forked from sooftware/kospeech
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Deep Speech 2.html
341 lines (222 loc) · 20.1 KB
/
Deep Speech 2.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
<!DOCTYPE html>
<html class="writer-html5" lang="en" >
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Deep Speech 2 — KoSpeech latest documentation</title>
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<!--[if lt IE 9]>
<script src="_static/js/html5shiv.min.js"></script>
<![endif]-->
<script type="text/javascript" id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
<script src="_static/jquery.js"></script>
<script src="_static/underscore.js"></script>
<script src="_static/doctools.js"></script>
<script src="_static/language_data.js"></script>
<script type="text/javascript" src="_static/js/theme.js"></script>
<link rel="index" title="Index" href="genindex.html" />
<link rel="search" title="Search" href="search.html" />
<link rel="next" title="Listen, Attend and Spell" href="Listen%20Attend%20Spell.html" />
<link rel="prev" title="Interface" href="Interface.html" />
</head>
<body class="wy-body-for-nav">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side">
<div class="wy-side-scroll">
<div class="wy-side-nav-search" >
<a href="index.html" class="icon icon-home"> KoSpeech
</a>
<div class="version">
latest
</div>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="search.html" method="get">
<input type="text" name="q" placeholder="Search docs" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div>
<div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<p class="caption"><span class="caption-text">GETTING STARTED</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html">What’s New</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#note">Note</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#kospeech-open-source-toolkit-for-end-to-end-korean-speech-recognition-paper"><em><strong>KoSpeech: Open-Source Toolkit for End-to-End Korean Speech Recognition [Paper]</strong></em></a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#supported-models">Supported Models</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#pre-processed-transcripts">Pre-processed Transcripts</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#introduction">Introduction</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#roadmap">Roadmap</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#installation">Installation</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#get-started">Get Started</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#troubleshoots-and-contributing">Troubleshoots and Contributing</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/intro.html#citation">Citation</a></li>
<li class="toctree-l1"><a class="reference internal" href="notes/Preparation.html">KsponSpeech-preprocess</a></li>
</ul>
<p class="caption"><span class="caption-text">ARCHITECTURE</span></p>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="Interface.html">Interface</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">Deep Speech 2</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#id1">Deep Speech 2</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="Listen%20Attend%20Spell.html">Listen, Attend and Spell</a></li>
<li class="toctree-l1"><a class="reference internal" href="RNN%20Transducer.html">RNN Transducer</a></li>
<li class="toctree-l1"><a class="reference internal" href="Speech%20Transformer.html">Speech Transformer</a></li>
<li class="toctree-l1"><a class="reference internal" href="Jasper.html">Jasper</a></li>
<li class="toctree-l1"><a class="reference internal" href="Conformer.html">Conformer</a></li>
<li class="toctree-l1"><a class="reference internal" href="Modules.html">Modules</a></li>
</ul>
<p class="caption"><span class="caption-text">LIBRARY REFERENCE</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="Checkpoint.html">Checkpoint</a></li>
<li class="toctree-l1"><a class="reference internal" href="Criterion.html">Criterion</a></li>
<li class="toctree-l1"><a class="reference internal" href="Data.html">Data</a></li>
<li class="toctree-l1"><a class="reference internal" href="Decode.html">Decode</a></li>
<li class="toctree-l1"><a class="reference internal" href="Evaluator.html">Evaluator</a></li>
<li class="toctree-l1"><a class="reference internal" href="Optim.html">Optim</a></li>
<li class="toctree-l1"><a class="reference internal" href="Learning%20Rate%20Schedulers.html">Learning Rate Scheduler</a></li>
<li class="toctree-l1"><a class="reference internal" href="Trainer.html">Trainer</a></li>
<li class="toctree-l1"><a class="reference internal" href="Vocabs.html">Vocabs</a></li>
<li class="toctree-l1"><a class="reference internal" href="Etc.html">Etc</a></li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
<nav class="wy-nav-top" aria-label="top navigation">
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="index.html">KoSpeech</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="index.html" class="icon icon-home"></a> »</li>
<li>Deep Speech 2</li>
<li class="wy-breadcrumbs-aside">
<a href="_sources/Deep Speech 2.rst.txt" rel="nofollow"> View page source</a>
</li>
</ul>
<hr/>
</div>
<div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
<div itemprop="articleBody">
<div class="section" id="deep-speech-2">
<h1>Deep Speech 2<a class="headerlink" href="#deep-speech-2" title="Permalink to this headline">¶</a></h1>
<div class="section" id="id1">
<h2>Deep Speech 2<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h2>
<span class="target" id="module-kospeech.models.deepspeech2.model"></span><dl class="py class">
<dt id="kospeech.models.deepspeech2.model.BNReluRNN">
<em class="property">class </em><code class="sig-prename descclassname">kospeech.models.deepspeech2.model.</code><code class="sig-name descname">BNReluRNN</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">input_size</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span></em>, <em class="sig-param"><span class="n">hidden_state_dim</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span> <span class="o">=</span> <span class="default_value">512</span></em>, <em class="sig-param"><span class="n">rnn_type</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.9)">str</a></span> <span class="o">=</span> <span class="default_value">'gru'</span></em>, <em class="sig-param"><span class="n">bidirectional</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.9)">bool</a></span> <span class="o">=</span> <span class="default_value">True</span></em>, <em class="sig-param"><span class="n">dropout_p</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.9)">float</a></span> <span class="o">=</span> <span class="default_value">0.1</span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/kospeech/models/deepspeech2/model.html#BNReluRNN"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#kospeech.models.deepspeech2.model.BNReluRNN" title="Permalink to this definition">¶</a></dt>
<dd><p>Recurrent neural network with batch normalization layer & ReLU activation function.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_size</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a>) – size of input</p></li>
<li><p><strong>hidden_state_dim</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a>) – the number of features in the hidden state <cite>h</cite></p></li>
<li><p><strong>rnn_type</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.9)"><em>str</em></a><em>, </em><em>optional</em>) – type of RNN cell (default: gru)</p></li>
<li><p><strong>bidirectional</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.9)"><em>bool</em></a><em>, </em><em>optional</em>) – if True, becomes a bidirectional encoder (defulat: True)</p></li>
<li><p><strong>dropout_p</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.9)"><em>float</em></a><em>, </em><em>optional</em>) – dropout probability (default: 0.1)</p></li>
</ul>
</dd>
</dl>
<dl class="simple">
<dt>Inputs: inputs, input_lengths</dt><dd><ul class="simple">
<li><p><strong>inputs</strong> (batch, time, dim): Tensor containing input vectors</p></li>
<li><p><strong>input_lengths</strong>: Tensor containing containing sequence lengths</p></li>
</ul>
</dd>
<dt>Returns: outputs</dt><dd><ul class="simple">
<li><p><strong>outputs</strong>: Tensor produced by the BNReluRNN module</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="py class">
<dt id="kospeech.models.deepspeech2.model.DeepSpeech2">
<em class="property">class </em><code class="sig-prename descclassname">kospeech.models.deepspeech2.model.</code><code class="sig-name descname">DeepSpeech2</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">input_dim</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span></em>, <em class="sig-param"><span class="n">num_classes</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span></em>, <em class="sig-param"><span class="n">rnn_type</span><span class="o">=</span><span class="default_value">'gru'</span></em>, <em class="sig-param"><span class="n">num_rnn_layers</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span> <span class="o">=</span> <span class="default_value">5</span></em>, <em class="sig-param"><span class="n">rnn_hidden_dim</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)">int</a></span> <span class="o">=</span> <span class="default_value">512</span></em>, <em class="sig-param"><span class="n">dropout_p</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.9)">float</a></span> <span class="o">=</span> <span class="default_value">0.1</span></em>, <em class="sig-param"><span class="n">bidirectional</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.9)">bool</a></span> <span class="o">=</span> <span class="default_value">True</span></em>, <em class="sig-param"><span class="n">activation</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.9)">str</a></span> <span class="o">=</span> <span class="default_value">'hardtanh'</span></em>, <em class="sig-param"><span class="n">device</span><span class="p">:</span> <span class="n">torch.device</span> <span class="o">=</span> <span class="default_value">'cuda'</span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/kospeech/models/deepspeech2/model.html#DeepSpeech2"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#kospeech.models.deepspeech2.model.DeepSpeech2" title="Permalink to this definition">¶</a></dt>
<dd><p>Deep Speech2 model with configurable encoder and decoder.
Paper: <a class="reference external" href="https://arxiv.org/abs/1512.02595">https://arxiv.org/abs/1512.02595</a></p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_dim</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a>) – dimension of input vector</p></li>
<li><p><strong>num_classes</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a>) – number of classfication</p></li>
<li><p><strong>rnn_type</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.9)"><em>str</em></a><em>, </em><em>optional</em>) – type of RNN cell (default: gru)</p></li>
<li><p><strong>num_rnn_layers</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a><em>, </em><em>optional</em>) – number of recurrent layers (default: 5)</p></li>
<li><p><strong>rnn_hidden_dim</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.9)"><em>int</em></a>) – the number of features in the hidden state <cite>h</cite></p></li>
<li><p><strong>dropout_p</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.9)"><em>float</em></a><em>, </em><em>optional</em>) – dropout probability (default: 0.1)</p></li>
<li><p><strong>bidirectional</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.9)"><em>bool</em></a><em>, </em><em>optional</em>) – if True, becomes a bidirectional encoder (defulat: True)</p></li>
<li><p><strong>activation</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.9)"><em>str</em></a>) – type of activation function (default: hardtanh)</p></li>
<li><p><strong>device</strong> (<em>torch.device</em>) – device - ‘cuda’ or ‘cpu’</p></li>
</ul>
</dd>
</dl>
<dl class="simple">
<dt>Inputs: inputs, input_lengths</dt><dd><ul class="simple">
<li><p><strong>inputs</strong>: list of sequences, whose length is the batch size and within which each sequence is list of tokens</p></li>
<li><p><strong>input_lengths</strong>: list of sequence lengths</p></li>
</ul>
</dd>
<dt>Returns: output</dt><dd><ul class="simple">
<li><p><strong>output</strong>: tensor containing the encoded features of the input sequence</p></li>
</ul>
</dd>
</dl>
<dl class="py method">
<dt id="kospeech.models.deepspeech2.model.DeepSpeech2.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">inputs</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://pytorch.org/docs/master/tensors.html#torch.Tensor" title="(in PyTorch vmaster (1.8.0a0+4b6fea9 ))">torch.Tensor</a></span></em>, <em class="sig-param"><span class="n">input_lengths</span><span class="p">:</span> <span class="n"><a class="reference external" href="https://pytorch.org/docs/master/tensors.html#torch.Tensor" title="(in PyTorch vmaster (1.8.0a0+4b6fea9 ))">torch.Tensor</a></span></em><span class="sig-paren">)</span> → Tuple<span class="p">[</span><a class="reference external" href="https://pytorch.org/docs/master/tensors.html#torch.Tensor" title="(in PyTorch vmaster (1.8.0a0+4b6fea9 ))">torch.Tensor</a><span class="p">, </span><a class="reference external" href="https://pytorch.org/docs/master/tensors.html#torch.Tensor" title="(in PyTorch vmaster (1.8.0a0+4b6fea9 ))">torch.Tensor</a><span class="p">]</span><a class="reference internal" href="_modules/kospeech/models/deepspeech2/model.html#DeepSpeech2.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#kospeech.models.deepspeech2.model.DeepSpeech2.forward" title="Permalink to this definition">¶</a></dt>
<dd><p>Forward propagate a <cite>inputs</cite> for ctc training.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>inputs</strong> (<em>torch.FloatTensor</em>) – A input sequence passed to encoder. Typically for inputs this will be a padded
<cite>FloatTensor</cite> of size <code class="docutils literal notranslate"><span class="pre">(batch,</span> <span class="pre">seq_length,</span> <span class="pre">dimension)</span></code>.</p></li>
<li><p><strong>input_lengths</strong> (<em>torch.LongTensor</em>) – The length of input tensor. <code class="docutils literal notranslate"><span class="pre">(batch)</span></code></p></li>
</ul>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p><ul class="simple">
<li><p>predicted_log_prob (torch.FloatTensor)s: Log probability of model predictions.</p></li>
<li><p>output_lengths (torch.LongTensor): The length of output tensor <code class="docutils literal notranslate"><span class="pre">(batch)</span></code></p></li>
</ul>
</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>(Tensor, Tensor)</p>
</dd>
</dl>
</dd></dl>
</dd></dl>
</div>
</div>
</div>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="Listen%20Attend%20Spell.html" class="btn btn-neutral float-right" title="Listen, Attend and Spell" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right" aria-hidden="true"></span></a>
<a href="Interface.html" class="btn btn-neutral float-left" title="Interface" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left" aria-hidden="true"></span> Previous</a>
</div>
<hr/>
<div role="contentinfo">
<p>
© Copyright 2020, Soohwan Kim.
</p>
</div>
Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
<a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<script type="text/javascript">
jQuery(function () {
SphinxRtdTheme.Navigation.enable(true);
});
</script>
</body>
</html>