Repository navigation
Expand file tree
/
Copy pathLanguage-Modeling-with-RNNs.html
More file actions
1576 lines (1216 loc) · 198 KB
/
Copy pathLanguage-Modeling-with-RNNs.html
File metadata and controls
1576 lines (1216 loc) · 198 KB
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
{% load static %}
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>Deep Learning: RNNs</title>
{# <script src="https://cdnjs.cloudflare.com/ajax/libs/require.js/2.1.10/require.min.js"></script>#}
{# <script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/2.0.3/jquery.min.js"></script>#}
<!-- Loading mathjax macro -->
<!-- Load mathjax -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS_HTML"></script>
<!-- MathJax configuration -->
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
tex2jax: {
inlineMath: [ ['$','$'], ["\\(","\\)"] ],
displayMath: [ ['$$','$$'], ["\\[","\\]"] ],
processEscapes: true,
processEnvironments: true
},
// Center justify equations in code and markdown cells. Elsewhere
// we use CSS to left justify single line equations in code cells.
displayAlign: 'center',
"HTML-CSS": {
styles: {'.MathJax_Display': {"margin": 0}},
linebreaks: { automatic: true }
}
});
</script>
<!-- Custom stylesheet, it must be in the same directory as the html file -->
{% include 'layout/components/css-head.html' %}
<link rel="stylesheet" href="{% static 'projects/css/ipynb.custom.css' %}">
<style>
a.dropdown-item {
color: #FAACA8;
font-size:1.5rem;
}
#notebook-container > p {
font-family: monospace, monospace;
}
</style>
</head>
<body>
{% include 'layout/components/navbar.html' %}
<div tabindex="-1" id="notebook" class="border-box-sizing bg-gradient">
<div class="container" id="notebook-container">
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="Language-Modeling-with-RNNs">Language Modeling with RNNs<a class="anchor-link" href="#Language-Modeling-with-RNNs">¶</a></h1><ul>
<li><b>Objective:</b> Understand Recurrent Neural Networks by implementing a vanilla RNN and an LSTM to train a model that can generate text</li>
<li>See <code>lib/layer_utils.py</code> for the definitions different layer type classes (RNN and LSTM)</li>
<li>See <code>lib/rnn.py</code> for the implementation of the text generation model.</li>
</ul>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [1]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="kn">from</span> <span class="nn">lib.rnn</span> <span class="k">import</span> <span class="o">*</span>
<span class="kn">from</span> <span class="nn">lib.layer_utils</span> <span class="k">import</span> <span class="o">*</span>
<span class="kn">from</span> <span class="nn">lib.grad_check</span> <span class="k">import</span> <span class="o">*</span>
<span class="kn">from</span> <span class="nn">lib.optim</span> <span class="k">import</span> <span class="o">*</span>
<span class="kn">from</span> <span class="nn">lib.train</span> <span class="k">import</span> <span class="o">*</span>
<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="kn">import</span> <span class="nn">matplotlib.pyplot</span> <span class="k">as</span> <span class="nn">plt</span>
<span class="o">%</span><span class="k">matplotlib</span> inline
<span class="n">plt</span><span class="o">.</span><span class="n">rcParams</span><span class="p">[</span><span class="s1">'figure.figsize'</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="mf">10.0</span><span class="p">,</span> <span class="mf">8.0</span><span class="p">)</span> <span class="c1"># set default size of plots</span>
<span class="n">plt</span><span class="o">.</span><span class="n">rcParams</span><span class="p">[</span><span class="s1">'image.interpolation'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'nearest'</span>
<span class="n">plt</span><span class="o">.</span><span class="n">rcParams</span><span class="p">[</span><span class="s1">'image.cmap'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'gray'</span>
<span class="c1"># for auto-reloading external modules</span>
<span class="c1"># see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython</span>
<span class="o">%</span><span class="k">load_ext</span> autoreload
<span class="o">%</span><span class="k">autoreload</span> 2
</pre></div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="Vanilla-RNN">Vanilla RNN<a class="anchor-link" href="#Vanilla-RNN">¶</a></h1><h2 id="Vanilla-RNN:-step-forward">Vanilla RNN: step forward<a class="anchor-link" href="#Vanilla-RNN:-step-forward">¶</a></h2><p>Testing the forward pass for a single timestep
of a basic RNN cell defined in <code>lib/layer_utils.py</code></p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [2]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">4</span>
<span class="n">rnn</span> <span class="o">=</span> <span class="n">VanillaRNN</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"rnn_test"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.4</span><span class="p">,</span> <span class="mf">0.7</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">D</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">prev_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.2</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.1</span><span class="p">,</span> <span class="mf">0.9</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">D</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.3</span><span class="p">,</span> <span class="mf">0.7</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">H</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.2</span><span class="p">,</span> <span class="mf">0.4</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">H</span><span class="p">)</span>
<span class="n">next_h</span><span class="p">,</span> <span class="n">_</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">)</span>
<span class="n">expected_next_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.58172089</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.50182032</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.41232771</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.31410098</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.66854692</span><span class="p">,</span> <span class="mf">0.79562378</span><span class="p">,</span> <span class="mf">0.87755553</span><span class="p">,</span> <span class="mf">0.92795967</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.97934501</span><span class="p">,</span> <span class="mf">0.99144213</span><span class="p">,</span> <span class="mf">0.99646691</span><span class="p">,</span> <span class="mf">0.99854353</span><span class="p">]])</span>
<span class="nb">print</span><span class="p">(</span><span class="n">next_h</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'next_h error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">expected_next_h</span><span class="p">,</span> <span class="n">next_h</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (10, 4)
self.params[self.wh_name]: (4, 4)
self.params[self.b_name]: (4,)
[[-0.58172089 -0.50182032 -0.41232771 -0.31410098]
[ 0.66854692 0.79562378 0.87755553 0.92795967]
[ 0.97934501 0.99144213 0.99646691 0.99854353]]
next_h error: 6.292421426471037e-09
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="Vanilla-RNN:-step-backward">Vanilla RNN: step backward<a class="anchor-link" href="#Vanilla-RNN:-step-backward">¶</a></h2><p>Checking the gradient calculations of the <code>VanillaRNN</code> class in the file <code>lib/layer_utils.py</code></p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [3]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">5</span><span class="p">,</span> <span class="mi">6</span>
<span class="n">rnn</span> <span class="o">=</span> <span class="n">VanillaRNN</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"rnn_test"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wx</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wh</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">out</span><span class="p">,</span> <span class="n">meta</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)</span>
<span class="n">dnext_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">dx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span> <span class="n">x</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span>
<span class="n">dprev_h_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">h</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span> <span class="n">h</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span>
<span class="n">dWx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">Wx</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span> <span class="n">Wx</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span>
<span class="n">dWh_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">Wh</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span> <span class="n">Wh</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span>
<span class="n">db_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">b</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span> <span class="n">b</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span>
<span class="n">dx</span><span class="p">,</span> <span class="n">dprev_h</span><span class="p">,</span> <span class="n">dWx</span><span class="p">,</span> <span class="n">dWh</span><span class="p">,</span> <span class="n">db</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">step_backward</span><span class="p">(</span><span class="n">dnext_h</span><span class="p">,</span> <span class="n">meta</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dx_num</span><span class="p">,</span> <span class="n">dx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dprev_h error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dprev_h_num</span><span class="p">,</span> <span class="n">dprev_h</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWx_num</span><span class="p">,</span> <span class="n">dWx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWh error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWh_num</span><span class="p">,</span> <span class="n">dWh</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'db error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">db_num</span><span class="p">,</span> <span class="n">db</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (5, 6)
self.params[self.wh_name]: (6, 6)
self.params[self.b_name]: (6,)
dx error: 9.956054783068812e-10
dprev_h error: 2.349821195946612e-10
dWx error: 1.9370107669274995e-10
dWh error: 3.186343591539138e-10
db error: 6.415453358481749e-11
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="Vanilla-RNN:-forward">Vanilla RNN: forward<a class="anchor-link" href="#Vanilla-RNN:-forward">¶</a></h2><p>Test <code>VanillaRNN</code> by processing a sequence of data.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [4]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">2</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="mi">5</span>
<span class="n">rnn</span> <span class="o">=</span> <span class="n">VanillaRNN</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"rnn_test"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.1</span><span class="p">,</span> <span class="mf">0.3</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">T</span><span class="o">*</span><span class="n">D</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">h0</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.3</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.2</span><span class="p">,</span> <span class="mf">0.4</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">D</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.4</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">H</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.7</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wx</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wh</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">h</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span>
<span class="n">expected_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([</span>
<span class="p">[</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.42070749</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.27279261</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.11074945</span><span class="p">,</span> <span class="mf">0.05740409</span><span class="p">,</span> <span class="mf">0.22236251</span><span class="p">],</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.39525808</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.22554661</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.0409454</span><span class="p">,</span> <span class="mf">0.14649412</span><span class="p">,</span> <span class="mf">0.32397316</span><span class="p">],</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.42305111</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.24223728</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.04287027</span><span class="p">,</span> <span class="mf">0.15997045</span><span class="p">,</span> <span class="mf">0.35014525</span><span class="p">],</span>
<span class="p">],</span>
<span class="p">[</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.55857474</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.39065825</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.19198182</span><span class="p">,</span> <span class="mf">0.02378408</span><span class="p">,</span> <span class="mf">0.23735671</span><span class="p">],</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.27150199</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.07088804</span><span class="p">,</span> <span class="mf">0.13562939</span><span class="p">,</span> <span class="mf">0.33099728</span><span class="p">,</span> <span class="mf">0.50158768</span><span class="p">],</span>
<span class="p">[</span><span class="o">-</span><span class="mf">0.51014825</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.30524429</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.06755202</span><span class="p">,</span> <span class="mf">0.17806392</span><span class="p">,</span> <span class="mf">0.40333043</span><span class="p">]]])</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"expected_h: "</span><span class="p">,</span><span class="n">expected_h</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'h error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">expected_h</span><span class="p">,</span> <span class="n">h</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (4, 5)
self.params[self.wh_name]: (5, 5)
self.params[self.b_name]: (5,)
expected_h: (2, 3, 5)
h error: 7.728466180186066e-08
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="Vanilla-RNN:-backward">Vanilla RNN: backward<a class="anchor-link" href="#Vanilla-RNN:-backward">¶</a></h2><p>Test back-propagation over the entire sequence by calling the <code>step_backward</code> function</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [5]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">5</span>
<span class="n">rnn</span> <span class="o">=</span> <span class="n">VanillaRNN</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"rnn_test"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">h0</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">H</span><span class="p">)</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wx</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wh</span>
<span class="n">rnn</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">out</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span>
<span class="n">dout</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">dx</span><span class="p">,</span> <span class="n">dh0</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">dout</span><span class="p">)</span>
<span class="n">dx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">x</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dh0_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">h0</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">h0</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dWx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">Wx</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">Wx</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dWh_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">Wh</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">Wh</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">db_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">b</span><span class="p">:</span> <span class="n">rnn</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">b</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dWx</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span>
<span class="n">dWh</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span>
<span class="n">db</span> <span class="o">=</span> <span class="n">rnn</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">rnn</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dx_num</span><span class="p">,</span> <span class="n">dx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dh0 error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dh0_num</span><span class="p">,</span> <span class="n">dh0</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWx_num</span><span class="p">,</span> <span class="n">dWx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWh error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWh_num</span><span class="p">,</span> <span class="n">dWh</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'db error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">db_num</span><span class="p">,</span> <span class="n">db</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="n">dh0_num</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">dh0</span><span class="p">)</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (3, 5)
self.params[self.wh_name]: (5, 5)
self.params[self.b_name]: (5,)
dx error: 2.736928435887175e-08
dh0 error: 8.231409890331713e-10
dWx error: 2.0789178600982087e-08
dWh error: 1.5210912998171804e-08
db error: 2.77914406963907e-10
[[-4.28153694 -2.74230889 0.71964976 -1.18508456 -0.8895025 ]
[ 0.5942948 -0.86422636 -1.14307499 -0.07620721 -1.10608234]]
[[-4.28153694 -2.74230889 0.71964976 -1.18508456 -0.8895025 ]
[ 0.5942948 -0.86422636 -1.14307499 -0.07620721 -1.10608234]]
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="Word-embedding">Word embedding<a class="anchor-link" href="#Word-embedding">¶</a></h1><h2 id="Word-embedding:-forward">Word embedding: forward<a class="anchor-link" href="#Word-embedding:-forward">¶</a></h2><p>Checking implementation of the function <code>forward</code> in the <code>word_embedding</code> class</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [6]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span><span class="p">,</span> <span class="n">D</span> <span class="o">=</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">3</span>
<span class="n">we</span> <span class="o">=</span> <span class="n">word_embedding</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"we"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([[</span><span class="mi">0</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="mi">2</span><span class="p">],</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">0</span><span class="p">,</span> <span class="mi">3</span><span class="p">]])</span>
<span class="n">W</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">V</span><span class="o">*</span><span class="n">D</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">we</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">we</span><span class="o">.</span><span class="n">w_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">W</span>
<span class="n">out</span> <span class="o">=</span> <span class="n">we</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">out</span><span class="p">)</span>
<span class="n">expected_out</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([</span>
<span class="p">[[</span> <span class="mf">0.</span><span class="p">,</span> <span class="mf">0.07142857</span><span class="p">,</span> <span class="mf">0.14285714</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.64285714</span><span class="p">,</span> <span class="mf">0.71428571</span><span class="p">,</span> <span class="mf">0.78571429</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.21428571</span><span class="p">,</span> <span class="mf">0.28571429</span><span class="p">,</span> <span class="mf">0.35714286</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.42857143</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.57142857</span><span class="p">]],</span>
<span class="p">[[</span> <span class="mf">0.42857143</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.57142857</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.21428571</span><span class="p">,</span> <span class="mf">0.28571429</span><span class="p">,</span> <span class="mf">0.35714286</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.</span><span class="p">,</span> <span class="mf">0.07142857</span><span class="p">,</span> <span class="mf">0.14285714</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.64285714</span><span class="p">,</span> <span class="mf">0.71428571</span><span class="p">,</span> <span class="mf">0.78571429</span><span class="p">]]])</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'out error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">expected_out</span><span class="p">,</span> <span class="n">out</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>[[[0. 0.07142857 0.14285714]
[0.64285714 0.71428571 0.78571429]
[0.21428571 0.28571429 0.35714286]
[0.42857143 0.5 0.57142857]]
[[0.42857143 0.5 0.57142857]
[0.21428571 0.28571429 0.35714286]
[0. 0.07142857 0.14285714]
[0.64285714 0.71428571 0.78571429]]]
out error: 1.0000000094736443e-08
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="Word-embedding:-backward">Word embedding: backward<a class="anchor-link" href="#Word-embedding:-backward">¶</a></h2><p>Checking implementation of the function <code>backward</code> in the <code>word_embedding</code> class</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [7]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span><span class="p">,</span> <span class="n">D</span> <span class="o">=</span> <span class="mi">50</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">6</span>
<span class="n">we</span> <span class="o">=</span> <span class="n">word_embedding</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s2">"we"</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">))</span>
<span class="n">W</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">we</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">we</span><span class="o">.</span><span class="n">w_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">W</span>
<span class="n">out</span> <span class="o">=</span> <span class="n">we</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">dout</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">we</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">dout</span><span class="p">)</span>
<span class="n">dW</span> <span class="o">=</span> <span class="n">we</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">we</span><span class="o">.</span><span class="n">w_name</span><span class="p">]</span>
<span class="n">f</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">W</span><span class="p">:</span> <span class="n">we</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">dW_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="n">f</span><span class="p">,</span> <span class="n">W</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dW error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dW</span><span class="p">,</span> <span class="n">dW_num</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>dW error: 3.2759440934795915e-12
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="Temporal-Fully-Connected-layer">Temporal Fully Connected layer<a class="anchor-link" href="#Temporal-Fully-Connected-layer">¶</a></h1><p>Checking implementation of the <code>temporal_fc</code> class, which defines a layer that uses an affine function to transform the RNN hidden vector at that timestep into scores for each word in the vocabulary</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [8]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="c1"># Gradient check for temporal affine layer</span>
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">M</span> <span class="o">=</span> <span class="mi">2</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="mi">5</span>
<span class="n">t_fc</span> <span class="o">=</span> <span class="n">temporal_fc</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">M</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">'test_t_fc'</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">w</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">M</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">M</span><span class="p">)</span>
<span class="n">t_fc</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">t_fc</span><span class="o">.</span><span class="n">w_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">w</span>
<span class="n">t_fc</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">t_fc</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">out</span> <span class="o">=</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="n">dout</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">dx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">),</span> <span class="n">x</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dw_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">w</span><span class="p">:</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">),</span> <span class="n">w</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">db_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span><span class="p">(</span><span class="k">lambda</span> <span class="n">b</span><span class="p">:</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">),</span> <span class="n">b</span><span class="p">,</span> <span class="n">dout</span><span class="p">)</span>
<span class="n">dx</span> <span class="o">=</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">dout</span><span class="p">)</span>
<span class="n">dw</span> <span class="o">=</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">t_fc</span><span class="o">.</span><span class="n">w_name</span><span class="p">]</span>
<span class="n">db</span> <span class="o">=</span> <span class="n">t_fc</span><span class="o">.</span><span class="n">grads</span><span class="p">[</span><span class="n">t_fc</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dx_num</span><span class="p">,</span> <span class="n">dx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dw error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dw_num</span><span class="p">,</span> <span class="n">dw</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'db error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">db_num</span><span class="p">,</span> <span class="n">db</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>dx error: 3.2269470390098687e-10
dw error: 3.8595619942595054e-11
db error: 1.1455396263586309e-11
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="Temporal-Softmax-Cross-Entropy-loss">Temporal Softmax Cross-Entropy loss<a class="anchor-link" href="#Temporal-Softmax-Cross-Entropy-loss">¶</a></h1><p>When rolling out a RNN language model to generate a sentence, a score for each word in the vocabulary is produced at every timestep. This score is propotional to the predicted likelihood of this word appearing at the particular timestep in the sentence. Because the ground-truth word at each timestep is known, softmax cross-entropy loss function is used to:</p>
<ul>
<li>Compute a proper probability distribution over the words in the vocabulary at every time step</li>
<li>Compute loss and gradient at each timestep. We sum the losses over time and average them over the minibatch.</li>
</ul>
<p>(Loss function: <code>temporal_softmax_CE_loss</code> in<code>lib/layer_utils.py</code>)</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [9]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">loss_func</span> <span class="o">=</span> <span class="n">temporal_softmax_CE_loss</span><span class="p">()</span>
<span class="c1"># Sanity check for temporal softmax loss</span>
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span> <span class="o">=</span> <span class="mi">100</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">10</span>
<span class="k">def</span> <span class="nf">check_loss</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span><span class="p">,</span> <span class="n">p</span><span class="p">):</span>
<span class="n">x</span> <span class="o">=</span> <span class="mf">0.001</span> <span class="o">*</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span><span class="p">)</span>
<span class="n">y</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">))</span>
<span class="n">mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">)</span> <span class="o"><=</span> <span class="n">p</span>
<span class="nb">print</span><span class="p">(</span><span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">mask</span><span class="p">))</span>
<span class="n">check_loss</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">)</span> <span class="c1"># Should be about 2.3</span>
<span class="n">check_loss</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">)</span> <span class="c1"># Should be about 23</span>
<span class="n">check_loss</span><span class="p">(</span><span class="mi">5000</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mf">0.1</span><span class="p">)</span> <span class="c1"># Should be about 2.3</span>
<span class="c1"># Gradient check for temporal softmax loss</span>
<span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span> <span class="o">=</span> <span class="mi">7</span><span class="p">,</span> <span class="mi">8</span><span class="p">,</span> <span class="mi">9</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">V</span><span class="p">)</span>
<span class="n">y</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="n">V</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">))</span>
<span class="n">mask</span> <span class="o">=</span> <span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">)</span> <span class="o">></span> <span class="mf">0.5</span><span class="p">)</span>
<span class="n">loss</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">mask</span><span class="p">)</span>
<span class="n">dx</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
<span class="n">dx_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient</span><span class="p">(</span><span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">mask</span><span class="p">),</span> <span class="n">x</span><span class="p">,</span> <span class="n">verbose</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dx</span><span class="p">,</span> <span class="n">dx_num</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>2.3026547279318357
23.026307039328714
2.2989009292538665
dx error: 4.0464746298031226e-08
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="RNN-for-language-modeling">RNN for language modeling<a class="anchor-link" href="#RNN-for-language-modeling">¶</a></h1><p>Check the forward and backward pass of the <code>TestRNN</code> class using a small test case</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [10]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">20</span><span class="p">,</span> <span class="mi">40</span>
<span class="n">V</span> <span class="o">=</span> <span class="mi">4</span>
<span class="n">T</span> <span class="o">=</span> <span class="mi">13</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">TestRNN</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">cell_type</span><span class="o">=</span><span class="s1">'rnn'</span><span class="p">)</span>
<span class="n">loss_func</span> <span class="o">=</span> <span class="n">temporal_softmax_CE_loss</span><span class="p">()</span>
<span class="c1"># Set all model parameters to fixed values</span>
<span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">model</span><span class="o">.</span><span class="n">params</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="n">model</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">k</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">1.4</span><span class="p">,</span> <span class="mf">1.3</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">v</span><span class="o">.</span><span class="n">size</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="o">*</span><span class="n">v</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">model</span><span class="o">.</span><span class="n">assign_params</span><span class="p">()</span>
<span class="n">features</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">1.5</span><span class="p">,</span> <span class="mf">0.3</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="p">(</span><span class="n">N</span> <span class="o">*</span> <span class="n">D</span> <span class="o">*</span> <span class="n">T</span><span class="p">))</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">h0</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">1.5</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="p">(</span><span class="n">N</span><span class="o">*</span><span class="n">H</span><span class="p">))</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">labels</span> <span class="o">=</span> <span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">N</span> <span class="o">*</span> <span class="n">T</span><span class="p">)</span> <span class="o">%</span> <span class="n">V</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">)</span>
<span class="n">pred</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">features</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span>
<span class="c1"># You'll need this</span>
<span class="n">mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">N</span><span class="p">,</span> <span class="n">T</span><span class="p">))</span>
<span class="n">loss</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">pred</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mask</span><span class="p">)</span>
<span class="n">dLoss</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
<span class="n">expected_loss</span> <span class="o">=</span> <span class="mf">51.0949189134</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'loss: '</span><span class="p">,</span> <span class="n">loss</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'expected loss: '</span><span class="p">,</span> <span class="n">expected_loss</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'difference: '</span><span class="p">,</span> <span class="nb">abs</span><span class="p">(</span><span class="n">loss</span> <span class="o">-</span> <span class="n">expected_loss</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (20, 40)
self.params[self.wh_name]: (40, 40)
self.params[self.b_name]: (40,)
loss: 51.094918913361184
expected loss: 51.0949189134
difference: 3.881694965457427e-11
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>Detailed gradient checking on the backward pass of the <code>TestRNN</code> class</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [11]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="n">batch_size</span> <span class="o">=</span> <span class="mi">2</span>
<span class="n">timesteps</span> <span class="o">=</span> <span class="mi">3</span>
<span class="n">input_dim</span> <span class="o">=</span> <span class="mi">4</span>
<span class="n">hidden_dim</span> <span class="o">=</span> <span class="mi">6</span>
<span class="n">label_size</span> <span class="o">=</span> <span class="mi">4</span>
<span class="n">labels</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="n">label_size</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">timesteps</span><span class="p">))</span>
<span class="n">features</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">timesteps</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">)</span>
<span class="n">h0</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">hidden_dim</span><span class="p">)</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">TestRNN</span><span class="p">(</span><span class="n">input_dim</span><span class="p">,</span> <span class="n">hidden_dim</span><span class="p">,</span> <span class="n">cell_type</span><span class="o">=</span><span class="s1">'rnn'</span><span class="p">)</span>
<span class="n">loss_func</span> <span class="o">=</span> <span class="n">temporal_softmax_CE_loss</span><span class="p">()</span>
<span class="n">pred</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">features</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span>
<span class="c1"># You'll need this</span>
<span class="n">mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">timesteps</span><span class="p">))</span>
<span class="n">loss</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">pred</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mask</span><span class="p">)</span>
<span class="n">dLoss</span> <span class="o">=</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
<span class="n">dout</span><span class="p">,</span> <span class="n">dh0</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">dLoss</span><span class="p">)</span>
<span class="n">grads</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">grads</span>
<span class="k">for</span> <span class="n">param_name</span> <span class="ow">in</span> <span class="nb">sorted</span><span class="p">(</span><span class="n">grads</span><span class="p">):</span>
<span class="n">f</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">_</span><span class="p">:</span> <span class="n">loss_func</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">features</span><span class="p">,</span> <span class="n">h0</span><span class="p">),</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mask</span><span class="p">)</span>
<span class="n">param_grad_num</span> <span class="o">=</span> <span class="n">eval_numerical_gradient</span><span class="p">(</span><span class="n">f</span><span class="p">,</span> <span class="n">model</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">param_name</span><span class="p">],</span> <span class="n">verbose</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">h</span><span class="o">=</span><span class="mf">1e-6</span><span class="p">)</span>
<span class="n">e</span> <span class="o">=</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">param_grad_num</span><span class="p">,</span> <span class="n">grads</span><span class="p">[</span><span class="n">param_name</span><span class="p">])</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'</span><span class="si">%s</span><span class="s1"> relative error: </span><span class="si">%e</span><span class="s1">'</span> <span class="o">%</span> <span class="p">(</span><span class="n">param_name</span><span class="p">,</span> <span class="n">e</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>self.params[self.wx_name] (4, 6)
self.params[self.wh_name]: (6, 6)
self.params[self.b_name]: (6,)
vanilla_rnn_b relative error: 9.451394e-08
vanilla_rnn_wh relative error: 3.221744e-08
vanilla_rnn_wx relative error: 9.508480e-08
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h1 id="LSTM:-Theory">LSTM: Theory<a class="anchor-link" href="#LSTM:-Theory">¶</a></h1><p>Vanilla RNNs can be tough to train on long sequences due to vanishing and exploding gradiants. LSTMs solve this problem by replacing the simple update rule in the forward step of the vanilla RNN with a gating mechanism as follows.</p>
<p>Similar to the vanilla RNN, at each timestep we receive an input $x_t\in\mathbb{R}^D$ and the previous hidden state $h_{t-1}\in\mathbb{R}^H$. Crucially, the LSTM also maintains an $H$-dimensional <em>cell state</em>, so we also receive the previous cell state $c_{t-1}\in\mathbb{R}^H$. The learnable parameters of the LSTM are an <em>input-to-hidden</em> matrix $W_x\in\mathbb{R}^{4H\times D}$, a <em>hidden-to-hidden</em> matrix $W_h\in\mathbb{R}^{4H\times H}$ and a <em>bias vector</em> $b\in\mathbb{R}^{4H}$.</p>
<p>At each timestep we first compute an <em>activation vector</em> $a\in\mathbb{R}^{4H}$ as $a=W_xx_t + W_hh_{t-1}+b$. We then divide this into four vectors $a_i,a_f,a_o,a_g\in\mathbb{R}^H$ where $a_i$ consists of the first $H$ elements of $a$, $a_f$ is the next $H$ elements of $a$, etc. We then compute the <em>input gate</em> $g\in\mathbb{R}^H$, <em>forget gate</em> $f\in\mathbb{R}^H$, <em>output gate</em> $o\in\mathbb{R}^H$ and <em>gate gate</em> $g\in\mathbb{R}^H$ as</p>
$$
\begin{align*}
i = \sigma(a_i) \hspace{2pc}
f = \sigma(a_f) \hspace{2pc}
o = \sigma(a_o) \hspace{2pc}
g = \tanh(a_g)
\end{align*}
$$<p>where $\sigma$ is the sigmoid function and $\tanh$ is the hyperbolic tangent, both applied elementwise.</p>
<p>Finally we compute the next cell state $c_t$ and next hidden state $h_t$ as</p>
$$
c_{t} = f\odot c_{t-1} + i\odot g \hspace{4pc}
h_t = o\odot\tanh(c_t)
$$<p>where $\odot$ is the elementwise product of vectors.</p>
<p>In the rest of the notebook we will implement the LSTM update rule and apply it to the text generation task.</p>
<p>In the code, we assume that data is stored in batches so that $X_t \in \mathbb{R}^{N\times D}$, and will work with <em>transposed</em> versions of the parameters: $W_x \in \mathbb{R}^{D \times 4H}$, $W_h \in \mathbb{R}^{H\times 4H}$ so that activations $A \in \mathbb{R}^{N\times 4H}$ can be computed efficiently as $A = X_t W_x + H_{t-1} W_h$</p>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="LSTM:-step-forward">LSTM: step forward<a class="anchor-link" href="#LSTM:-step-forward">¶</a></h2><p>Testing <code>lstm.step_forward</code> for a single timestep</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [12]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">5</span>
<span class="n">lstm</span> <span class="o">=</span> <span class="n">LSTM</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">'test_lstm'</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.4</span><span class="p">,</span> <span class="mf">1.2</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">D</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">prev_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.3</span><span class="p">,</span> <span class="mf">0.7</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">prev_c</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.4</span><span class="p">,</span> <span class="mf">0.9</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="n">N</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">2.1</span><span class="p">,</span> <span class="mf">1.3</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="mi">4</span><span class="o">*</span><span class="n">D</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="mi">4</span> <span class="o">*</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="o">-</span><span class="mf">0.7</span><span class="p">,</span> <span class="mf">2.2</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="mi">4</span><span class="o">*</span><span class="n">H</span><span class="o">*</span><span class="n">H</span><span class="p">)</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="mi">4</span> <span class="o">*</span> <span class="n">H</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mf">0.3</span><span class="p">,</span> <span class="mf">0.7</span><span class="p">,</span> <span class="n">num</span><span class="o">=</span><span class="mi">4</span><span class="o">*</span><span class="n">H</span><span class="p">)</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wx</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wh</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">next_h</span><span class="p">,</span> <span class="n">next_c</span><span class="p">,</span> <span class="n">cache</span> <span class="o">=</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)</span>
<span class="n">expected_next_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([</span>
<span class="p">[</span> <span class="mf">0.24635157</span><span class="p">,</span> <span class="mf">0.28610883</span><span class="p">,</span> <span class="mf">0.32240467</span><span class="p">,</span> <span class="mf">0.35525807</span><span class="p">,</span> <span class="mf">0.38474904</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.49223563</span><span class="p">,</span> <span class="mf">0.55611431</span><span class="p">,</span> <span class="mf">0.61507696</span><span class="p">,</span> <span class="mf">0.66844003</span><span class="p">,</span> <span class="mf">0.7159181</span> <span class="p">],</span>
<span class="p">[</span> <span class="mf">0.56735664</span><span class="p">,</span> <span class="mf">0.66310127</span><span class="p">,</span> <span class="mf">0.74419266</span><span class="p">,</span> <span class="mf">0.80889665</span><span class="p">,</span> <span class="mf">0.858299</span> <span class="p">]])</span>
<span class="n">expected_next_c</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">([</span>
<span class="p">[</span> <span class="mf">0.32986176</span><span class="p">,</span> <span class="mf">0.39145139</span><span class="p">,</span> <span class="mf">0.451556</span><span class="p">,</span> <span class="mf">0.51014116</span><span class="p">,</span> <span class="mf">0.56717407</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.66382255</span><span class="p">,</span> <span class="mf">0.76674007</span><span class="p">,</span> <span class="mf">0.87195994</span><span class="p">,</span> <span class="mf">0.97902709</span><span class="p">,</span> <span class="mf">1.08751345</span><span class="p">],</span>
<span class="p">[</span> <span class="mf">0.74192008</span><span class="p">,</span> <span class="mf">0.90592151</span><span class="p">,</span> <span class="mf">1.07717006</span><span class="p">,</span> <span class="mf">1.25120233</span><span class="p">,</span> <span class="mf">1.42395676</span><span class="p">]])</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'next_h error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">expected_next_h</span><span class="p">,</span> <span class="n">next_h</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'next_c error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">expected_next_c</span><span class="p">,</span> <span class="n">next_c</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="prompt"></div>
<div class="output_subarea output_stream output_stdout output_text">
<pre>next_h error: 5.7054131185818695e-09
next_c error: 5.8143123088804145e-09
</pre>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt">
</div><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="LSTM:-step-backward">LSTM: step backward<a class="anchor-link" href="#LSTM:-step-backward">¶</a></h2><p>Testing <code>lstm.step_backward</code> for a single timestep</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="prompt input_prompt">In [13]:</div>
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="o">%</span><span class="k">reload_ext</span> autoreload
<span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">231</span><span class="p">)</span>
<span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">H</span> <span class="o">=</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">5</span><span class="p">,</span> <span class="mi">6</span>
<span class="n">lstm</span> <span class="o">=</span> <span class="n">LSTM</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="n">H</span><span class="p">,</span> <span class="n">init_scale</span><span class="o">=</span><span class="mf">0.02</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">'test_lstm'</span><span class="p">)</span>
<span class="n">x</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">D</span><span class="p">)</span>
<span class="n">prev_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">prev_c</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">D</span><span class="p">,</span> <span class="mi">4</span> <span class="o">*</span> <span class="n">H</span><span class="p">)</span>
<span class="n">Wh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="n">H</span><span class="p">,</span> <span class="mi">4</span> <span class="o">*</span> <span class="n">H</span><span class="p">)</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="mi">4</span> <span class="o">*</span> <span class="n">H</span><span class="p">)</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">wx_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wx</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">wh_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">Wh</span>
<span class="n">lstm</span><span class="o">.</span><span class="n">params</span><span class="p">[</span><span class="n">lstm</span><span class="o">.</span><span class="n">b_name</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
<span class="n">next_h</span><span class="p">,</span> <span class="n">next_c</span><span class="p">,</span> <span class="n">cache</span> <span class="o">=</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)</span>
<span class="n">dnext_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">next_h</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">dnext_c</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="o">*</span><span class="n">next_c</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
<span class="n">fx_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fh_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">h</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fc_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">c</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fWx_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">Wx</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fWh_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">Wh</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fb_h</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">b</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">fx_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">fh_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">h</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">fc_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">c</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">fWx_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">Wx</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">fWh_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">Wh</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">fb_c</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">b</span><span class="p">:</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">)[</span><span class="mi">1</span><span class="p">]</span>
<span class="n">num_grad</span> <span class="o">=</span> <span class="n">eval_numerical_gradient_array</span>
<span class="n">dx_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fx_h</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fx_c</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">dh_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fh_h</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fh_c</span><span class="p">,</span> <span class="n">prev_h</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">dc_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fc_h</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fc_c</span><span class="p">,</span> <span class="n">prev_c</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">dWx_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fWx_h</span><span class="p">,</span> <span class="n">Wx</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fWx_c</span><span class="p">,</span> <span class="n">Wx</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">dWh_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fWh_h</span><span class="p">,</span> <span class="n">Wh</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fWh_c</span><span class="p">,</span> <span class="n">Wh</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">db_num</span> <span class="o">=</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fb_h</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">dnext_h</span><span class="p">)</span> <span class="o">+</span> <span class="n">num_grad</span><span class="p">(</span><span class="n">fb_c</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">)</span>
<span class="n">dx</span><span class="p">,</span> <span class="n">dh</span><span class="p">,</span> <span class="n">dc</span><span class="p">,</span> <span class="n">dWx</span><span class="p">,</span> <span class="n">dWh</span><span class="p">,</span> <span class="n">db</span> <span class="o">=</span> <span class="n">lstm</span><span class="o">.</span><span class="n">step_backward</span><span class="p">(</span><span class="n">dnext_h</span><span class="p">,</span> <span class="n">dnext_c</span><span class="p">,</span> <span class="n">cache</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dx_num</span><span class="p">,</span> <span class="n">dx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dh error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dh_num</span><span class="p">,</span> <span class="n">dh</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dc error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dc_num</span><span class="p">,</span> <span class="n">dc</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWx error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWx_num</span><span class="p">,</span> <span class="n">dWx</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'dWh error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">dWh_num</span><span class="p">,</span> <span class="n">dWh</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s1">'db error: '</span><span class="p">,</span> <span class="n">rel_error</span><span class="p">(</span><span class="n">db_num</span><span class="p">,</span> <span class="n">db</span><span class="p">))</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">