| <!DOCTYPE html> |
| |
| <html lang="en"> |
| <head> |
| <meta charset="utf-8"/> |
| <meta content="IE=edge" http-equiv="X-UA-Compatible"/> |
| <meta content="width=device-width, initial-scale=1" name="viewport"/> |
| <title>Gluon Recurrent Neural Network API — mxnet documentation</title> |
| <link crossorigin="anonymous" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/css/bootstrap.min.css" integrity="sha384-1q8mTJOASx8j1Au+a5WDVnPi2lkFfwwEAa8hDDdjZlpLegxhjVME1fgjWPGmkzs7" rel="stylesheet"/> |
| <link href="https://maxcdn.bootstrapcdn.com/font-awesome/4.5.0/css/font-awesome.min.css" rel="stylesheet"/> |
| <link href="../../../_static/basic.css" rel="stylesheet" type="text/css"> |
| <link href="../../../_static/pygments.css" rel="stylesheet" type="text/css"> |
| <link href="../../../_static/mxnet.css" rel="stylesheet" type="text/css"/> |
| <script type="text/javascript"> |
| var DOCUMENTATION_OPTIONS = { |
| URL_ROOT: '../../../', |
| VERSION: '', |
| COLLAPSE_INDEX: false, |
| FILE_SUFFIX: '.html', |
| HAS_SOURCE: true, |
| SOURCELINK_SUFFIX: '' |
| }; |
| </script> |
| <script src="https://code.jquery.com/jquery-1.11.1.min.js" type="text/javascript"></script> |
| <script src="../../../_static/underscore.js" type="text/javascript"></script> |
| <script src="../../../_static/searchtools_custom.js" type="text/javascript"></script> |
| <script src="../../../_static/doctools.js" type="text/javascript"></script> |
| <script src="../../../_static/selectlang.js" type="text/javascript"></script> |
| <script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML" type="text/javascript"></script> |
| <script type="text/javascript"> jQuery(function() { Search.loadIndex("/searchindex.js"); Search.init();}); </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-96378503-1', 'auto'); |
| ga('send', 'pageview'); |
| |
| </script> |
| <!-- --> |
| <!-- <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="https://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script> --> |
| <!-- --> |
| <link href="gluon.html" rel="up" title="Gluon Package"> |
| <link href="loss.html" rel="next" title="Gluon Loss API"/> |
| <link href="nn.html" rel="prev" title="Gluon Neural Network Layers"/> |
| <link href="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet-icon.png" rel="icon" type="image/png"/> |
| </link></link></link></head> |
| <body background="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet-background-compressed.jpeg" role="document"> |
| <div class="content-block"><div class="navbar navbar-fixed-top"> |
| <div class="container" id="navContainer"> |
| <div class="innder" id="header-inner"> |
| <h1 id="logo-wrap"> |
| <a href="../../../" id="logo"><img src="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet_logo.png"/></a> |
| </h1> |
| <nav class="nav-bar" id="main-nav"> |
| <a class="main-nav-link" href="../../../install/index.html">Install</a> |
| <span id="dropdown-menu-position-anchor"> |
| <a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Gluon <span class="caret"></span></a> |
| <ul class="dropdown-menu navbar-menu" id="package-dropdown-menu"> |
| <li><a class="main-nav-link" href="../../../gluon/index.html">About</a></li> |
| <li><a class="main-nav-link" href="http://gluon.mxnet.io">Tutorials</a></li> |
| </ul> |
| </span> |
| <span id="dropdown-menu-position-anchor"> |
| <a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">API <span class="caret"></span></a> |
| <ul class="dropdown-menu navbar-menu" id="package-dropdown-menu"> |
| <li><a class="main-nav-link" href="../../../api/python/index.html">Python</a></li> |
| <li><a class="main-nav-link" href="../../../api/scala/index.html">Scala</a></li> |
| <li><a class="main-nav-link" href="../../../api/r/index.html">R</a></li> |
| <li><a class="main-nav-link" href="../../../api/julia/index.html">Julia</a></li> |
| <li><a class="main-nav-link" href="../../../api/c++/index.html">C++</a></li> |
| <li><a class="main-nav-link" href="../../../api/perl/index.html">Perl</a></li> |
| </ul> |
| </span> |
| <span id="dropdown-menu-position-anchor-docs"> |
| <a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Docs <span class="caret"></span></a> |
| <ul class="dropdown-menu navbar-menu" id="package-dropdown-menu-docs"> |
| <li><a class="main-nav-link" href="../../../tutorials/index.html">Tutorials</a> |
| <li><a class="main-nav-link" href="../../../faq/index.html">FAQ</a></li> |
| <li><a class="main-nav-link" href="../../../architecture/index.html">Architecture</a></li> |
| <li><a class="main-nav-link" href="https://github.com/apache/incubator-mxnet/tree/master/example">Examples</a></li> |
| <li><a class="main-nav-link" href="../../../model_zoo/index.html">Model Zoo</a></li> |
| </li></ul> |
| </span> |
| <a class="main-nav-link" href="https://github.com/dmlc/mxnet">Github</a> |
| <span id="dropdown-menu-position-anchor-community"> |
| <a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Community <span class="caret"></span></a> |
| <ul class="dropdown-menu navbar-menu" id="package-dropdown-menu-community"> |
| <li><a class="main-nav-link" href="../../../community/index.html">Community</a></li> |
| <li><a class="main-nav-link" href="../../../community/contribute.html">Contribute</a></li> |
| <li><a class="main-nav-link" href="../../../community/powered_by.html">Powered By</a></li> |
| <li><a class="main-nav-link" href="http://discuss.mxnet.io">Discuss</a></li> |
| </ul> |
| </span> |
| <span id="dropdown-menu-position-anchor-version" style="position: relative"><a href="#" class="main-nav-link dropdown-toggle" data-toggle="dropdown" role="button" aria-haspopup="true" aria-expanded="true">Versions(master)<span class="caret"></span></a><ul id="package-dropdown-menu" class="dropdown-menu"><li><a class="main-nav-link" href=https://mxnet.incubator.apache.org/>1.0.0</a></li><li><a class="main-nav-link" href=https://mxnet.incubator.apache.org/versions/0.12.1/index.html>0.12.1</a></li><li><a class="main-nav-link" href=https://mxnet.incubator.apache.org/versions/0.12.0/index.html>0.12.0</a></li><li><a class="main-nav-link" href=https://mxnet.incubator.apache.org/versions/0.11.0/index.html>0.11.0</a></li><li><a class="main-nav-link" href=https://mxnet.incubator.apache.org/versions/master/index.html>master</a></li></ul></span></nav> |
| <script> function getRootPath(){ return "../../../" } </script> |
| <div class="burgerIcon dropdown"> |
| <a class="dropdown-toggle" data-toggle="dropdown" href="#" role="button">☰</a> |
| <ul class="dropdown-menu" id="burgerMenu"> |
| <li><a href="../../../install/index.html">Install</a></li> |
| <li><a class="main-nav-link" href="../../../tutorials/index.html">Tutorials</a></li> |
| <li class="dropdown-submenu"> |
| <a href="#" tabindex="-1">Community</a> |
| <ul class="dropdown-menu"> |
| <li><a href="../../../community/index.html" tabindex="-1">Community</a></li> |
| <li><a href="../../../community/contribute.html" tabindex="-1">Contribute</a></li> |
| <li><a href="../../../community/powered_by.html" tabindex="-1">Powered By</a></li> |
| </ul> |
| </li> |
| <li class="dropdown-submenu"> |
| <a href="#" tabindex="-1">API</a> |
| <ul class="dropdown-menu"> |
| <li><a href="../../../api/python/index.html" tabindex="-1">Python</a> |
| </li> |
| <li><a href="../../../api/scala/index.html" tabindex="-1">Scala</a> |
| </li> |
| <li><a href="../../../api/r/index.html" tabindex="-1">R</a> |
| </li> |
| <li><a href="../../../api/julia/index.html" tabindex="-1">Julia</a> |
| </li> |
| <li><a href="../../../api/c++/index.html" tabindex="-1">C++</a> |
| </li> |
| <li><a href="../../../api/perl/index.html" tabindex="-1">Perl</a> |
| </li> |
| </ul> |
| </li> |
| <li class="dropdown-submenu"> |
| <a href="#" tabindex="-1">Docs</a> |
| <ul class="dropdown-menu"> |
| <li><a href="../../../tutorials/index.html" tabindex="-1">Tutorials</a></li> |
| <li><a href="../../../faq/index.html" tabindex="-1">FAQ</a></li> |
| <li><a href="../../../architecture/index.html" tabindex="-1">Architecture</a></li> |
| <li><a href="https://github.com/apache/incubator-mxnet/tree/master/example" tabindex="-1">Examples</a></li> |
| <li><a href="../../../model_zoo/index.html" tabindex="-1">Model Zoo</a></li> |
| </ul> |
| </li> |
| <li><a href="../../../architecture/index.html">Architecture</a></li> |
| <li><a class="main-nav-link" href="https://github.com/dmlc/mxnet">Github</a></li> |
| <li id="dropdown-menu-position-anchor-version-mobile" class="dropdown-submenu" style="position: relative"><a href="#" tabindex="-1">Versions(master)</a><ul class="dropdown-menu"><li><a tabindex="-1" href=https://mxnet.incubator.apache.org/>1.0.0</a></li><li><a tabindex="-1" href=https://mxnet.incubator.apache.org/versions/0.12.1/index.html>0.12.1</a></li><li><a tabindex="-1" href=https://mxnet.incubator.apache.org/versions/0.12.0/index.html>0.12.0</a></li><li><a tabindex="-1" href=https://mxnet.incubator.apache.org/versions/0.11.0/index.html>0.11.0</a></li><li><a tabindex="-1" href=https://mxnet.incubator.apache.org/versions/master/index.html>master</a></li></ul></li></ul> |
| </div> |
| <div class="plusIcon dropdown"> |
| <a class="dropdown-toggle" data-toggle="dropdown" href="#" role="button"><span aria-hidden="true" class="glyphicon glyphicon-plus"></span></a> |
| <ul class="dropdown-menu dropdown-menu-right" id="plusMenu"></ul> |
| </div> |
| <div id="search-input-wrap"> |
| <form action="../../../search.html" autocomplete="off" class="" method="get" role="search"> |
| <div class="form-group inner-addon left-addon"> |
| <i class="glyphicon glyphicon-search"></i> |
| <input class="form-control" name="q" placeholder="Search" type="text"/> |
| </div> |
| <input name="check_keywords" type="hidden" value="yes"> |
| <input name="area" type="hidden" value="default"/> |
| </input></form> |
| <div id="search-preview"></div> |
| </div> |
| <div id="searchIcon"> |
| <span aria-hidden="true" class="glyphicon glyphicon-search"></span> |
| </div> |
| <!-- <div id="lang-select-wrap"> --> |
| <!-- <label id="lang-select-label"> --> |
| <!-- <\!-- <i class="fa fa-globe"></i> -\-> --> |
| <!-- <span></span> --> |
| <!-- </label> --> |
| <!-- <select id="lang-select"> --> |
| <!-- <option value="en">Eng</option> --> |
| <!-- <option value="zh">中文</option> --> |
| <!-- </select> --> |
| <!-- </div> --> |
| <!-- <a id="mobile-nav-toggle"> |
| <span class="mobile-nav-toggle-bar"></span> |
| <span class="mobile-nav-toggle-bar"></span> |
| <span class="mobile-nav-toggle-bar"></span> |
| </a> --> |
| </div> |
| </div> |
| </div> |
| <script type="text/javascript"> |
| $('body').css('background', 'white'); |
| </script> |
| <div class="container"> |
| <div class="row"> |
| <div aria-label="main navigation" class="sphinxsidebar leftsidebar" role="navigation"> |
| <div class="sphinxsidebarwrapper"> |
| <ul class="current"> |
| <li class="toctree-l1 current"><a class="reference internal" href="../index.html">Python Documents</a><ul class="current"> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#ndarray-api">NDArray API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#symbol-api">Symbol API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#module-api">Module API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#autograd-api">Autograd API</a></li> |
| <li class="toctree-l2 current"><a class="reference internal" href="../index.html#gluon-api">Gluon API</a><ul class="current"> |
| <li class="toctree-l3 current"><a class="reference internal" href="gluon.html">Gluon Package</a><ul class="current"> |
| <li class="toctree-l4 current"><a class="reference internal" href="gluon.html#overview">Overview</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="gluon.html#parameter">Parameter</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="gluon.html#containers">Containers</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="gluon.html#trainer">Trainer</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="gluon.html#utilities">Utilities</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="gluon.html#api-reference">API Reference</a></li> |
| </ul> |
| </li> |
| <li class="toctree-l3"><a class="reference internal" href="nn.html">Gluon Neural Network Layers</a></li> |
| <li class="toctree-l3 current"><a class="current reference internal" href="">Gluon Recurrent Neural Network API</a><ul> |
| <li class="toctree-l4"><a class="reference internal" href="#overview">Overview</a></li> |
| <li class="toctree-l4"><a class="reference internal" href="#api-reference">API Reference</a></li> |
| </ul> |
| </li> |
| <li class="toctree-l3"><a class="reference internal" href="loss.html">Gluon Loss API</a></li> |
| <li class="toctree-l3"><a class="reference internal" href="data.html">Gluon Data API</a></li> |
| <li class="toctree-l3"><a class="reference internal" href="model_zoo.html">Gluon Model Zoo</a></li> |
| <li class="toctree-l3"><a class="reference internal" href="contrib.html">Gluon Contrib API</a></li> |
| </ul> |
| </li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#kvstore-api">KVStore API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#io-api">IO API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#image-api">Image API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#optimization-api">Optimization API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#callback-api">Callback API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#metric-api">Metric API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#run-time-compilation-api">Run-Time Compilation API</a></li> |
| <li class="toctree-l2"><a class="reference internal" href="../index.html#contrib-package">Contrib Package</a></li> |
| </ul> |
| </li> |
| <li class="toctree-l1"><a class="reference internal" href="../../r/index.html">R Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../julia/index.html">Julia Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../c++/index.html">C++ Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../scala/index.html">Scala Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../perl/index.html">Perl Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../../faq/index.html">HowTo Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../../architecture/index.html">System Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../../tutorials/index.html">Tutorials</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../../community/index.html">Community</a></li> |
| </ul> |
| </div> |
| </div> |
| <div class="content"> |
| <div class="page-tracker"></div> |
| <div class="section" id="gluon-recurrent-neural-network-api"> |
| <span id="gluon-recurrent-neural-network-api"></span><h1>Gluon Recurrent Neural Network API<a class="headerlink" href="#gluon-recurrent-neural-network-api" title="Permalink to this headline">¶</a></h1> |
| <div class="section" id="overview"> |
| <span id="overview"></span><h2>Overview<a class="headerlink" href="#overview" title="Permalink to this headline">¶</a></h2> |
| <p>This document lists the recurrent neural network API in Gluon:</p> |
| <div class="section" id="recurrent-layers"> |
| <span id="recurrent-layers"></span><h3>Recurrent Layers<a class="headerlink" href="#recurrent-layers" title="Permalink to this headline">¶</a></h3> |
| <p>Recurrent layers can be used in <code class="docutils literal"><span class="pre">Sequential</span></code> with other regular neural network layers. |
| For example, to construct a sequence labeling model where a prediction is made for each |
| time-step:</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">model</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">nn</span><span class="o">.</span><span class="n">Sequential</span><span class="p">()</span> |
| <span class="k">with</span> <span class="n">model</span><span class="o">.</span><span class="n">name_scope</span><span class="p">():</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">nn</span><span class="o">.</span><span class="n">Embedding</span><span class="p">(</span><span class="mi">30</span><span class="p">,</span> <span class="mi">10</span><span class="p">))</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">LSTM</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="n">flatten</span><span class="o">=</span><span class="bp">False</span><span class="p">))</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">initialize</span><span class="p">()</span> |
| <span class="n">model</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">2</span><span class="p">,</span><span class="mi">3</span><span class="p">)))</span> |
| </pre></div> |
| </div> |
| <table border="1" class="longtable docutils"> |
| <colgroup> |
| <col width="10%"/> |
| <col width="90%"/> |
| </colgroup> |
| <tbody valign="top"> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.RNN" title="mxnet.gluon.rnn.RNN"><code class="xref py py-obj docutils literal"><span class="pre">RNN</span></code></a></td> |
| <td>Applies a multi-layer Elman RNN with <cite>tanh</cite> or <cite>ReLU</cite> non-linearity to an input sequence.</td> |
| </tr> |
| <tr class="row-even"><td><a class="reference internal" href="#mxnet.gluon.rnn.LSTM" title="mxnet.gluon.rnn.LSTM"><code class="xref py py-obj docutils literal"><span class="pre">LSTM</span></code></a></td> |
| <td>Applies a multi-layer long short-term memory (LSTM) RNN to an input sequence.</td> |
| </tr> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.GRU" title="mxnet.gluon.rnn.GRU"><code class="xref py py-obj docutils literal"><span class="pre">GRU</span></code></a></td> |
| <td>Applies a multi-layer gated recurrent unit (GRU) RNN to an input sequence.</td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| <div class="section" id="recurrent-cells"> |
| <span id="recurrent-cells"></span><h3>Recurrent Cells<a class="headerlink" href="#recurrent-cells" title="Permalink to this headline">¶</a></h3> |
| <p>Recurrent cells allows fine-grained control when defining recurrent models. User |
| can explicit step and unroll to construct complex networks. It provides more |
| flexibility but is slower than recurrent layers. Recurrent cells can be stacked |
| with <code class="docutils literal"><span class="pre">SequentialRNNCell</span></code>:</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">model</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">SequentialRNNCell</span><span class="p">()</span> |
| <span class="k">with</span> <span class="n">model</span><span class="o">.</span><span class="n">name_scope</span><span class="p">():</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">LSTMCell</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span> |
| <span class="n">model</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">LSTMCell</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span> |
| <span class="n">states</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">begin_state</span><span class="p">(</span><span class="n">batch_size</span><span class="o">=</span><span class="mi">32</span><span class="p">)</span> |
| <span class="n">inputs</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="mi">32</span><span class="p">,</span> <span class="mi">10</span><span class="p">))</span> |
| <span class="n">outputs</span> <span class="o">=</span> <span class="p">[]</span> |
| <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">5</span><span class="p">):</span> |
| <span class="n">output</span><span class="p">,</span> <span class="n">states</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">[</span><span class="n">i</span><span class="p">],</span> <span class="n">states</span><span class="p">)</span> |
| <span class="n">outputs</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">output</span><span class="p">)</span> |
| </pre></div> |
| </div> |
| <table border="1" class="longtable docutils"> |
| <colgroup> |
| <col width="10%"/> |
| <col width="90%"/> |
| </colgroup> |
| <tbody valign="top"> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.RNNCell" title="mxnet.gluon.rnn.RNNCell"><code class="xref py py-obj docutils literal"><span class="pre">RNNCell</span></code></a></td> |
| <td>Elman RNN recurrent neural network cell.</td> |
| </tr> |
| <tr class="row-even"><td><a class="reference internal" href="#mxnet.gluon.rnn.LSTMCell" title="mxnet.gluon.rnn.LSTMCell"><code class="xref py py-obj docutils literal"><span class="pre">LSTMCell</span></code></a></td> |
| <td>Long-Short Term Memory (LSTM) network cell.</td> |
| </tr> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.GRUCell" title="mxnet.gluon.rnn.GRUCell"><code class="xref py py-obj docutils literal"><span class="pre">GRUCell</span></code></a></td> |
| <td>Gated Rectified Unit (GRU) network cell.</td> |
| </tr> |
| <tr class="row-even"><td><a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell" title="mxnet.gluon.rnn.RecurrentCell"><code class="xref py py-obj docutils literal"><span class="pre">RecurrentCell</span></code></a></td> |
| <td>Abstract base class for RNN cells</td> |
| </tr> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.SequentialRNNCell" title="mxnet.gluon.rnn.SequentialRNNCell"><code class="xref py py-obj docutils literal"><span class="pre">SequentialRNNCell</span></code></a></td> |
| <td>Sequentially stacking multiple RNN cells.</td> |
| </tr> |
| <tr class="row-even"><td><a class="reference internal" href="#mxnet.gluon.rnn.BidirectionalCell" title="mxnet.gluon.rnn.BidirectionalCell"><code class="xref py py-obj docutils literal"><span class="pre">BidirectionalCell</span></code></a></td> |
| <td>Bidirectional RNN cell.</td> |
| </tr> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.DropoutCell" title="mxnet.gluon.rnn.DropoutCell"><code class="xref py py-obj docutils literal"><span class="pre">DropoutCell</span></code></a></td> |
| <td>Applies dropout on input.</td> |
| </tr> |
| <tr class="row-even"><td><a class="reference internal" href="#mxnet.gluon.rnn.ZoneoutCell" title="mxnet.gluon.rnn.ZoneoutCell"><code class="xref py py-obj docutils literal"><span class="pre">ZoneoutCell</span></code></a></td> |
| <td>Applies Zoneout on base cell.</td> |
| </tr> |
| <tr class="row-odd"><td><a class="reference internal" href="#mxnet.gluon.rnn.ResidualCell" title="mxnet.gluon.rnn.ResidualCell"><code class="xref py py-obj docutils literal"><span class="pre">ResidualCell</span></code></a></td> |
| <td>Adds residual connection as described in Wu et al, 2016 (<a class="reference external" href="https://arxiv.org/abs/1609.08144">https://arxiv.org/abs/1609.08144</a>).</td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| </div> |
| <div class="section" id="api-reference"> |
| <span id="api-reference"></span><h2>API Reference<a class="headerlink" href="#api-reference" title="Permalink to this headline">¶</a></h2> |
| <script src="../../_static/js/auto_module_index.js" type="text/javascript"></script><span class="target" id="module-mxnet.gluon.rnn"></span><p>Recurrent neural network module.</p> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.BidirectionalCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">BidirectionalCell</code><span class="sig-paren">(</span><em>l_cell</em>, <em>r_cell</em>, <em>output_prefix='bi_'</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#BidirectionalCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.BidirectionalCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Bidirectional RNN cell.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>l_cell</strong> (<a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell" title="mxnet.gluon.rnn.RecurrentCell"><em>RecurrentCell</em></a>) – Cell for forward unrolling</li> |
| <li><strong>r_cell</strong> (<a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell" title="mxnet.gluon.rnn.RecurrentCell"><em>RecurrentCell</em></a>) – Cell for backward unrolling</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.DropoutCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">DropoutCell</code><span class="sig-paren">(</span><em>rate</em>, <em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#DropoutCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.DropoutCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Applies dropout on input.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><strong>rate</strong> (<em>float</em>) – Percentage of elements to drop out, which |
| is 1 - percentage to retain.</td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(batch_size, size)</cite>.</li> |
| <li><strong>states</strong>: a list of recurrent state tensors.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(batch_size, size)</cite>.</li> |
| <li><strong>next_states</strong>: returns input <cite>states</cite> directly.</li> |
| </ul> |
| </dd> |
| </dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.GRU"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">GRU</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>num_layers=1</em>, <em>layout='TNC'</em>, <em>dropout=0</em>, <em>bidirectional=False</em>, <em>input_size=0</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_layer.html#GRU"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.GRU" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Applies a multi-layer gated recurrent unit (GRU) RNN to an input sequence.</p> |
| <p>For each element in the input sequence, each layer computes the following |
| function:</p> |
| <div class="math"> |
| \[\begin{split}\begin{array}{ll} |
| r_t = sigmoid(W_{ir} x_t + b_{ir} + W_{hr} h_{(t-1)} + b_{hr}) \\ |
| i_t = sigmoid(W_{ii} x_t + b_{ii} + W_hi h_{(t-1)} + b_{hi}) \\ |
| n_t = \tanh(W_{in} x_t + b_{in} + r_t * (W_{hn} h_{(t-1)}+ b_{hn})) \\ |
| h_t = (1 - i_t) * n_t + i_t * h_{(t-1)} \\ |
| \end{array}\end{split}\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, <span class="math">\(x_t\)</span> is the hidden |
| state of the previous layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer, |
| and <span class="math">\(r_t\)</span>, <span class="math">\(i_t\)</span>, <span class="math">\(n_t\)</span> are the reset, input, and new gates, respectively.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – The number of features in the hidden state h</li> |
| <li><strong>num_layers</strong> (<em>int, default 1</em>) – Number of recurrent layers.</li> |
| <li><strong>layout</strong> (<em>str, default 'TNC'</em>) – The format of input and output tensors. T, N and C stand for |
| sequence length, batch size, and feature dimensions respectively.</li> |
| <li><strong>dropout</strong> (<em>float, default 0</em>) – If non-zero, introduces a dropout layer on the outputs of each |
| RNN layer except the last layer</li> |
| <li><strong>bidirectional</strong> (<em>bool, default False</em>) – If True, becomes a bidirectional RNN.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>input_size</strong> (<em>int, default 0</em>) – The number of expected features in the input x. |
| If not specified, it will be inferred from input.</li> |
| <li><strong>prefix</strong> (<em>str or None</em>) – Prefix of this <cite>Block</cite>.</li> |
| <li><strong>params</strong> (<em>ParameterDict or None</em>) – Shared Parameters for this <cite>Block</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(sequence_length, batch_size, input_size)</cite> |
| when <cite>layout</cite> is “TNC”. For other layouts dimensions are permuted accordingly.</li> |
| <li><strong>states</strong>: initial recurrent state tensor with shape |
| <cite>(num_layers, batch_size, num_hidden)</cite>. If <cite>bidirectional</cite> is True, |
| shape will instead be <cite>(2*num_layers, batch_size, num_hidden)</cite>. If |
| <cite>states</cite> is None, zeros will be used as default begin states.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(sequence_length, batch_size, num_hidden)</cite> |
| when <cite>layout</cite> is “TNC”. If <cite>bidirectional</cite> is True, output shape will instead |
| be <cite>(sequence_length, batch_size, 2*num_hidden)</cite></li> |
| <li><strong>out_states</strong>: output recurrent state tensor with the same shape as <cite>states</cite>. |
| If <cite>states</cite> is None <cite>out_states</cite> will not be returned.</li> |
| </ul> |
| </dd> |
| </dl> |
| <p class="rubric">Examples</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">layer</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">GRU</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="n">layer</span><span class="o">.</span><span class="n">initialize</span><span class="p">()</span> |
| <span class="gp">>>> </span><span class="nb">input</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">5</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="gp">>>> </span><span class="c1"># by default zeros are used as begin state</span> |
| <span class="gp">>>> </span><span class="n">output</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="c1"># manually specify begin state.</span> |
| <span class="gp">>>> </span><span class="n">h0</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span> |
| <span class="gp">>>> </span><span class="n">output</span><span class="p">,</span> <span class="n">hn</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span> |
| </pre></div> |
| </div> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.GRUCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">GRUCell</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>input_size=0</em>, <em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#GRUCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.GRUCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Gated Rectified Unit (GRU) network cell. |
| Note: this is an implementation of the cuDNN version of GRUs |
| (slight modification compared to Cho et al. 2014).</p> |
| <p>Each call computes the following function:</p> |
| <div class="math"> |
| \[\begin{split}\begin{array}{ll} |
| r_t = sigmoid(W_{ir} x_t + b_{ir} + W_{hr} h_{(t-1)} + b_{hr}) \\ |
| i_t = sigmoid(W_{ii} x_t + b_{ii} + W_hi h_{(t-1)} + b_{hi}) \\ |
| n_t = \tanh(W_{in} x_t + b_{in} + r_t * (W_{hn} h_{(t-1)}+ b_{hn})) \\ |
| h_t = (1 - i_t) * n_t + i_t * h_{(t-1)} \\ |
| \end{array}\end{split}\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, <span class="math">\(x_t\)</span> is the hidden |
| state of the previous layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer, |
| and <span class="math">\(r_t\)</span>, <span class="math">\(i_t\)</span>, <span class="math">\(n_t\)</span> are the reset, input, and new gates, respectively.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – Number of units in output symbol.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>prefix</strong> (str, default ‘<a href="#id3"><span class="problematic" id="id4">gru_</span></a>‘) – prefix for name of <cite>Block`s |
| (and name of weight if params is `None</cite>).</li> |
| <li><strong>params</strong> (<em>Parameter or None</em>) – Container for weight sharing between cells. |
| Created if <cite>None</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(batch_size, input_size)</cite>.</li> |
| <li><strong>states</strong>: a list of one initial recurrent state tensor with shape |
| <cite>(batch_size, num_hidden)</cite>.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(batch_size, num_hidden)</cite>.</li> |
| <li><strong>next_states</strong>: a list of one output recurrent state tensor with the |
| same shape as <cite>states</cite>.</li> |
| </ul> |
| </dd> |
| </dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.HybridRecurrentCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">HybridRecurrentCell</code><span class="sig-paren">(</span><em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#HybridRecurrentCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.HybridRecurrentCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>HybridRecurrentCell supports hybridize.</p> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.LSTM"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">LSTM</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>num_layers=1</em>, <em>layout='TNC'</em>, <em>dropout=0</em>, <em>bidirectional=False</em>, <em>input_size=0</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_layer.html#LSTM"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.LSTM" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Applies a multi-layer long short-term memory (LSTM) RNN to an input sequence.</p> |
| <p>For each element in the input sequence, each layer computes the following |
| function:</p> |
| <div class="math"> |
| \[\begin{split}\begin{array}{ll} |
| i_t = sigmoid(W_{ii} x_t + b_{ii} + W_{hi} h_{(t-1)} + b_{hi}) \\ |
| f_t = sigmoid(W_{if} x_t + b_{if} + W_{hf} h_{(t-1)} + b_{hf}) \\ |
| g_t = \tanh(W_{ig} x_t + b_{ig} + W_{hc} h_{(t-1)} + b_{hg}) \\ |
| o_t = sigmoid(W_{io} x_t + b_{io} + W_{ho} h_{(t-1)} + b_{ho}) \\ |
| c_t = f_t * c_{(t-1)} + i_t * g_t \\ |
| h_t = o_t * \tanh(c_t) |
| \end{array}\end{split}\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, <span class="math">\(c_t\)</span> is the |
| cell state at time <cite>t</cite>, <span class="math">\(x_t\)</span> is the hidden state of the previous |
| layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer, and <span class="math">\(i_t\)</span>, |
| <span class="math">\(f_t\)</span>, <span class="math">\(g_t\)</span>, <span class="math">\(o_t\)</span> are the input, forget, cell, and |
| out gates, respectively.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – The number of features in the hidden state h.</li> |
| <li><strong>num_layers</strong> (<em>int, default 1</em>) – Number of recurrent layers.</li> |
| <li><strong>layout</strong> (<em>str, default 'TNC'</em>) – The format of input and output tensors. T, N and C stand for |
| sequence length, batch size, and feature dimensions respectively.</li> |
| <li><strong>dropout</strong> (<em>float, default 0</em>) – If non-zero, introduces a dropout layer on the outputs of each |
| RNN layer except the last layer.</li> |
| <li><strong>bidirectional</strong> (<em>bool, default False</em>) – If <cite>True</cite>, becomes a bidirectional RNN.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer, default 'lstmbias'</em>) – Initializer for the bias vector. By default, bias for the forget |
| gate is initialized to 1 while all other biases are initialized |
| to zero.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>input_size</strong> (<em>int, default 0</em>) – The number of expected features in the input x. |
| If not specified, it will be inferred from input.</li> |
| <li><strong>prefix</strong> (<em>str or None</em>) – Prefix of this <cite>Block</cite>.</li> |
| <li><strong>params</strong> (<cite>ParameterDict</cite> or <cite>None</cite>) – Shared Parameters for this <cite>Block</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(sequence_length, batch_size, input_size)</cite> |
| when <cite>layout</cite> is “TNC”. For other layouts dimensions are permuted accordingly.</li> |
| <li><strong>states</strong>: a list of two initial recurrent state tensors. Each has shape |
| <cite>(num_layers, batch_size, num_hidden)</cite>. If <cite>bidirectional</cite> is True, |
| shape will instead be <cite>(2*num_layers, batch_size, num_hidden)</cite>. If |
| <cite>states</cite> is None, zeros will be used as default begin states.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(sequence_length, batch_size, num_hidden)</cite> |
| when <cite>layout</cite> is “TNC”. If <cite>bidirectional</cite> is True, output shape will instead |
| be <cite>(sequence_length, batch_size, 2*num_hidden)</cite></li> |
| <li><strong>out_states</strong>: a list of two output recurrent state tensors with the same |
| shape as in <cite>states</cite>. If <cite>states</cite> is None <cite>out_states</cite> will not be returned.</li> |
| </ul> |
| </dd> |
| </dl> |
| <p class="rubric">Examples</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">layer</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">LSTM</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="n">layer</span><span class="o">.</span><span class="n">initialize</span><span class="p">()</span> |
| <span class="gp">>>> </span><span class="nb">input</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">5</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="gp">>>> </span><span class="c1"># by default zeros are used as begin state</span> |
| <span class="gp">>>> </span><span class="n">output</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="c1"># manually specify begin state.</span> |
| <span class="gp">>>> </span><span class="n">h0</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span> |
| <span class="gp">>>> </span><span class="n">c0</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span> |
| <span class="gp">>>> </span><span class="n">output</span><span class="p">,</span> <span class="n">hn</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="p">[</span><span class="n">h0</span><span class="p">,</span> <span class="n">c0</span><span class="p">])</span> |
| </pre></div> |
| </div> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.LSTMCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">LSTMCell</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>input_size=0</em>, <em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#LSTMCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.LSTMCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Long-Short Term Memory (LSTM) network cell.</p> |
| <p>Each call computes the following function:</p> |
| <div class="math"> |
| \[\begin{split}\begin{array}{ll} |
| i_t = sigmoid(W_{ii} x_t + b_{ii} + W_{hi} h_{(t-1)} + b_{hi}) \\ |
| f_t = sigmoid(W_{if} x_t + b_{if} + W_{hf} h_{(t-1)} + b_{hf}) \\ |
| g_t = \tanh(W_{ig} x_t + b_{ig} + W_{hc} h_{(t-1)} + b_{hg}) \\ |
| o_t = sigmoid(W_{io} x_t + b_{io} + W_{ho} h_{(t-1)} + b_{ho}) \\ |
| c_t = f_t * c_{(t-1)} + i_t * g_t \\ |
| h_t = o_t * \tanh(c_t) |
| \end{array}\end{split}\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, <span class="math">\(c_t\)</span> is the |
| cell state at time <cite>t</cite>, <span class="math">\(x_t\)</span> is the hidden state of the previous |
| layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer, and <span class="math">\(i_t\)</span>, |
| <span class="math">\(f_t\)</span>, <span class="math">\(g_t\)</span>, <span class="math">\(o_t\)</span> are the input, forget, cell, and |
| out gates, respectively.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – Number of units in output symbol.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer, default 'lstmbias'</em>) – Initializer for the bias vector. By default, bias for the forget |
| gate is initialized to 1 while all other biases are initialized |
| to zero.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>prefix</strong> (str, default ‘<a href="#id5"><span class="problematic" id="id6">lstm_</span></a>‘) – Prefix for name of <cite>Block`s |
| (and name of weight if params is `None</cite>).</li> |
| <li><strong>params</strong> (<em>Parameter or None</em>) – Container for weight sharing between cells. |
| Created if <cite>None</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(batch_size, input_size)</cite>.</li> |
| <li><strong>states</strong>: a list of two initial recurrent state tensors. Each has shape |
| <cite>(batch_size, num_hidden)</cite>.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(batch_size, num_hidden)</cite>.</li> |
| <li><strong>next_states</strong>: a list of two output recurrent state tensors. Each has |
| the same shape as <cite>states</cite>.</li> |
| </ul> |
| </dd> |
| </dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.ModifierCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">ModifierCell</code><span class="sig-paren">(</span><em>base_cell</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#ModifierCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.ModifierCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Base class for modifier cells. A modifier |
| cell takes a base cell, apply modifications |
| on it (e.g. Zoneout), and returns a new cell.</p> |
| <p>After applying modifiers the base cell should |
| no longer be called directly. The modifier cell |
| should be used instead.</p> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.RNN"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">RNN</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>num_layers=1</em>, <em>activation='relu'</em>, <em>layout='TNC'</em>, <em>dropout=0</em>, <em>bidirectional=False</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>input_size=0</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_layer.html#RNN"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RNN" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Applies a multi-layer Elman RNN with <cite>tanh</cite> or <cite>ReLU</cite> non-linearity to an input sequence.</p> |
| <p>For each element in the input sequence, each layer computes the following |
| function:</p> |
| <div class="math"> |
| \[h_t = \tanh(w_{ih} * x_t + b_{ih} + w_{hh} * h_{(t-1)} + b_{hh})\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, and <span class="math">\(x_t\)</span> is the hidden |
| state of the previous layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer. |
| If nonlinearity=’relu’, then <cite>ReLU</cite> is used instead of <cite>tanh</cite>.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – The number of features in the hidden state h.</li> |
| <li><strong>num_layers</strong> (<em>int, default 1</em>) – Number of recurrent layers.</li> |
| <li><strong>activation</strong> (<em>{'relu' or 'tanh'}, default 'tanh'</em>) – The activation function to use.</li> |
| <li><strong>layout</strong> (<em>str, default 'TNC'</em>) – The format of input and output tensors. T, N and C stand for |
| sequence length, batch size, and feature dimensions respectively.</li> |
| <li><strong>dropout</strong> (<em>float, default 0</em>) – If non-zero, introduces a dropout layer on the outputs of each |
| RNN layer except the last layer.</li> |
| <li><strong>bidirectional</strong> (<em>bool, default False</em>) – If <cite>True</cite>, becomes a bidirectional RNN.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>input_size</strong> (<em>int, default 0</em>) – The number of expected features in the input x. |
| If not specified, it will be inferred from input.</li> |
| <li><strong>prefix</strong> (<em>str or None</em>) – Prefix of this <cite>Block</cite>.</li> |
| <li><strong>params</strong> (<em>ParameterDict or None</em>) – Shared Parameters for this <cite>Block</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(sequence_length, batch_size, input_size)</cite> |
| when <cite>layout</cite> is “TNC”. For other layouts dimensions are permuted accordingly.</li> |
| <li><strong>states</strong>: initial recurrent state tensor with shape |
| <cite>(num_layers, batch_size, num_hidden)</cite>. If <cite>bidirectional</cite> is True, |
| shape will instead be <cite>(2*num_layers, batch_size, num_hidden)</cite>. If |
| <cite>states</cite> is None, zeros will be used as default begin states.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(sequence_length, batch_size, num_hidden)</cite> |
| when <cite>layout</cite> is “TNC”. If <cite>bidirectional</cite> is True, output shape will instead |
| be <cite>(sequence_length, batch_size, 2*num_hidden)</cite></li> |
| <li><strong>out_states</strong>: output recurrent state tensor with the same shape as <cite>states</cite>. |
| If <cite>states</cite> is None <cite>out_states</cite> will not be returned.</li> |
| </ul> |
| </dd> |
| </dl> |
| <p class="rubric">Examples</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">layer</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">gluon</span><span class="o">.</span><span class="n">rnn</span><span class="o">.</span><span class="n">RNN</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="n">layer</span><span class="o">.</span><span class="n">initialize</span><span class="p">()</span> |
| <span class="gp">>>> </span><span class="nb">input</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">5</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="gp">>>> </span><span class="c1"># by default zeros are used as begin state</span> |
| <span class="gp">>>> </span><span class="n">output</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">)</span> |
| <span class="gp">>>> </span><span class="c1"># manually specify begin state.</span> |
| <span class="gp">>>> </span><span class="n">h0</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">100</span><span class="p">))</span> |
| <span class="gp">>>> </span><span class="n">output</span><span class="p">,</span> <span class="n">hn</span> <span class="o">=</span> <span class="n">layer</span><span class="p">(</span><span class="nb">input</span><span class="p">,</span> <span class="n">h0</span><span class="p">)</span> |
| </pre></div> |
| </div> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.RNNCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">RNNCell</code><span class="sig-paren">(</span><em>hidden_size</em>, <em>activation='tanh'</em>, <em>i2h_weight_initializer=None</em>, <em>h2h_weight_initializer=None</em>, <em>i2h_bias_initializer='zeros'</em>, <em>h2h_bias_initializer='zeros'</em>, <em>input_size=0</em>, <em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RNNCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RNNCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Elman RNN recurrent neural network cell.</p> |
| <p>Each call computes the following function:</p> |
| <div class="math"> |
| \[h_t = \tanh(w_{ih} * x_t + b_{ih} + w_{hh} * h_{(t-1)} + b_{hh})\]</div> |
| <p>where <span class="math">\(h_t\)</span> is the hidden state at time <cite>t</cite>, and <span class="math">\(x_t\)</span> is the hidden |
| state of the previous layer at time <cite>t</cite> or <span class="math">\(input_t\)</span> for the first layer. |
| If nonlinearity=’relu’, then <cite>ReLU</cite> is used instead of <cite>tanh</cite>.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>hidden_size</strong> (<em>int</em>) – Number of units in output symbol</li> |
| <li><strong>activation</strong> (<em>str or Symbol, default 'tanh'</em>) – Type of activation function.</li> |
| <li><strong>i2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the input weights matrix, used for the linear |
| transformation of the inputs.</li> |
| <li><strong>h2h_weight_initializer</strong> (<em>str or Initializer</em>) – Initializer for the recurrent weights matrix, used for the linear |
| transformation of the recurrent state.</li> |
| <li><strong>i2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>h2h_bias_initializer</strong> (<em>str or Initializer</em>) – Initializer for the bias vector.</li> |
| <li><strong>prefix</strong> (str, default ‘<a href="#id7"><span class="problematic" id="id8">rnn_</span></a>‘) – Prefix for name of <cite>Block`s |
| (and name of weight if params is `None</cite>).</li> |
| <li><strong>params</strong> (<em>Parameter or None</em>) – Container for weight sharing between cells. |
| Created if <cite>None</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="docutils"> |
| <dt>Inputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>data</strong>: input tensor with shape <cite>(batch_size, input_size)</cite>.</li> |
| <li><strong>states</strong>: a list of one initial recurrent state tensor with shape |
| <cite>(batch_size, num_hidden)</cite>.</li> |
| </ul> |
| </dd> |
| <dt>Outputs:</dt> |
| <dd><ul class="first last simple"> |
| <li><strong>out</strong>: output tensor with shape <cite>(batch_size, num_hidden)</cite>.</li> |
| <li><strong>next_states</strong>: a list of one output recurrent state tensor with the |
| same shape as <cite>states</cite>.</li> |
| </ul> |
| </dd> |
| </dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">RecurrentCell</code><span class="sig-paren">(</span><em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Abstract base class for RNN cells</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first last simple"> |
| <li><strong>prefix</strong> (<em>str, optional</em>) – Prefix for names of <cite>Block`s |
| (this prefix is also used for names of weights if `params</cite> is <cite>None</cite> |
| i.e. if <cite>params</cite> are being created and not reused)</li> |
| <li><strong>params</strong> (<em>Parameter or None, optional</em>) – Container for weight sharing between cells. |
| A new Parameter container is created if <cite>params</cite> is <cite>None</cite>.</li> |
| </ul> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell.begin_state"> |
| <code class="descname">begin_state</code><span class="sig-paren">(</span><em>batch_size=0</em>, <em>func=<function zeros></em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell.begin_state"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell.begin_state" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Initial state for this cell.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first simple"> |
| <li><strong>func</strong> (<em>callable, default symbol.zeros</em>) – <p>Function for creating initial state.</p> |
| <p>For Symbol API, func can be <cite>symbol.zeros</cite>, <cite>symbol.uniform</cite>, |
| <cite>symbol.var etc</cite>. Use <cite>symbol.var</cite> if you want to directly |
| feed input as states.</p> |
| <p>For NDArray API, func can be <cite>ndarray.zeros</cite>, <cite>ndarray.ones</cite>, etc.</p> |
| </li> |
| <li><strong>batch_size</strong> (<em>int, default 0</em>) – Only required for NDArray API. Size of the batch (‘N’ in layout) |
| dimension of input.</li> |
| <li><strong>**kwargs</strong> – <p>Additional keyword arguments passed to func. For example |
| <cite>mean</cite>, <cite>std</cite>, <cite>dtype</cite>, etc.</p> |
| </li> |
| </ul> |
| </td> |
| </tr> |
| <tr class="field-even field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>states</strong> – |
| Starting states for the first RNN step.</p> |
| </td> |
| </tr> |
| <tr class="field-odd field"><th class="field-name">Return type:</th><td class="field-body"><p class="first last">nested list of Symbol</p> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| </dd></dl> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell.forward"> |
| <code class="descname">forward</code><span class="sig-paren">(</span><em>inputs</em>, <em>states</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell.forward" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Unrolls the recurrent cell for one time step.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first simple"> |
| <li><strong>inputs</strong> (<em>sym.Variable</em>) – Input symbol, 2D, of shape (batch_size * num_units).</li> |
| <li><strong>states</strong> (<em>list of sym.Variable</em>) – RNN state from previous step or the output of begin_state().</li> |
| </ul> |
| </td> |
| </tr> |
| <tr class="field-even field"><th class="field-name">Returns:</th><td class="field-body"><p class="first last"><ul class="simple"> |
| <li><strong>output</strong> (<em>Symbol</em>) – |
| Symbol corresponding to the output from the RNN when unrolling |
| for a single time step.</li> |
| <li><strong>states</strong> (<em>list of Symbol</em>) – |
| The new state of this RNN after this unrolling. |
| The type of this symbol is same as the output of <cite>begin_state()</cite>. |
| This can be used as an input state to the next time step |
| of this RNN.</li> |
| </ul> |
| </p> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| <div class="admonition seealso"> |
| <p class="first admonition-title">See also</p> |
| <dl class="last docutils"> |
| <dt><a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell.begin_state" title="mxnet.gluon.rnn.RecurrentCell.begin_state"><code class="xref py py-meth docutils literal"><span class="pre">begin_state()</span></code></a></dt> |
| <dd>This function can provide the states for the first time step.</dd> |
| <dt><a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell.unroll" title="mxnet.gluon.rnn.RecurrentCell.unroll"><code class="xref py py-meth docutils literal"><span class="pre">unroll()</span></code></a></dt> |
| <dd>This function unrolls an RNN for a given number of (>=1) time steps.</dd> |
| </dl> |
| </div> |
| </dd></dl> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell.reset"> |
| <code class="descname">reset</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell.reset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell.reset" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Reset before re-using the cell for another graph.</p> |
| </dd></dl> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell.state_info"> |
| <code class="descname">state_info</code><span class="sig-paren">(</span><em>batch_size=0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell.state_info"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell.state_info" title="Permalink to this definition">¶</a></dt> |
| <dd><p>shape and layout information of states</p> |
| </dd></dl> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.RecurrentCell.unroll"> |
| <code class="descname">unroll</code><span class="sig-paren">(</span><em>length</em>, <em>inputs</em>, <em>begin_state=None</em>, <em>layout='NTC'</em>, <em>merge_outputs=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#RecurrentCell.unroll"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.RecurrentCell.unroll" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Unrolls an RNN cell across time steps.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><ul class="first simple"> |
| <li><strong>length</strong> (<em>int</em>) – Number of steps to unroll.</li> |
| <li><strong>inputs</strong> (<em>Symbol, list of Symbol, or None</em>) – <p>If <cite>inputs</cite> is a single Symbol (usually the output |
| of Embedding symbol), it should have shape |
| (batch_size, length, ...) if <cite>layout</cite> is ‘NTC’, |
| or (length, batch_size, ...) if <cite>layout</cite> is ‘TNC’.</p> |
| <p>If <cite>inputs</cite> is a list of symbols (usually output of |
| previous unroll), they should all have shape |
| (batch_size, ...).</p> |
| </li> |
| <li><strong>begin_state</strong> (<em>nested list of Symbol, optional</em>) – Input states created by <cite>begin_state()</cite> |
| or output state of another cell. |
| Created from <cite>begin_state()</cite> if <cite>None</cite>.</li> |
| <li><strong>layout</strong> (<em>str, optional</em>) – <cite>layout</cite> of input symbol. Only used if inputs |
| is a single Symbol.</li> |
| <li><strong>merge_outputs</strong> (<em>bool, optional</em>) – If <cite>False</cite>, returns outputs as a list of Symbols. |
| If <cite>True</cite>, concatenates output across time steps |
| and returns a single symbol with shape |
| (batch_size, length, ...) if layout is ‘NTC’, |
| or (length, batch_size, ...) if layout is ‘TNC’. |
| If <cite>None</cite>, output whatever is faster.</li> |
| </ul> |
| </td> |
| </tr> |
| <tr class="field-even field"><th class="field-name">Returns:</th><td class="field-body"><p class="first last"><ul class="simple"> |
| <li><strong>outputs</strong> (<em>list of Symbol or Symbol</em>) – |
| Symbol (if <cite>merge_outputs</cite> is True) or list of Symbols |
| (if <cite>merge_outputs</cite> is False) corresponding to the output from |
| the RNN from this unrolling.</li> |
| <li><strong>states</strong> (<em>list of Symbol</em>) – |
| The new state of this RNN after this unrolling. |
| The type of this symbol is same as the output of <cite>begin_state()</cite>.</li> |
| </ul> |
| </p> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| </dd></dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.ResidualCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">ResidualCell</code><span class="sig-paren">(</span><em>base_cell</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#ResidualCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.ResidualCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Adds residual connection as described in Wu et al, 2016 |
| (<a class="reference external" href="https://arxiv.org/abs/1609.08144">https://arxiv.org/abs/1609.08144</a>). |
| Output of the cell is output of the base cell plus input.</p> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.SequentialRNNCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">SequentialRNNCell</code><span class="sig-paren">(</span><em>prefix=None</em>, <em>params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#SequentialRNNCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.SequentialRNNCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Sequentially stacking multiple RNN cells.</p> |
| <dl class="method"> |
| <dt id="mxnet.gluon.rnn.SequentialRNNCell.add"> |
| <code class="descname">add</code><span class="sig-paren">(</span><em>cell</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#SequentialRNNCell.add"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.SequentialRNNCell.add" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Appends a cell into the stack.</p> |
| <table class="docutils field-list" frame="void" rules="none"> |
| <col class="field-name"/> |
| <col class="field-body"/> |
| <tbody valign="top"> |
| <tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><strong>cell</strong> (<a class="reference internal" href="#mxnet.gluon.rnn.RecurrentCell" title="mxnet.gluon.rnn.RecurrentCell"><em>RecurrentCell</em></a>) – The cell to add.</td> |
| </tr> |
| </tbody> |
| </table> |
| </dd></dl> |
| </dd></dl> |
| <dl class="class"> |
| <dt id="mxnet.gluon.rnn.ZoneoutCell"> |
| <em class="property">class </em><code class="descclassname">mxnet.gluon.rnn.</code><code class="descname">ZoneoutCell</code><span class="sig-paren">(</span><em>base_cell</em>, <em>zoneout_outputs=0.0</em>, <em>zoneout_states=0.0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/rnn/rnn_cell.html#ZoneoutCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.rnn.ZoneoutCell" title="Permalink to this definition">¶</a></dt> |
| <dd><p>Applies Zoneout on base cell.</p> |
| </dd></dl> |
| <script>auto_index("api-reference");</script></div> |
| </div> |
| </div> |
| </div> |
| <div aria-label="main navigation" class="sphinxsidebar rightsidebar" role="navigation"> |
| <div class="sphinxsidebarwrapper"> |
| <h3><a href="../../../index.html">Table Of Contents</a></h3> |
| <ul> |
| <li><a class="reference internal" href="#">Gluon Recurrent Neural Network API</a><ul> |
| <li><a class="reference internal" href="#overview">Overview</a><ul> |
| <li><a class="reference internal" href="#recurrent-layers">Recurrent Layers</a></li> |
| <li><a class="reference internal" href="#recurrent-cells">Recurrent Cells</a></li> |
| </ul> |
| </li> |
| <li><a class="reference internal" href="#api-reference">API Reference</a></li> |
| </ul> |
| </li> |
| </ul> |
| </div> |
| </div> |
| </div><div class="footer"> |
| <div class="section-disclaimer"> |
| <div class="container"> |
| <div> |
| <img height="60" src="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/apache_incubator_logo.png"/> |
| <p> |
| Apache MXNet is an effort undergoing incubation at The Apache Software Foundation (ASF), <strong>sponsored by the <i>Apache Incubator</i></strong>. Incubation is required of all newly accepted projects until a further review indicates that the infrastructure, communications, and decision making process have stabilized in a manner consistent with other successful ASF projects. While incubation status is not necessarily a reflection of the completeness or stability of the code, it does indicate that the project has yet to be fully endorsed by the ASF. |
| </p> |
| <p> |
| "Copyright © 2017-2018, The Apache Software Foundation |
| Apache MXNet, MXNet, Apache, the Apache feather, and the Apache MXNet project logo are either registered trademarks or trademarks of the Apache Software Foundation." |
| </p> |
| </div> |
| </div> |
| </div> |
| </div> <!-- pagename != index --> |
| </div> |
| <script crossorigin="anonymous" integrity="sha384-0mSbJDEHialfmuBBQP6A4Qrprq5OVfW37PRR3j5ELqxss1yVqOtnepnHVP9aJ7xS" src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/js/bootstrap.min.js"></script> |
| <script src="../../../_static/js/sidebar.js" type="text/javascript"></script> |
| <script src="../../../_static/js/search.js" type="text/javascript"></script> |
| <script src="../../../_static/js/navbar.js" type="text/javascript"></script> |
| <script src="../../../_static/js/clipboard.min.js" type="text/javascript"></script> |
| <script src="../../../_static/js/copycode.js" type="text/javascript"></script> |
| <script src="../../../_static/js/page.js" type="text/javascript"></script> |
| <script type="text/javascript"> |
| $('body').ready(function () { |
| $('body').css('visibility', 'visible'); |
| }); |
| </script> |
| </body> |
| </html> |
