch-nn">torch.nn</h1>
<h2 id="avgpool2d">AvgPool2d</h2>
<p>Applies a 2D average pooling over an input signal composed of several input</p>
<pre class="highlight python"><code><span class="n">The</span> <span class="n">output</span> <span class="n">value</span> <span class="n">of</span> <span class="n">the</span> <span class="n">layer</span> <span class="k">with</span> <span class="nb">input</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">C</span> <span class="n">x</span> <span class="n">H</span> <span class="n">x</span> <span class="n">W</span><span class="p">)</span> <span class="ow">and</span> <span class="n">output</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">C</span> <span class="n">x</span> <span class="n">oH</span> <span class="n">x</span> <span class="n">oW</span><span class="p">)</span>
<span class="n">can</span> <span class="n">be</span> <span class="n">precisely</span> <span class="n">described</span> <span class="k">as</span><span class="p">:</span>
<span class="n">output</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">c_i</span><span class="p">][</span><span class="n">h_i</span><span class="p">][</span><span class="n">w_i</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="mi">1</span> <span class="o">/</span> <span class="n">K</span><span class="p">)</span> <span class="o">*</span> <span class="n">sum_</span><span class="p">{</span><span class="n">kh</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">KH</span><span class="p">}</span> <span class="n">sum_</span><span class="p">{</span><span class="n">kw</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">kW</span><span class="p">}</span> <span class="nb">input</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">c_i</span><span class="p">][</span><span class="n">stride_h</span> <span class="o">*</span> <span class="n">h_i</span> <span class="o">+</span> <span class="n">kh</span><span class="p">)][</span><span class="n">stride_w</span> <span class="o">*</span> <span class="n">w_i</span> <span class="o">+</span> <span class="n">kw</span><span class="p">)]</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># pool of square window of size=3, stride=2</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">AvgPool2d</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="n">stride</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="c"># pool of non-square window</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">AvgPool2d</span><span class="p">((</span><span class="mi">3</span><span class="p">,</span> <span class="mi">2</span><span class="p">),</span> <span class="n">stride</span><span class="o">=</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">1</span><span class="p">))</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">50</span><span class="p">,</span> <span class="mi">32</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the window. Can be a single number k (for a square kernel of k x k) or a tuple (kh x kw)</td>
</tr>
<tr>
<td>stride</td>
<td>kernel_size</td>
<td>the stride of the window. Can be a single number s or a tuple (sh x sw).</td>
</tr>
<tr>
<td>padding</td>
<td>0</td>
<td>implicit padding to be added. Can be a single number or a tuple.</td>
</tr>
<tr>
<td>ceil_mode</td>
<td></td>
<td>when True, will use “ceil” instead of “floor” to compute the output shape</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , * , *, * ]</td>
<td>Input is minibatch x channels x iH x iW</td>
</tr>
<tr>
<td>output</td>
<td>[ * , * , *, * ]</td>
<td>Output shape = minibatch x channels x floor((iH + 2*padH - kH) / sH + 1) x floor((iW + 2*padW - kW) / sW + 1)</td>
</tr>
</tbody></table>
<h2 id="avgpool3d">AvgPool3d</h2>
<p>Applies a 3D average pooling over an input signal composed of several input</p>
<pre class="highlight python"><code><span class="c"># pool of square window of size=3, stride=2</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">AvgPool3d</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="n">stride</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="c"># pool of non-square window</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">AvgPool3d</span><span class="p">((</span><span class="mi">3</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="n">stride</span><span class="o">=</span><span class="p">(</span><span class="mi">2</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="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">50</span><span class="p">,</span><span class="mi">44</span><span class="p">,</span> <span class="mi">31</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the window to take a average over. Can be a single number k (for a square kernel of k x k x k) or a tuple (kt x kh x kw)</td>
</tr>
<tr>
<td>stride</td>
<td>kernel_size</td>
<td>the stride of the window. Can be a single number s or a tuple (st x sh x sw).</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , * , *, *, * ]</td>
<td>Input is minibatch x channels x iT x iH x iW</td>
</tr>
<tr>
<td>output</td>
<td>[ * , * , *, *, * ]</td>
<td>Output shape = minibatch x channels x floor((iT + 2*padT - kT) / sT + 1) x floor((iH + 2*padH - kH) / sH + 1) x floor((iW + 2*padW - kW) / sW + 1)</td>
</tr>
</tbody></table>
<h2 id="batchnorm1d">BatchNorm1d</h2>
<p>Applies Batch Normalization over a 2d input that is seen as a mini-batch of 1d inputs</p>
<pre class="highlight python"><code> <span class="n">x</span> <span class="o">-</span> <span class="n">mean</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">y</span> <span class="o">=</span> <span class="o">-----------------------------</span> <span class="o">*</span> <span class="n">gamma</span> <span class="o">+</span> <span class="n">beta</span>
<span class="n">standard_deviation</span><span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="o">+</span> <span class="n">eps</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># With Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm1d</span><span class="p">(</span><span class="mi">100</span><span class="p">)</span>
<span class="c"># Without Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm1d</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="n">affine</span><span class="o">=</span><span class="bp">False</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">100</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The mean and standard-deviation are calculated per-dimension over
the mini-batches and gamma and beta are learnable parameter vectors
of size N (where N is the input size).</p>
<p>During training, this layer keeps a running estimate of its computed mean
and variance. The running sum is kept with a default momentum of 0.1
During evaluation, this running mean/variance is used for normalization.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>num_features</td>
<td></td>
<td>the size of each 1D input in the mini-batch</td>
</tr>
<tr>
<td>eps</td>
<td>1e-5</td>
<td>a value added to the denominator for numerical stability.</td>
</tr>
<tr>
<td>momentum</td>
<td>0.1</td>
<td>the value used for the running_mean and running_var computation.</td>
</tr>
<tr>
<td>affine</td>
<td></td>
<td>a boolean value that when set to true, gives the layer learnable affine parameters.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , num_features ]</td>
<td>2D Tensor of nBatches x num_features</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output has the same shape as input</td>
</tr>
</tbody></table>
<h3 id="returns">Returns</h3>
<p>a normalized tensor in the batch dimension</p>
<h2 id="batchnorm2d">BatchNorm2d</h2>
<p>Applies Batch Normalization over a 4d input that is seen as a mini-batch of 3d inputs</p>
<pre class="highlight python"><code> <span class="n">x</span> <span class="o">-</span> <span class="n">mean</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">y</span> <span class="o">=</span> <span class="o">-----------------------------</span> <span class="o">*</span> <span class="n">gamma</span> <span class="o">+</span> <span class="n">beta</span>
<span class="n">standard_deviation</span><span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="o">+</span> <span class="n">eps</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># With Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm2d</span><span class="p">(</span><span class="mi">100</span><span class="p">)</span>
<span class="c"># Without Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm2d</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="n">affine</span><span class="o">=</span><span class="bp">False</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">100</span><span class="p">,</span> <span class="mi">35</span><span class="p">,</span> <span class="mi">45</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The mean and standard-deviation are calculated per-dimension over
the mini-batches and gamma and beta are learnable parameter vectors
of size N (where N is the input size).</p>
<p>During training, this layer keeps a running estimate of its computed mean
and variance. The running sum is kept with a default momentum of 0.1
During evaluation, this running mean/variance is used for normalization.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>num_features</td>
<td></td>
<td>num_features from an expected input of size batch_size x num_features x height x width</td>
</tr>
<tr>
<td>eps</td>
<td>1e-5</td>
<td>a value added to the denominator for numerical stability.</td>
</tr>
<tr>
<td>momentum</td>
<td>0.1</td>
<td>the value used for the running_mean and running_var computation.</td>
</tr>
<tr>
<td>affine</td>
<td></td>
<td>a boolean value that when set to true, gives the layer learnable affine parameters.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , num_features , *, * ]</td>
<td>4D Tensor of batch_size x num_features x height x width</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output has the same shape as input</td>
</tr>
</tbody></table>
<h3 id="returns">Returns</h3>
<p>a normalized tensor in the batch dimension</p>
<h2 id="batchnorm3d">BatchNorm3d</h2>
<p>Applies Batch Normalization over a 5d input that is seen as a mini-batch of 4d inputs</p>
<pre class="highlight python"><code> <span class="n">x</span> <span class="o">-</span> <span class="n">mean</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">y</span> <span class="o">=</span> <span class="o">-----------------------------</span> <span class="o">*</span> <span class="n">gamma</span> <span class="o">+</span> <span class="n">beta</span>
<span class="n">standard_deviation</span><span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="o">+</span> <span class="n">eps</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># With Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm3d</span><span class="p">(</span><span class="mi">100</span><span class="p">)</span>
<span class="c"># Without Learnable Parameters</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm3d</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="n">affine</span><span class="o">=</span><span class="bp">False</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">100</span><span class="p">,</span> <span class="mi">35</span><span class="p">,</span> <span class="mi">45</span><span class="p">,</span> <span class="mi">10</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The mean and standard-deviation are calculated per-dimension over
the mini-batches and gamma and beta are learnable parameter vectors
of size N (where N is the input size).</p>
<p>During training, this layer keeps a running estimate of its computed mean
and variance. The running sum is kept with a default momentum of 0.1
During evaluation, this running mean/variance is used for normalization.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>num_features</td>
<td></td>
<td>num_features from an expected input of size batch_size x num_features x height x width</td>
</tr>
<tr>
<td>eps</td>
<td>1e-5</td>
<td>a value added to the denominator for numerical stability.</td>
</tr>
<tr>
<td>momentum</td>
<td>0.1</td>
<td>the value used for the running_mean and running_var computation.</td>
</tr>
<tr>
<td>affine</td>
<td></td>
<td>a boolean value that when set to true, gives the layer learnable affine parameters.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , num_features , * , * , * ]</td>
<td>5D Tensor of batch_size x num_features x depth x height x width</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output has the same shape as input</td>
</tr>
</tbody></table>
<h3 id="returns">Returns</h3>
<p>a normalized tensor in the batch dimension</p>
<h2 id="container">Container</h2>
<p>This is the base container class for all neural networks you would define.</p>
<pre class="highlight python"><code><span class="c"># Example of using Container</span>
<span class="k">class</span> <span class="nc">Net</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Container</span><span class="p">):</span>
<span class="k">def</span> <span class="nf">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="nb">super</span><span class="p">(</span><span class="n">Net</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="n">__init__</span><span class="p">(</span>
<span class="n">conv1</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv2d</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">20</span><span class="p">,</span> <span class="mi">5</span><span class="p">),</span>
<span class="n">relu</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">ReLU</span><span class="p">()</span>
<span class="p">)</span>
<span class="k">def</span> <span class="nf">__call__</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="n">output</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">conv1</span><span class="p">(</span><span class="n">x</span><span class="p">))</span>
<span class="k">return</span> <span class="n">output</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">Net</span><span class="p">()</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># one can add modules to the container after construction</span>
<span class="n">model</span><span class="o">.</span><span class="n">add_module</span><span class="p">(</span><span class="s">'pool1'</span><span class="p">,</span> <span class="n">nn</span><span class="o">.</span><span class="n">MaxPool2d</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
</code></pre>
<p>You will subclass your container from this class.
In the constructor you define the modules that you would want to use,
and in the <strong>call</strong> function you use the constructed modules in
your operations.</p>
<p>To make it easier to understand, given is a small example.</p>
<p>One can also add new modules to a container after construction.
You can do this with the add_module function.</p>
<p>The container has one additional method <code class="prettyprint">parameters()</code> which
returns the list of learnable parameters in the container instance.</p>
<h2 id="conv1d">Conv1d</h2>
<p>Applies a 1D convolution over an input signal composed of several input</p>
<pre class="highlight python"><code><span class="n">The</span> <span class="n">output</span> <span class="n">value</span> <span class="n">of</span> <span class="n">the</span> <span class="n">layer</span> <span class="k">with</span> <span class="nb">input</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">iC</span> <span class="n">x</span> <span class="n">W</span><span class="p">)</span> <span class="ow">and</span> <span class="n">output</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">oC</span> <span class="n">x</span> <span class="n">oW</span><span class="p">)</span>
<span class="n">can</span> <span class="n">be</span> <span class="n">precisely</span> <span class="n">described</span> <span class="k">as</span><span class="p">:</span>
<span class="n">output</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">oc_i</span><span class="p">][</span><span class="n">w_i</span><span class="p">]</span> <span class="o">=</span> <span class="n">bias</span><span class="p">[</span><span class="n">oc_i</span><span class="p">]</span>
<span class="o">+</span> <span class="n">sum_iC</span> <span class="n">sum_</span><span class="p">{</span><span class="n">ow</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">oW</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span> <span class="n">sum_</span><span class="p">{</span><span class="n">kw</span> <span class="o">=</span> <span class="mi">0</span> <span class="n">to</span> <span class="n">kW</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span>
<span class="n">weight</span><span class="p">[</span><span class="n">oc_i</span><span class="p">][</span><span class="n">ic_i</span><span class="p">][</span><span class="n">kw</span><span class="p">]</span> <span class="o">*</span> <span class="nb">input</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">ic_i</span><span class="p">][</span><span class="n">stride_w</span> <span class="o">*</span> <span class="n">ow</span> <span class="o">+</span> <span class="n">kw</span><span class="p">)]</span>
</code></pre>
<pre class="highlight python"><code><span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv1d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="n">stride</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">50</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<p>Note that depending of the size of your kernel, several (of the last)
columns of the input might be lost. It is up to the user
to add proper padding.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>in_channels</td>
<td></td>
<td>The number of expected input channels in the image given as input</td>
</tr>
<tr>
<td>out_channels</td>
<td></td>
<td>The number of output channels the convolution layer will produce</td>
</tr>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the convolving kernel.</td>
</tr>
<tr>
<td>stride</td>
<td></td>
<td>the stride of the convolving kernel.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , in_channels , * ]</td>
<td>Input is minibatch x in_channels x iW</td>
</tr>
<tr>
<td>output</td>
<td>[ * , out_channels , * ]</td>
<td>Output shape is precisely minibatch x out_channels x floor((iW + 2*padW - kW) / dW + 1)</td>
</tr>
</tbody></table>
<p>Members:
weight: the learnable weights of the module of shape (out_channels x in_channels x kW)
bias: the learnable bias of the module of shape (out_channels)</p>
<h2 id="conv2d">Conv2d</h2>
<p>Applies a 2D convolution over an input image composed of several input</p>
<pre class="highlight python"><code><span class="n">The</span> <span class="n">output</span> <span class="n">value</span> <span class="n">of</span> <span class="n">the</span> <span class="n">layer</span> <span class="k">with</span> <span class="nb">input</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">iC</span> <span class="n">x</span> <span class="n">H</span> <span class="n">x</span> <span class="n">W</span><span class="p">)</span> <span class="ow">and</span> <span class="n">output</span> <span class="p">(</span><span class="n">b</span> <span class="n">x</span> <span class="n">oC</span> <span class="n">x</span> <span class="n">oH</span> <span class="n">x</span> <span class="n">oW</span><span class="p">)</span>
<span class="n">can</span> <span class="n">be</span> <span class="n">precisely</span> <span class="n">described</span> <span class="k">as</span><span class="p">:</span>
<span class="n">output</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">oc_i</span><span class="p">][</span><span class="n">h_i</span><span class="p">][</span><span class="n">w_i</span><span class="p">]</span> <span class="o">=</span> <span class="n">bias</span><span class="p">[</span><span class="n">oc_i</span><span class="p">]</span>
<span class="o">+</span> <span class="n">sum_iC</span> <span class="n">sum_</span><span class="p">{</span><span class="n">oh</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">oH</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span> <span class="n">sum_</span><span class="p">{</span><span class="n">ow</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">oW</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span> <span class="n">sum_</span><span class="p">{</span><span class="n">kh</span> <span class="o">=</span> <span class="mi">0</span> <span class="n">to</span> <span class="n">kH</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span> <span class="n">sum_</span><span class="p">{</span><span class="n">kw</span> <span class="o">=</span> <span class="mi">0</span> <span class="n">to</span> <span class="n">kW</span><span class="o">-</span><span class="mi">1</span><span class="p">}</span>
<span class="n">weight</span><span class="p">[</span><span class="n">oc_i</span><span class="p">][</span><span class="n">ic_i</span><span class="p">][</span><span class="n">kh</span><span class="p">][</span><span class="n">kw</span><span class="p">]</span> <span class="o">*</span> <span class="nb">input</span><span class="p">[</span><span class="n">b_i</span><span class="p">][</span><span class="n">ic_i</span><span class="p">][</span><span class="n">stride_h</span> <span class="o">*</span> <span class="n">oh</span> <span class="o">+</span> <span class="n">kh</span><span class="p">)][</span><span class="n">stride_w</span> <span class="o">*</span> <span class="n">ow</span> <span class="o">+</span> <span class="n">kw</span><span class="p">)]</span>
</code></pre>
<pre class="highlight python"><code><span class="c"># With square kernels and equal stride</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv2d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="n">stride</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="c"># non-square kernels and unequal stride and with padding</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv2d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">5</span><span class="p">),</span> <span class="n">stride</span><span class="o">=</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">padding</span><span class="o">=</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="mi">2</span><span class="p">))</span>
<span class="c"># non-square kernels and unequal stride and with padding and dilation</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv2d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">5</span><span class="p">),</span> <span class="n">stride</span><span class="o">=</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">padding</span><span class="o">=</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="mi">2</span><span class="p">),</span> <span class="n">dilation</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">1</span><span class="p">))</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">50</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<p>Note that depending of the size of your kernel, several (of the last)
columns or rows of the input image might be lost. It is up to the user
to add proper padding in images.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>in_channels</td>
<td></td>
<td>The number of expected input channels in the image given as input</td>
</tr>
<tr>
<td>out_channels</td>
<td></td>
<td>The number of output channels the convolution layer will produce</td>
</tr>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the convolving kernel. Can be a single number k (for a square kernel of k x k) or a tuple (kh x kw)</td>
</tr>
<tr>
<td>stride</td>
<td>1</td>
<td>the stride of the convolving kernel. Can be a single number s or a tuple (sh x sw).</td>
</tr>
<tr>
<td>padding</td>
<td>0</td>
<td>implicit zero padding on the input. Can be a single number s or a tuple.</td>
</tr>
<tr>
<td>dilation</td>
<td>None</td>
<td>If given, will do dilated (or atrous) convolutions. Can be a single number s or a tuple.</td>
</tr>
<tr>
<td>no_bias</td>
<td>False</td>
<td>If set to true, the layer will not learn an additive bias.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , in_channels , * , * ]</td>
<td>Input is minibatch x in_channels x iH x iW</td>
</tr>
<tr>
<td>output</td>
<td>[ * , out_channels , * , * ]</td>
<td>Output shape is precisely minibatch x out_channels x floor((iH + 2*padH - kH) / dH + 1) x floor((iW + 2*padW - kW) / dW + 1)</td>
</tr>
</tbody></table>
<p>Members:
weight: the learnable weights of the module of shape (out_channels x in_channels x kH x kW)
bias: the learnable bias of the module of shape (out_channels)</p>
<h2 id="conv3d">Conv3d</h2>
<p>Applies a 3D convolution over an input image composed of several input</p>
<pre class="highlight python"><code><span class="c"># With square kernels and equal stride</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv3d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="n">stride</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="c"># non-square kernels and unequal stride and with padding</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Conv3d</span><span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="mi">33</span><span class="p">,</span> <span class="p">(</span><span class="mi">3</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="n">stride</span><span class="o">=</span><span class="p">(</span><span class="mi">2</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="n">padding</span><span class="o">=</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">))</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">50</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<p>Note that depending of the size of your kernel, several (of the last)
columns or rows of the input image might be lost. It is up to the user
to add proper padding in images.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>in_channels</td>
<td></td>
<td>The number of expected input channels in the image given as input</td>
</tr>
<tr>
<td>out_channels</td>
<td></td>
<td>The number of output channels the convolution layer will produce</td>
</tr>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the convolving kernel. Can be a single number k (for a square kernel of k x k x k) or a tuple (kt x kh x kw)</td>
</tr>
<tr>
<td>stride</td>
<td>1</td>
<td>the stride of the convolving kernel. Can be a single number s or a tuple (kt x sh x sw).</td>
</tr>
<tr>
<td>padding</td>
<td>0</td>
<td>implicit zero padding on the input. Can be a single number s or a tuple.</td>
</tr>
<tr>
<td>no_bias</td>
<td>False</td>
<td>If set to true, the layer will not learn an additive bias.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ * , in_channels , * , * , * ]</td>
<td>Input is minibatch x in_channels x iT x iH x iW</td>
</tr>
<tr>
<td>output</td>
<td>[ * , out_channels , * , * , * ]</td>
<td>Output shape is precisely minibatch x out_channels x floor((iT + 2*padT - kT) / dT + 1) x floor((iH + 2*padH - kH) / dH + 1) x floor((iW + 2*padW - kW) / dW + 1)</td>
</tr>
</tbody></table>
<p>Members:
weight: the learnable weights of the module of shape (out_channels x in_channels x kT x kH x kW)
bias: the learnable bias of the module of shape (out_channels)</p>
<h2 id="dropout">Dropout</h2>
<p>Randomly zeroes some of the elements of the input tensor.</p>
<pre class="highlight python"><code><span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Dropout</span><span class="p">(</span><span class="n">p</span><span class="o">=</span><span class="mf">0.2</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The elements to zero are randomized on every forward call.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>p</td>
<td>0.5</td>
<td>probability of an element to be zeroed.</td>
</tr>
<tr>
<td>inplace</td>
<td>false</td>
<td>If set to True, will do this operation in-place.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>Any</td>
<td>Input can be of any shape</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output is of the same shape as input</td>
</tr>
</tbody></table>
<h2 id="dropout2d">Dropout2d</h2>
<p>Randomly zeroes whole channels of the input tensor.</p>
<pre class="highlight python"><code><span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Dropout2d</span><span class="p">(</span><span class="n">p</span><span class="o">=</span><span class="mf">0.2</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">32</span><span class="p">,</span> <span class="mi">32</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The input is 4D (batch x channels, height, width) and each channel
is of size (1, height, width).
The channels to zero are randomized on every forward call.
Usually the input comes from Conv2d modules.</p>
<p>As described in the paper “Efficient Object Localization Using Convolutional
Networks” (http:arxiv.org/abs/1411.4280), if adjacent pixels within
feature maps are strongly correlated (as is normally the case in early
convolution layers) then iid dropout will not regularize the activations
and will otherwise just result in an effective learning rate decrease.
In this case, nn.Dropout2d will help promote independence between
feature maps and should be used instead.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>p</td>
<td>0.5</td>
<td>probability of an element to be zeroed.</td>
</tr>
<tr>
<td>inplace</td>
<td>false</td>
<td>If set to True, will do this operation in-place.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[*, *, *, *]</td>
<td>Input can be of any sizes of 4D shape</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output is of the same shape as input</td>
</tr>
</tbody></table>
<h2 id="dropout3d">Dropout3d</h2>
<p>Randomly zeroes whole channels of the input tensor.</p>
<pre class="highlight python"><code><span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Dropout3d</span><span class="p">(</span><span class="n">p</span><span class="o">=</span><span class="mf">0.2</span><span class="p">)</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">32</span><span class="p">,</span> <span class="mi">32</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>The input is 5D (batch x channels, depth, height, width) and each channel
is of size (1, depth, height, width).
The channels to zero are randomized on every forward call.
Usually the input comes from Conv3d modules.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>p</td>
<td>0.5</td>
<td>probability of an element to be zeroed.</td>
</tr>
<tr>
<td>inplace</td>
<td>false</td>
<td>If set to True, will do this operation in-place.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[*, *, *, *, *]</td>
<td>Input can be of any sizes of 5D shape</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output is of the same shape as input</td>
</tr>
</tbody></table>
<h2 id="elu">ELU</h2>
<p>Applies element-wise, ELU(x) = max(0,x) + min(0, alpha * (exp(x) - 1))</p>
<pre class="highlight python"><code><span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">ELU</span><span class="p">()</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">2</span><span class="p">))</span>
<span class="k">print</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
<span class="k">print</span><span class="p">(</span><span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">))</span>
</code></pre>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>alpha</td>
<td>1.0</td>
<td>the alpha value for the ELU formulation.</td>
</tr>
<tr>
<td>inplace</td>
<td></td>
<td>can optionally do the operation in-place</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>Any</td>
<td>Tensor of any size and dimension</td>
</tr>
<tr>
<td>output</td>
<td>Same</td>
<td>Output has the same shape as input</td>
</tr>
</tbody></table>
<h3 id="returns">Returns</h3>
<p>a Tensor of the same dimension and shape as the input</p>
<p><img src="image/elu.png" ></p>
<h2 id="embedding">Embedding</h2>
<p>A simple lookup table that stores embeddings of a fixed dictionary and size</p>
<pre class="highlight python"><code><span class="c"># an Embedding module containing 10 tensors of size 3</span>
<span class="n">embedding</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Embedding</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span>
<span class="c"># a batch of 2 samples of 4 indices each</span>
<span class="nb">input</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="mi">4</span><span class="p">,</span><span class="mi">5</span><span class="p">],[</span><span class="mi">4</span><span class="p">,</span><span class="mi">3</span><span class="p">,</span><span class="mi">2</span><span class="p">,</span><span class="mi">10</span><span class="p">]])</span>
<span class="k">print</span><span class="p">(</span><span class="n">embedding</span><span class="p">(</span><span class="nb">input</span><span class="p">))</span>
</code></pre>
<p>This module is often used to store word embeddings and retrieve them using indices.
The input to the module is a list of indices, and the output is the corresponding
word embeddings.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>num_embeddings</td>
<td></td>
<td>size of the dictionary of embeddings</td>
</tr>
<tr>
<td>embedding_dim</td>
<td></td>
<td>the size of each embedding vector</td>
</tr>
<tr>
<td>padding_idx</td>
<td>-1</td>
<td>If given, pads the output with zeros whenever it encounters the index.</td>
</tr>
<tr>
<td>max_norm</td>
<td>None</td>
<td>If given, will renormalize the embeddings to always have a norm lesser than this</td>
</tr>
<tr>
<td>norm_type</td>
<td></td>
<td>The p of the p-norm to compute for the max_norm option</td>
</tr>
<tr>
<td>scale_grad_by_freq</td>
<td></td>
<td>if given, this will scale gradients by the frequency of the words in the dictionary.</td>
</tr>
</tbody></table>
<h3 id="expected-shape">Expected Shape</h3>
<table><thead>
<tr>
<th></th>
<th>Shape</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>input</td>
<td>[ *, * ]</td>
<td>Input is a 2D mini_batch LongTensor of m x n indices to extract from the Embedding dictionary</td>
</tr>
<tr>
<td>output</td>
<td>[ * , *, * ]</td>
<td>Output shape = m x n x embedding_dim</td>
</tr>
</tbody></table>
<h2 id="fractionalmaxpool2d">FractionalMaxPool2d</h2>
<p>Applies a 2D fractional max pooling over an input signal composed of several input</p>
<pre class="highlight python"><code><span class="c"># pool of square window of size=3, and target output size 13x12</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">FractionalMaxPool2d</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="n">output_size</span><span class="o">=</span><span class="p">(</span><span class="mi">13</span><span class="p">,</span> <span class="mi">12</span><span class="p">))</span>
<span class="c"># pool of square window and target output size being half of input image size</span>
<span class="n">m</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">FractionalMaxPool2d</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="n">output_ratio</span><span class="o">=</span><span class="p">(</span><span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">))</span>
<span class="nb">input</span> <span class="o">=</span> <span class="n">autograd</span><span class="o">.</span><span class="n">Variable</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">16</span><span class="p">,</span> <span class="mi">50</span><span class="p">,</span> <span class="mi">32</span><span class="p">))</span>
<span class="n">output</span> <span class="o">=</span> <span class="n">m</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span>
</code></pre>
<p>planes. </p>
<p>Fractiona MaxPooling is described in detail in the paper <a href="http://arxiv.org/abs/1412.6071">“Fractional Max-Pooling” by Ben Graham</a>
The max-pooling operation is applied in kHxkW regions by a stochastic
step size determined by the target output size.
The number of output features is equal to the number of input planes.</p>
<h3 id="constructor-arguments">Constructor Arguments</h3>
<table><thead>
<tr>
<th>Parameter</th>
<th>Default</th>
<th>Description</th>
</tr>
</thead><tbody>
<tr>
<td>kernel_size</td>
<td></td>
<td>the size of the window to take a max over. Can be a single number k (for a square kernel of k x k) or a tuple (kh x kw)</td>
</tr>
<tr>
<td>output_size</td>
<td></td>
<td>the target output size of the image of the form oH x oW. Can be a tuple (oH, oW) or a single number oH for a square image oH x oH</td>
</tr>
<tr>
<td>output_ratio</td>
<td></td>
<td>If one wants to have an output size as a ratio of the input size, this option can be given. This has to be a number or tuple in the range (0, 1)</td>
</tr>
<tr>
<td>return_indices</td>
<td>False</td>
<td>if True, will return the indices along with the outputs. Useful to pass to nn.MaxUnpool2d .</td>
</tr>