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
|
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="canonical" href="http://keras.io/initializers/">
<link rel="shortcut icon" href="../img/favicon.ico">
<title>Initializers - Keras Documentation</title>
<link href='https://fonts.googleapis.com/css?family=Lato:400,700|Source+Sans+Pro:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>
<link rel="stylesheet" href="../css/theme.css" type="text/css" />
<link rel="stylesheet" href="../css/theme_extra.css" type="text/css" />
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
<script>
// Current page data
var mkdocs_page_name = "Initializers";
var mkdocs_page_input_path = "initializers.md";
var mkdocs_page_url = "/initializers/";
</script>
<script src="../js/jquery-2.1.1.min.js" defer></script>
<script src="../js/modernizr-2.8.3.min.js" defer></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
<script>
(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
})(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
ga('create', 'UA-61785484-1', 'keras.io');
ga('send', 'pageview');
</script>
</head>
<body class="wy-body-for-nav" role="document">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
<div class="wy-side-scroll">
<a href="">
<div class="keras-logo">
<img src="/img/keras-logo-small.jpg" class="keras-logo-img">
Keras Documentation
</div>
</a>
<div class="wy-side-nav-search">
<div role="search">
<form id ="rtd-search-form" class="wy-form" action="../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" title="Type search term here" />
</form>
</div>
</div>
<div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<ul>
<li class="toctree-l1"><a class="reference internal" href="..">Home</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../why-use-keras/">Why use Keras</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Getting started</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../getting-started/sequential-model-guide/">Guide to the Sequential model</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../getting-started/functional-api-guide/">Guide to the Functional API</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../getting-started/faq/">FAQ</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Models</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../models/about-keras-models/">About Keras models</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../models/sequential/">Sequential</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../models/model/">Model (functional API)</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Layers</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../layers/about-keras-layers/">About Keras layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/core/">Core Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/convolutional/">Convolutional Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/pooling/">Pooling Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/local/">Locally-connected Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/recurrent/">Recurrent Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/embeddings/">Embedding Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/merge/">Merge Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/advanced-activations/">Advanced Activations Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/normalization/">Normalization Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/noise/">Noise layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/wrappers/">Layer wrappers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../layers/writing-your-own-keras-layers/">Writing your own Keras layers</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Preprocessing</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../preprocessing/sequence/">Sequence Preprocessing</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../preprocessing/text/">Text Preprocessing</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../preprocessing/image/">Image Preprocessing</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../losses/">Losses</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../metrics/">Metrics</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../optimizers/">Optimizers</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../activations/">Activations</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../callbacks/">Callbacks</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../datasets/">Datasets</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../applications/">Applications</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../backend/">Backend</a>
</li>
</ul>
<ul class="current">
<li class="toctree-l1 current"><a class="reference internal current" href="./">Initializers</a>
<ul class="current">
</ul>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../regularizers/">Regularizers</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../constraints/">Constraints</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../visualization/">Visualization</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../scikit-learn-api/">Scikit-learn API</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../utils/">Utils</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../contributing/">Contributing</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Examples</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../examples/addition_rnn/">Addition RNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/antirectifier/">Custom layer - antirectifier</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/babi_rnn/">Baby RNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/babi_memnn/">Baby MemNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/cifar10_cnn/">CIFAR-10 CNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/cifar10_resnet/">CIFAR-10 ResNet</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/conv_filter_visualization/">Convolution filter visualization</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/conv_lstm/">Convolutional LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/deep_dream/">Deep Dream</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/image_ocr/">Image OCR</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/imdb_bidirectional_lstm/">Bidirectional LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/imdb_cnn/">1D CNN for text classification</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/imdb_cnn_lstm/">Sentiment classification CNN-LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/imdb_fasttext/">Fasttext for text classification</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/imdb_lstm/">Sentiment classification LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/lstm_seq2seq/">Sequence to sequence - training</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/lstm_seq2seq_restore/">Sequence to sequence - prediction</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/lstm_stateful/">Stateful LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/lstm_text_generation/">LSTM for text generation</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../examples/mnist_acgan/">Auxiliary Classifier GAN</a>
</li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
<nav class="wy-nav-top" role="navigation" aria-label="top navigation">
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="..">Keras Documentation</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="wy-breadcrumbs">
<li><a href="..">Docs</a> »</li>
<li>Initializers</li>
<li class="wy-breadcrumbs-aside">
<a href="https://github.com/keras-team/keras/tree/master/docs"
class="icon icon-github"> Edit on GitHub</a>
</li>
</ul>
<hr/>
</div>
<div role="main">
<div class="section">
<h2 id="usage-of-initializers">Usage of initializers</h2>
<p>Initializations define the way to set the initial random weights of Keras layers.</p>
<p>The keyword arguments used for passing initializers to layers will depend on the layer. Usually it is simply <code>kernel_initializer</code> and <code>bias_initializer</code>:</p>
<pre><code class="python">model.add(Dense(64,
kernel_initializer='random_uniform',
bias_initializer='zeros'))
</code></pre>
<h2 id="available-initializers">Available initializers</h2>
<p>The following built-in initializers are available as part of the <code>keras.initializers</code> module:</p>
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L14">[source]</a></span></p>
<h3 id="initializer">Initializer</h3>
<pre><code class="python">keras.initializers.Initializer()
</code></pre>
<p>Initializer base class: all initializers inherit from this class.</p>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L33">[source]</a></span></p>
<h3 id="zeros">Zeros</h3>
<pre><code class="python">keras.initializers.Zeros()
</code></pre>
<p>Initializer that generates tensors initialized to 0.</p>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L41">[source]</a></span></p>
<h3 id="ones">Ones</h3>
<pre><code class="python">keras.initializers.Ones()
</code></pre>
<p>Initializer that generates tensors initialized to 1.</p>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L49">[source]</a></span></p>
<h3 id="constant">Constant</h3>
<pre><code class="python">keras.initializers.Constant(value=0)
</code></pre>
<p>Initializer that generates tensors initialized to a constant value.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>value</strong>: float; the value of the generator tensors.</li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L66">[source]</a></span></p>
<h3 id="randomnormal">RandomNormal</h3>
<pre><code class="python">keras.initializers.RandomNormal(mean=0.0, stddev=0.05, seed=None)
</code></pre>
<p>Initializer that generates tensors with a normal distribution.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>mean</strong>: a python scalar or a scalar tensor. Mean of the random values
to generate.</li>
<li><strong>stddev</strong>: a python scalar or a scalar tensor. Standard deviation of the
random values to generate.</li>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L97">[source]</a></span></p>
<h3 id="randomuniform">RandomUniform</h3>
<pre><code class="python">keras.initializers.RandomUniform(minval=-0.05, maxval=0.05, seed=None)
</code></pre>
<p>Initializer that generates tensors with a uniform distribution.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>minval</strong>: A python scalar or a scalar tensor. Lower bound of the range
of random values to generate.</li>
<li><strong>maxval</strong>: A python scalar or a scalar tensor. Upper bound of the range
of random values to generate. Defaults to 1 for float types.</li>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L128">[source]</a></span></p>
<h3 id="truncatednormal">TruncatedNormal</h3>
<pre><code class="python">keras.initializers.TruncatedNormal(mean=0.0, stddev=0.05, seed=None)
</code></pre>
<p>Initializer that generates a truncated normal distribution.</p>
<p>These values are similar to values from a <code>RandomNormal</code>
except that values more than two standard deviations from the mean
are discarded and redrawn. This is the recommended initializer for
neural network weights and filters.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>mean</strong>: a python scalar or a scalar tensor. Mean of the random values
to generate.</li>
<li><strong>stddev</strong>: a python scalar or a scalar tensor. Standard deviation of the
random values to generate.</li>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L164">[source]</a></span></p>
<h3 id="variancescaling">VarianceScaling</h3>
<pre><code class="python">keras.initializers.VarianceScaling(scale=1.0, mode='fan_in', distribution='normal', seed=None)
</code></pre>
<p>Initializer capable of adapting its scale to the shape of weights.</p>
<p>With <code>distribution="normal"</code>, samples are drawn from a truncated normal
distribution centered on zero, with <code>stddev = sqrt(scale / n)</code> where n is:</p>
<ul>
<li>number of input units in the weight tensor, if mode = "fan_in"</li>
<li>number of output units, if mode = "fan_out"</li>
<li>average of the numbers of input and output units, if mode = "fan_avg"</li>
</ul>
<p>With <code>distribution="uniform"</code>,
samples are drawn from a uniform distribution
within [-limit, limit], with <code>limit = sqrt(3 * scale / n)</code>.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>scale</strong>: Scaling factor (positive float).</li>
<li><strong>mode</strong>: One of "fan_in", "fan_out", "fan_avg".</li>
<li><strong>distribution</strong>: Random distribution to use. One of "normal", "uniform".</li>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Raises</strong></p>
<ul>
<li><strong>ValueError</strong>: In case of an invalid value for the "scale", mode" or
"distribution" arguments.</li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L241">[source]</a></span></p>
<h3 id="orthogonal">Orthogonal</h3>
<pre><code class="python">keras.initializers.Orthogonal(gain=1.0, seed=None)
</code></pre>
<p>Initializer that generates a random orthogonal matrix.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>gain</strong>: Multiplicative factor to apply to the orthogonal matrix.</li>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>References</strong></p>
<ul>
<li><a href="http://arxiv.org/abs/1312.6120">Exact solutions to the nonlinear dynamics of learning in deep
linear neural networks</a></li>
</ul>
<hr />
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/initializers.py#L281">[source]</a></span></p>
<h3 id="identity">Identity</h3>
<pre><code class="python">keras.initializers.Identity(gain=1.0)
</code></pre>
<p>Initializer that generates the identity matrix.</p>
<p>Only use for 2D matrices.
If the desired matrix is not square, it gets padded
with zeros for the additional rows/columns.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>gain</strong>: Multiplicative factor to apply to the identity matrix.</li>
</ul>
<hr />
<h3 id="lecun_uniform">lecun_uniform</h3>
<pre><code class="python">keras.initializers.lecun_uniform(seed=None)
</code></pre>
<p>LeCun uniform initializer.</p>
<p>It draws samples from a uniform distribution within [-limit, limit]
where <code>limit</code> is <code>sqrt(3 / fan_in)</code>
where <code>fan_in</code> is the number of input units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="http://yann.lecun.com/exdb/publis/pdf/lecun-98b.pdf">Efficient BackProp</a></li>
</ul>
<hr />
<h3 id="glorot_normal">glorot_normal</h3>
<pre><code class="python">keras.initializers.glorot_normal(seed=None)
</code></pre>
<p>Glorot normal initializer, also called Xavier normal initializer.</p>
<p>It draws samples from a truncated normal distribution centered on 0
with <code>stddev = sqrt(2 / (fan_in + fan_out))</code>
where <code>fan_in</code> is the number of input units in the weight tensor
and <code>fan_out</code> is the number of output units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf">Understanding the difficulty of training deep feedforward neural
networks</a></li>
</ul>
<hr />
<h3 id="glorot_uniform">glorot_uniform</h3>
<pre><code class="python">keras.initializers.glorot_uniform(seed=None)
</code></pre>
<p>Glorot uniform initializer, also called Xavier uniform initializer.</p>
<p>It draws samples from a uniform distribution within [-limit, limit]
where <code>limit</code> is <code>sqrt(6 / (fan_in + fan_out))</code>
where <code>fan_in</code> is the number of input units in the weight tensor
and <code>fan_out</code> is the number of output units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf">Understanding the difficulty of training deep feedforward neural
networks</a></li>
</ul>
<hr />
<h3 id="he_normal">he_normal</h3>
<pre><code class="python">keras.initializers.he_normal(seed=None)
</code></pre>
<p>He normal initializer.</p>
<p>It draws samples from a truncated normal distribution centered on 0
with <code>stddev = sqrt(2 / fan_in)</code>
where <code>fan_in</code> is the number of input units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="http://arxiv.org/abs/1502.01852">Delving Deep into Rectifiers: Surpassing Human-Level Performance on
ImageNet Classification</a></li>
</ul>
<hr />
<h3 id="lecun_normal">lecun_normal</h3>
<pre><code class="python">keras.initializers.lecun_normal(seed=None)
</code></pre>
<p>LeCun normal initializer.</p>
<p>It draws samples from a truncated normal distribution centered on 0
with <code>stddev = sqrt(1 / fan_in)</code>
where <code>fan_in</code> is the number of input units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="https://arxiv.org/abs/1706.02515">Self-Normalizing Neural Networks</a></li>
<li><a href="http://yann.lecun.com/exdb/publis/pdf/lecun-98b.pdf">Efficient Backprop</a></li>
</ul>
<hr />
<h3 id="he_uniform">he_uniform</h3>
<pre><code class="python">keras.initializers.he_uniform(seed=None)
</code></pre>
<p>He uniform variance scaling initializer.</p>
<p>It draws samples from a uniform distribution within [-limit, limit]
where <code>limit</code> is <code>sqrt(6 / fan_in)</code>
where <code>fan_in</code> is the number of input units in the weight tensor.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: A Python integer. Used to seed the random generator.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An initializer.</p>
<p><strong>References</strong></p>
<ul>
<li><a href="http://arxiv.org/abs/1502.01852">Delving Deep into Rectifiers: Surpassing Human-Level Performance on
ImageNet Classification</a></li>
</ul>
<p>An initializer may be passed as a string (must match one of the available initializers above), or as a callable:</p>
<pre><code class="python">from keras import initializers
model.add(Dense(64, kernel_initializer=initializers.random_normal(stddev=0.01)))
# also works; will use the default parameters.
model.add(Dense(64, kernel_initializer='random_normal'))
</code></pre>
<h2 id="using-custom-initializers">Using custom initializers</h2>
<p>If passing a custom callable, then it must take the argument <code>shape</code> (shape of the variable to initialize) and <code>dtype</code> (dtype of generated values):</p>
<pre><code class="python">from keras import backend as K
def my_init(shape, dtype=None):
return K.random_normal(shape, dtype=dtype)
model.add(Dense(64, kernel_initializer=my_init))
</code></pre>
</div>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="../regularizers/" class="btn btn-neutral float-right" title="Regularizers">Next <span class="icon icon-circle-arrow-right"></span></a>
<a href="../backend/" class="btn btn-neutral" title="Backend"><span class="icon icon-circle-arrow-left"></span> Previous</a>
</div>
<hr/>
<div role="contentinfo">
<!-- Copyright etc -->
</div>
Built with <a href="https://www.mkdocs.org/">MkDocs</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<div class="rst-versions" role="note" aria-label="versions">
<span class="rst-current-version" data-toggle="rst-current-version">
<a href="http://github.com/keras-team/keras/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
<span><a href="../backend/" style="color: #fcfcfc;">« Previous</a></span>
<span style="margin-left: 15px"><a href="../regularizers/" style="color: #fcfcfc">Next »</a></span>
</span>
</div>
<script>var base_url = '..';</script>
<script src="../js/theme.js" defer></script>
<script src="../search/main.js" defer></script>
<script type="text/javascript" defer>
window.onload = function () {
SphinxRtdTheme.Navigation.enable(true);
};
</script>
</body>
</html>
|
