hierarchical_layer_cn.html 18.7 KB
Newer Older
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


<!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 name="viewport" content="width=device-width, initial-scale=1.0">
  
  <title>支持双层序列作为输入的Layer &mdash; PaddlePaddle  文档</title>
  

  
  

  

  
  
    

  

  
  
27
    <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
28 29 30 31 32
  

  
  
        <link rel="index" title="索引"
33 34 35
              href="../../genindex.html"/>
        <link rel="search" title="搜索" href="../../search.html"/>
    <link rel="top" title="PaddlePaddle  文档" href="../../index.html"/>
36
        <link rel="up" title="RNN模型" href="index_cn.html"/>
37 38 39 40
        <link rel="next" title="单双层RNN API对比介绍" href="hrnn_rnn_api_compare_cn.html"/>
        <link rel="prev" title="Recurrent Group教程" href="recurrent_group_cn.html"/> 

  <link rel="stylesheet" href="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/css/perfect-scrollbar.min.css" type="text/css" />
41
  <link rel="stylesheet" href="../../_static/css/override.css" type="text/css" />
42 43 44 45 46 47 48 49 50 51 52 53 54
  <script>
  var _hmt = _hmt || [];
  (function() {
    var hm = document.createElement("script");
    hm.src = "//hm.baidu.com/hm.js?b9a314ab40d04d805655aab1deee08ba";
    var s = document.getElementsByTagName("script")[0]; 
    s.parentNode.insertBefore(hm, s);
  })();
  </script>

  

  
55
  <script src="../../_static/js/modernizr.min.js"></script>
56 57 58 59 60 61 62 63

</head>

<body class="wy-body-for-nav" role="document">

  
  <header class="site-header">
    <div class="site-logo">
64
      <a href="/"><img src="../../_static/images/PP_w.png"></a>
65 66 67
    </div>
    <div class="site-nav-links">
      <div class="site-menu">
68
        <a class="fork-on-github" href="https://github.com/PaddlePaddle/Paddle" target="_blank"><i class="fa fa-github"></i>Fork me on Github</a>
69 70 71 72 73 74 75 76 77 78 79 80
        <div class="language-switcher dropdown">
          <a type="button" data-toggle="dropdown">
            <span>English</span>
            <i class="fa fa-angle-up"></i>
            <i class="fa fa-angle-down"></i>
          </a>
          <ul class="dropdown-menu">
            <li><a href="/doc_cn">中文</a></li>
            <li><a href="/doc">English</a></li>
          </ul>
        </div>
        <ul class="site-page-links">
81
          <li><a href="/">Home</a></li>
82 83 84 85 86
        </ul>
      </div>
      <div class="doc-module">
        
        <ul class="current">
87 88 89 90 91
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_cn.html">安装与编译</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="../index_cn.html">进阶使用</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_cn.html">开发标准</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</a></li>
92 93 94 95
</ul>

        
<div role="search">
96
  <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
    <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>
  </header>
  
  <div class="main-content-wrap">

    
    <nav class="doc-menu-vertical" role="navigation">
        
          
          <ul class="current">
113 114 115
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/quickstart_cn.html">快速开始</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/concepts/use_concepts_cn.html">基本使用概念</a></li>
116 117
</ul>
</li>
118 119 120 121 122
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_cn.html">安装与编译</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/pip_install_cn.html">使用pip安装</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/docker_install_cn.html">使用Docker安装运行</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/build_cn.html">用Docker编译和测试PaddlePaddle</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/build_from_source_cn.html">从源码编译</a></li>
123 124
</ul>
</li>
125 126 127 128 129
<li class="toctree-l1 current"><a class="reference internal" href="../index_cn.html">进阶使用</a><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="../cmd_parameter/index_cn.html">命令行参数设置</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/use_case_cn.html">使用案例</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/arguments_cn.html">参数概述</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/detail_introduction_cn.html">细节描述</a></li>
130 131
</ul>
</li>
132 133 134 135 136 137 138 139 140
<li class="toctree-l2"><a class="reference internal" href="../cluster/index_cn.html">分布式训练</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../cluster/preparations_cn.html">环境准备</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cluster/cmd_argument_cn.html">启动参数说明</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cluster/multi_cluster/index_cn.html">在不同集群中运行</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../cluster/multi_cluster/fabric_cn.html">使用fabric启动集群训练</a></li>
<li class="toctree-l4"><a class="reference internal" href="../cluster/multi_cluster/openmpi_cn.html">在OpenMPI集群中提交训练作业</a></li>
<li class="toctree-l4"><a class="reference internal" href="../cluster/multi_cluster/k8s_cn.html">Kubernetes单机训练</a></li>
<li class="toctree-l4"><a class="reference internal" href="../cluster/multi_cluster/k8s_distributed_cn.html">Kubernetes分布式训练</a></li>
<li class="toctree-l4"><a class="reference internal" href="../cluster/multi_cluster/k8s_aws_cn.html">Distributed PaddlePaddle Training on AWS with Kubernetes</a></li>
141 142
</ul>
</li>
143 144
</ul>
</li>
145 146 147 148 149 150
<li class="toctree-l2"><a class="reference internal" href="../capi/index_cn.html">C-API预测库</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../capi/compile_paddle_lib_cn.html">安装与编译C-API预测库</a></li>
<li class="toctree-l3"><a class="reference internal" href="../capi/organization_of_the_inputs_cn.html">输入/输出数据组织</a></li>
<li class="toctree-l3"><a class="reference internal" href="../capi/workflow_of_capi_cn.html">C-API使用流程</a></li>
</ul>
</li>
151
<li class="toctree-l2 current"><a class="reference internal" href="index_cn.html">RNN模型</a><ul class="current">
152
<li class="toctree-l3"><a class="reference internal" href="rnn_config_cn.html">RNN配置</a></li>
153 154 155 156 157
<li class="toctree-l3"><a class="reference internal" href="recurrent_group_cn.html">Recurrent Group教程</a></li>
<li class="toctree-l3 current"><a class="current reference internal" href="#">支持双层序列作为输入的Layer</a></li>
<li class="toctree-l3"><a class="reference internal" href="hrnn_rnn_api_compare_cn.html">单双层RNN API对比介绍</a></li>
</ul>
</li>
158 159 160 161 162 163
<li class="toctree-l2"><a class="reference internal" href="../optimization/gpu_profiling_cn.html">GPU性能调优</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_cn.html">开发标准</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../dev/contribute_to_paddle_cn.html">如何贡献代码</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../dev/write_docs_cn.html">如何贡献文档</a></li>
164 165
</ul>
</li>
166 167 168 169 170 171
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../faq/build_and_install/index_cn.html">编译安装与单元测试</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/model/index_cn.html">模型配置</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/parameter/index_cn.html">参数设置</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/local/index_cn.html">本地训练与预测</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/cluster/index_cn.html">集群训练与预测</a></li>
172 173
</ul>
</li>
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
</ul>

        
    </nav>
    
    <section class="doc-content-wrap">

      

 







<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
194
        <li><a href="../index_cn.html">进阶使用</a> > </li>
195
      
196
        <li><a href="index_cn.html">RNN模型</a> > </li>
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
      
    <li>支持双层序列作为输入的Layer</li>
  </ul>
</div>
      
      <div class="wy-nav-content" id="doc-content">
        <div class="rst-content">
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="layer">
<h1><a class="toc-backref" href="#id2">支持双层序列作为输入的Layer</a><a class="headerlink" href="#layer" title="永久链接至标题"></a></h1>
<div class="contents topic" id="contents">
<p class="topic-title first">Contents</p>
<ul class="simple">
<li><a class="reference internal" href="#layer" id="id2">支持双层序列作为输入的Layer</a><ul>
<li><a class="reference internal" href="#id1" id="id3">概述</a></li>
214
<li><a class="reference internal" href="#pooling" id="id4">pooling</a></li>
215
<li><a class="reference internal" href="#last-seq-first-seq" id="id5">last_seq 和 first_seq</a></li>
216
<li><a class="reference internal" href="#expand" id="id6">expand</a></li>
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
</ul>
</li>
</ul>
</div>
<div class="section" id="id1">
<h2><a class="toc-backref" href="#id3">概述</a><a class="headerlink" href="#id1" title="永久链接至标题"></a></h2>
<p>在自然语言处理任务中,序列是一种常见的数据类型。一个独立的词语,可以看作是一个非序列输入,或者,我们称之为一个0层的序列;由词语构成的句子,是一个单层序列;若干个句子构成一个段落,是一个双层的序列。</p>
<p>双层序列是一个嵌套的序列,它的每一个元素,又是一个单层的序列。这是一种非常灵活的数据组织方式,帮助我们构造一些复杂的输入信息。</p>
<p>我们可以按照如下层次定义非序列,单层序列,以及双层序列。</p>
<ul class="simple">
<li>0层序列:一个独立的元素,类型可以是PaddlePaddle支持的任意输入数据类型</li>
<li>单层序列:排成一列的多个元素,每个元素是一个0层序列,元素之间的顺序是重要的输入信息</li>
<li>双层序列:排成一列的多个元素,每个元素是一个单层序列,称之为双层序列的一个子序列(subseq),subseq的每个元素是一个0层序列</li>
</ul>
<p>在 PaddlePaddle中,下面这些Layer能够接受双层序列作为输入,完成相应的计算。</p>
</div>
233 234
<div class="section" id="pooling">
<h2><a class="toc-backref" href="#id4">pooling</a><a class="headerlink" href="#pooling" title="永久链接至标题"></a></h2>
235
<p>pooling 的使用示例如下,详细见 <a class="reference internal" href="../../api/v2/config/layer.html#api-v2-layer-pooling"><span class="std std-ref">pooling</span></a> 配置API。</p>
236 237
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nv">seq_pool</span> <span class="o">=</span> pooling<span class="o">(</span><span class="nv">input</span><span class="o">=</span>layer,
                   <span class="nv">pooling_type</span><span class="o">=</span>pooling.Max<span class="o">()</span>,
238
                   <span class="nv">agg_level</span><span class="o">=</span>AggregateLevel.TO_SEQUENCE<span class="o">)</span>
239 240 241
</pre></div>
</div>
<ul class="simple">
242
<li><cite>pooling_type</cite> 目前支持两种,分别是:pooling.Max()和pooling.Avg()。</li>
243
<li><cite>agg_level=AggregateLevel.TO_NO_SEQUENCE</cite> 时(默认值):<ul>
244 245 246 247 248
<li>作用:双层序列经过运算变成一个0层序列,或单层序列经过运算变成一个0层序列</li>
<li>输入:一个双层序列,或一个单层序列</li>
<li>输出:一个0层序列,即整个输入序列(单层或双层)的平均值(或最大值)</li>
</ul>
</li>
249
<li><cite>agg_level=AggregateLevel.TO_SEQUENCE</cite> 时:<ul>
250 251 252 253 254 255 256 257 258
<li>作用:一个双层序列经过运算变成一个单层序列</li>
<li>输入:必须是一个双层序列</li>
<li>输出:一个单层序列,序列的每个元素是原来双层序列每个subseq元素的平均值(或最大值)</li>
</ul>
</li>
</ul>
</div>
<div class="section" id="last-seq-first-seq">
<h2><a class="toc-backref" href="#id5">last_seq 和 first_seq</a><a class="headerlink" href="#last-seq-first-seq" title="永久链接至标题"></a></h2>
259
<p>last_seq 的使用示例如下( <a class="reference internal" href="../../api/v2/config/layer.html#api-v2-layer-first-seq"><span class="std std-ref">first_seq</span></a> 类似),详细见 <a class="reference internal" href="../../api/v2/config/layer.html#api-v2-layer-last-seq"><span class="std std-ref">last_seq</span></a> 配置API。</p>
260
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nv">last</span> <span class="o">=</span> last_seq<span class="o">(</span><span class="nv">input</span><span class="o">=</span>layer,
261
                <span class="nv">agg_level</span><span class="o">=</span>AggregateLevel.TO_SEQUENCE<span class="o">)</span>
262 263 264
</pre></div>
</div>
<ul class="simple">
265
<li><cite>agg_level=AggregateLevel.TO_NO_SEQUENCE</cite> 时(默认值):<ul>
266 267 268 269 270
<li>作用:一个双层序列经过运算变成一个0层序列,或一个单层序列经过运算变成一个0层序列</li>
<li>输入:一个双层序列或一个单层序列</li>
<li>输出:一个0层序列,即整个输入序列(双层或者单层)最后一个,或第一个元素。</li>
</ul>
</li>
271
<li><cite>agg_level=AggregateLevel.TO_SEQUENCE</cite> 时:
272 273 274 275 276
- 作用:一个双层序列经过运算变成一个单层序列
- 输入:必须是一个双层序列
- 输出:一个单层序列,其中每个元素是双层序列中每个subseq最后一个(或第一个)元素。</li>
</ul>
</div>
277 278
<div class="section" id="expand">
<h2><a class="toc-backref" href="#id6">expand</a><a class="headerlink" href="#expand" title="永久链接至标题"></a></h2>
279
<p>expand 的使用示例如下,详细见 <a class="reference internal" href="../../api/v2/config/layer.html#api-v2-layer-expand"><span class="std std-ref">ExpandLevel</span></a> 配置API。</p>
280 281
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nv">ex</span> <span class="o">=</span> expand<span class="o">(</span><span class="nv">input</span><span class="o">=</span>layer1,
            <span class="nv">expand_as</span><span class="o">=</span>layer2,
282
            <span class="nv">expand_level</span><span class="o">=</span>ExpandLevel.FROM_NO_SEQUENCE<span class="o">)</span>
283 284 285
</pre></div>
</div>
<ul class="simple">
286
<li><cite>expand_level=ExpandLevel.FROM_NO_SEQUENCE</cite> 时(默认值):<ul>
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
<li>作用:一个0层序列经过运算扩展成一个单层序列,或者一个双层序列</li>
<li>输入:layer1必须是一个0层序列,是待扩展的数据;layer2 可以是一个单层序列,或者是一个双层序列,提供扩展的长度信息</li>
<li>输出:一个单层序列或一个双层序列,输出序列的类型(双层序列或单层序列)和序列中含有元素的数目同 layer2 一致。若输出是单层序列,单层序列的每个元素(0层序列),都是对layer1元素的拷贝;若输出是双层序列,双层序列每个subseq中每个元素(0层序列),都是对layer1元素的拷贝</li>
</ul>
</li>
<li><cite>expand_level=ExpandLevel.FROM_SEQUENCE</cite> 时:<ul>
<li>作用:一个单层序列经过运算扩展成一个双层序列</li>
<li>输入:layer1必须是一个单层序列,是待扩展的数据;layer2 必须是一个双层序列,提供扩展的长度信息</li>
<li>输出:一个双层序列,序列中含有元素的数目同 layer2 一致。要求单层序列含有元素的数目(0层序列)和双层序列含有subseq 的数目一致。单层序列第i个元素(0层序列),被扩展为一个单层序列,构成了输出双层序列的第i个 subseq 。</li>
</ul>
</li>
</ul>
</div>
</div>


           </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
        <a href="hrnn_rnn_api_compare_cn.html" class="btn btn-neutral float-right" title="单双层RNN API对比介绍" accesskey="n">Next <span class="fa fa-arrow-circle-right"></span></a>
      
      
        <a href="recurrent_group_cn.html" class="btn btn-neutral" title="Recurrent Group教程" accesskey="p"><span class="fa fa-arrow-circle-left"></span> Previous</a>
      
    </div>
  

  <hr/>

  <div role="contentinfo">
    <p>
        &copy; Copyright 2016, PaddlePaddle developers.

    </p>
  </div>
  Built with <a href="http://sphinx-doc.org/">Sphinx</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>
  


  

    <script type="text/javascript">
        var DOCUMENTATION_OPTIONS = {
342
            URL_ROOT:'../../',
343 344 345
            VERSION:'',
            COLLAPSE_INDEX:false,
            FILE_SUFFIX:'.html',
346 347
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
348 349
        };
    </script>
350 351 352 353
      <script type="text/javascript" src="../../_static/jquery.js"></script>
      <script type="text/javascript" src="../../_static/underscore.js"></script>
      <script type="text/javascript" src="../../_static/doctools.js"></script>
      <script type="text/javascript" src="../../_static/translations.js"></script>
354
      <script type="text/javascript" src="https://cdn.bootcss.com/mathjax/2.7.0/MathJax.js"></script>
355 356 357 358 359
       
  

  
  
360
    <script type="text/javascript" src="../../_static/js/theme.js"></script>
361 362 363 364
  
  
  <script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js" integrity="sha384-Tc5IQib027qvyjSMfHjOMaLkfuWVxZxUPnCJA7l2mCWNIpG9mGCD8wGNIcPD7Txa" crossorigin="anonymous"></script>
  <script src="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/js/perfect-scrollbar.jquery.min.js"></script>
365
  <script src="../../_static/js/paddle_doc_init.js"></script> 
366 367

</body>
368
</html>