| <!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>Char RNN Example — 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="../../_static/jquery-1.11.1.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://cdn.mathjax.org/mathjax/latest/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="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet-icon.png" rel="icon" type="image/png"/> |
| </link></link></head> |
| <body role="document"><!-- Previous Navbar Layout |
| <div class="navbar navbar-default navbar-fixed-top"> |
| <div class="container"> |
| <div class="navbar-header"> |
| <button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#navbar" aria-expanded="false" aria-controls="navbar"> |
| <span class="sr-only">Toggle navigation</span> |
| <span class="icon-bar"></span> |
| <span class="icon-bar"></span> |
| <span class="icon-bar"></span> |
| </button> |
| <a href="../../" class="navbar-brand"> |
| <img src="http://data.mxnet.io/theme/mxnet.png"> |
| </a> |
| </div> |
| <div id="navbar" class="navbar-collapse collapse"> |
| <ul id="navbar" class="navbar navbar-left"> |
| |
| <li> <a href="../../get_started/index.html">Get Started</a> </li> |
| |
| <li> <a href="../../tutorials/index.html">Tutorials</a> </li> |
| |
| <li> <a href="../../how_to/index.html">How To</a> </li> |
| |
| |
| <li class="dropdown"> |
| <a href="#" class="dropdown-toggle" data-toggle="dropdown" role="button" aria-haspopup="true" aria-expanded="true">Packages <span class="caret"></span></a> |
| <ul class="dropdown-menu"> |
| |
| <li><a href="../../packages/python/index.html"> |
| Python |
| </a></li> |
| |
| <li><a href="../../packages/r/index.html"> |
| R |
| </a></li> |
| |
| <li><a href="../../packages/julia/index.html"> |
| Julia |
| </a></li> |
| |
| <li><a href="../../packages/c++/index.html"> |
| C++ |
| </a></li> |
| |
| <li><a href="../../packages/scala/index.html"> |
| Scala |
| </a></li> |
| |
| <li><a href="../../packages/perl/index.html"> |
| Perl |
| </a></li> |
| |
| </ul> |
| </li> |
| |
| <li> <a href="../../system/index.html">System</a> </li> |
| <li> |
| <form class="" role="search" action="../../search.html" method="get" autocomplete="off"> |
| <div class="form-group inner-addon left-addon"> |
| <i class="glyphicon glyphicon-search"></i> |
| <input type="text" name="q" class="form-control" placeholder="Search"> |
| </div> |
| <input type="hidden" name="check_keywords" value="yes" /> |
| <input type="hidden" name="area" value="default" /> |
| |
| </form> </li> |
| </ul> |
| <ul id="navbar" class="navbar navbar-right"> |
| <li> <a href="../../index.html"><span class="flag-icon flag-icon-us"></span></a> </li> |
| <li> <a href="../..//zh/index.html"><span class="flag-icon flag-icon-cn"></span></a> </li> |
| </ul> |
| </div> |
| </div> |
| </div> |
| Previous Navbar Layout End --> |
| <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="http://data.mxnet.io/theme/mxnet.png"/></a> |
| </h1> |
| <nav class="nav-bar" id="main-nav"> |
| <a class="main-nav-link" href="../../get_started/install.html">Install</a> |
| <a class="main-nav-link" href="../../tutorials/index.html">Tutorials</a> |
| <a class="main-nav-link" href="../../how_to/index.html">How To</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">API <span class="caret"></span></a> |
| <ul class="dropdown-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> |
| <a class="main-nav-link" href="../../architecture/index.html">Architecture</a> |
| <!-- <a class="main-nav-link" href="../../community/index.html">Community</a> --> |
| <a class="main-nav-link" href="https://github.com/dmlc/mxnet">Github</a> |
| <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=http://mxnet.incubator.apache.org/test/>v0.10.14</a></li><li><a class="main-nav-link" href=http://mxnet.incubator.apache.org/test/versions/0.10/index.html>0.10</a></li><li><a class="main-nav-link" href=http://mxnet.incubator.apache.org/test/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 dropdown-menu-right" id="burgerMenu"> |
| <li><a href="../../get_started/install.html">Install</a></li> |
| <li><a href="../../tutorials/index.html">Tutorials</a></li> |
| <li><a href="../../how_to/index.html">How To</a></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><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=http://mxnet.incubator.apache.org/test/>v0.10.14</a></li><li><a tabindex="-1" href=http://mxnet.incubator.apache.org/test/versions/0.10/index.html>0.10</a></li><li><a tabindex="-1" href=http://mxnet.incubator.apache.org/test/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> |
| <div class="container"> |
| <div class="row"> |
| <div aria-label="main navigation" class="sphinxsidebar leftsidebar" role="navigation"> |
| <div class="sphinxsidebarwrapper"> |
| <ul> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/python/index.html">Python Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/r/index.html">R Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/julia/index.html">Julia Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/c++/index.html">C++ Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/scala/index.html">Scala Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../api/perl/index.html">Perl Documents</a></li> |
| <li class="toctree-l1"><a class="reference internal" href="../../how_to/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="../index.html">Tutorials</a></li> |
| </ul> |
| </div> |
| </div> |
| <div class="content"> |
| <div class="section" id="char-rnn-example"> |
| <span id="char-rnn-example"></span><h1>Char RNN Example<a class="headerlink" href="#char-rnn-example" title="Permalink to this headline">¶</a></h1> |
| <p>This tutorial shows how to use an LSTM model to build a char-level language model, and generate text from it. For demonstration purposes, we use a Shakespearean text. You can find the data on <a class="reference external" href="https://github.com/dmlc/web-data/tree/master/mxnet/tinyshakespeare">GitHub</a>.</p> |
| <div class="section" id="load-the-data"> |
| <span id="load-the-data"></span><h2>Load the Data<a class="headerlink" href="#load-the-data" title="Permalink to this headline">¶</a></h2> |
| <p>Load in the data and preprocess it:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> <span class="kn">require</span><span class="p">(</span>mxnet<span class="p">)</span> |
| </pre></div> |
| </div> |
| <div class="highlight-python"><div class="highlight"><pre><span></span> <span class="c1">## Loading required package: mxnet</span> |
| </pre></div> |
| </div> |
| <div class="highlight-python"><div class="highlight"><pre><span></span> <span class="c1">## Loading required package: methods</span> |
| </pre></div> |
| </div> |
| <p>Set the basic network parameters:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> batch.size <span class="o">=</span> <span class="m">32</span> |
| seq.len <span class="o">=</span> <span class="m">32</span> |
| num.hidden <span class="o">=</span> <span class="m">16</span> |
| num.embed <span class="o">=</span> <span class="m">16</span> |
| num.lstm.layer <span class="o">=</span> <span class="m">1</span> |
| num.round <span class="o">=</span> <span class="m">1</span> |
| learning.rate<span class="o">=</span> <span class="m">0.1</span> |
| wd<span class="o">=</span><span class="m">0.00001</span> |
| clip_gradient<span class="o">=</span><span class="m">1</span> |
| update.period <span class="o">=</span> <span class="m">1</span> |
| </pre></div> |
| </div> |
| <p>Download the data:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> download.data <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>data_dir<span class="p">)</span> <span class="p">{</span> |
| <span class="kp">dir.create</span><span class="p">(</span>data_dir<span class="p">,</span> showWarnings <span class="o">=</span> <span class="kc">FALSE</span><span class="p">)</span> |
| <span class="kr">if</span> <span class="p">(</span><span class="o">!</span><span class="kp">file.exists</span><span class="p">(</span><span class="kp">paste0</span><span class="p">(</span>data_dir<span class="p">,</span><span class="s">'input.txt'</span><span class="p">)))</span> <span class="p">{</span> |
| download.file<span class="p">(</span>url<span class="o">=</span><span class="s">'https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/tinyshakespeare/input.txt'</span><span class="p">,</span> |
| destfile<span class="o">=</span><span class="kp">paste0</span><span class="p">(</span>data_dir<span class="p">,</span><span class="s">'input.txt'</span><span class="p">),</span> method<span class="o">=</span><span class="s">'wget'</span><span class="p">)</span> |
| <span class="p">}</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Make a dictionary from the text:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> make.dict <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>text<span class="p">,</span> max.vocab<span class="o">=</span><span class="m">10000</span><span class="p">)</span> <span class="p">{</span> |
| text <span class="o"><-</span> <span class="kp">strsplit</span><span class="p">(</span>text<span class="p">,</span> <span class="s">''</span><span class="p">)</span> |
| dic <span class="o"><-</span> <span class="kt">list</span><span class="p">()</span> |
| idx <span class="o"><-</span> <span class="m">1</span> |
| <span class="kr">for</span> <span class="p">(</span><span class="kt">c</span> <span class="kr">in</span> text<span class="p">[[</span><span class="m">1</span><span class="p">]])</span> <span class="p">{</span> |
| <span class="kr">if</span> <span class="p">(</span><span class="o">!</span><span class="p">(</span><span class="kt">c</span> <span class="o">%in%</span> <span class="kp">names</span><span class="p">(</span>dic<span class="p">)))</span> <span class="p">{</span> |
| dic<span class="p">[[</span><span class="kt">c</span><span class="p">]]</span> <span class="o"><-</span> idx |
| idx <span class="o"><-</span> idx <span class="o">+</span> <span class="m">1</span> |
| <span class="p">}</span> |
| <span class="p">}</span> |
| <span class="kr">if</span> <span class="p">(</span><span class="kp">length</span><span class="p">(</span>dic<span class="p">)</span> <span class="o">==</span> max.vocab <span class="o">-</span> <span class="m">1</span><span class="p">)</span> |
| dic<span class="p">[[</span><span class="s">"UNKNOWN"</span><span class="p">]]</span> <span class="o"><-</span> idx |
| <span class="kp">cat</span><span class="p">(</span><span class="kp">paste0</span><span class="p">(</span><span class="s">"Total unique char: "</span><span class="p">,</span> <span class="kp">length</span><span class="p">(</span>dic<span class="p">),</span> <span class="s">"\n"</span><span class="p">))</span> |
| <span class="kr">return</span> <span class="p">(</span>dic<span class="p">)</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Transfer the text into a data feature:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> make.data <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span><span class="kp">file.path</span><span class="p">,</span> seq.len<span class="o">=</span><span class="m">32</span><span class="p">,</span> max.vocab<span class="o">=</span><span class="m">10000</span><span class="p">,</span> dic<span class="o">=</span><span class="kc">NULL</span><span class="p">)</span> <span class="p">{</span> |
| fi <span class="o"><-</span> <span class="kp">file</span><span class="p">(</span><span class="kp">file.path</span><span class="p">,</span> <span class="s">"r"</span><span class="p">)</span> |
| text <span class="o"><-</span> <span class="kp">paste</span><span class="p">(</span><span class="kp">readLines</span><span class="p">(</span>fi<span class="p">),</span> collapse<span class="o">=</span><span class="s">"\n"</span><span class="p">)</span> |
| <span class="kp">close</span><span class="p">(</span>fi<span class="p">)</span> |
| |
| <span class="kr">if</span> <span class="p">(</span><span class="kp">is.null</span><span class="p">(</span>dic<span class="p">))</span> |
| dic <span class="o"><-</span> make.dict<span class="p">(</span>text<span class="p">,</span> max.vocab<span class="p">)</span> |
| lookup.table <span class="o"><-</span> <span class="kt">list</span><span class="p">()</span> |
| <span class="kr">for</span> <span class="p">(</span><span class="kt">c</span> <span class="kr">in</span> <span class="kp">names</span><span class="p">(</span>dic<span class="p">))</span> <span class="p">{</span> |
| idx <span class="o"><-</span> dic<span class="p">[[</span><span class="kt">c</span><span class="p">]]</span> |
| lookup.table<span class="p">[[</span>idx<span class="p">]]</span> <span class="o"><-</span> <span class="kt">c</span> |
| <span class="p">}</span> |
| |
| char.lst <span class="o"><-</span> <span class="kp">strsplit</span><span class="p">(</span>text<span class="p">,</span> <span class="s">''</span><span class="p">)[[</span><span class="m">1</span><span class="p">]]</span> |
| num.seq <span class="o"><-</span> <span class="kp">as.integer</span><span class="p">(</span><span class="kp">length</span><span class="p">(</span>char.lst<span class="p">)</span> <span class="o">/</span> seq.len<span class="p">)</span> |
| char.lst <span class="o"><-</span> char.lst<span class="p">[</span><span class="m">1</span><span class="o">:</span><span class="p">(</span>num.seq <span class="o">*</span> seq.len<span class="p">)]</span> |
| data <span class="o"><-</span> <span class="kt">array</span><span class="p">(</span><span class="m">0</span><span class="p">,</span> dim<span class="o">=</span><span class="kt">c</span><span class="p">(</span>seq.len<span class="p">,</span> num.seq<span class="p">))</span> |
| idx <span class="o"><-</span> <span class="m">1</span> |
| <span class="kr">for</span> <span class="p">(</span>i <span class="kr">in</span> <span class="m">1</span><span class="o">:</span>num.seq<span class="p">)</span> <span class="p">{</span> |
| <span class="kr">for</span> <span class="p">(</span>j <span class="kr">in</span> <span class="m">1</span><span class="o">:</span>seq.len<span class="p">)</span> <span class="p">{</span> |
| <span class="kr">if</span> <span class="p">(</span>char.lst<span class="p">[</span>idx<span class="p">]</span> <span class="o">%in%</span> <span class="kp">names</span><span class="p">(</span>dic<span class="p">))</span> |
| data<span class="p">[</span>j<span class="p">,</span> i<span class="p">]</span> <span class="o"><-</span> dic<span class="p">[[</span> char.lst<span class="p">[</span>idx<span class="p">]</span> <span class="p">]]</span><span class="m">-1</span> |
| <span class="kr">else</span> <span class="p">{</span> |
| data<span class="p">[</span>j<span class="p">,</span> i<span class="p">]</span> <span class="o"><-</span> dic<span class="p">[[</span><span class="s">"UNKNOWN"</span><span class="p">]]</span><span class="m">-1</span> |
| <span class="p">}</span> |
| idx <span class="o"><-</span> idx <span class="o">+</span> <span class="m">1</span> |
| <span class="p">}</span> |
| <span class="p">}</span> |
| <span class="kr">return</span> <span class="p">(</span><span class="kt">list</span><span class="p">(</span>data<span class="o">=</span>data<span class="p">,</span> dic<span class="o">=</span>dic<span class="p">,</span> lookup.table<span class="o">=</span>lookup.table<span class="p">))</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Move the tail text:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> drop.tail <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>X<span class="p">,</span> batch.size<span class="p">)</span> <span class="p">{</span> |
| shape <span class="o"><-</span> <span class="kp">dim</span><span class="p">(</span>X<span class="p">)</span> |
| nstep <span class="o"><-</span> <span class="kp">as.integer</span><span class="p">(</span>shape<span class="p">[</span><span class="m">2</span><span class="p">]</span> <span class="o">/</span> batch.size<span class="p">)</span> |
| <span class="kr">return</span> <span class="p">(</span>X<span class="p">[,</span> <span class="m">1</span><span class="o">:</span><span class="p">(</span>nstep <span class="o">*</span> batch.size<span class="p">)])</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Get the label of X:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> get.label <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>X<span class="p">)</span> <span class="p">{</span> |
| label <span class="o"><-</span> <span class="kt">array</span><span class="p">(</span><span class="m">0</span><span class="p">,</span> dim<span class="o">=</span><span class="kp">dim</span><span class="p">(</span>X<span class="p">))</span> |
| d <span class="o"><-</span> <span class="kp">dim</span><span class="p">(</span>X<span class="p">)[</span><span class="m">1</span><span class="p">]</span> |
| w <span class="o"><-</span> <span class="kp">dim</span><span class="p">(</span>X<span class="p">)[</span><span class="m">2</span><span class="p">]</span> |
| <span class="kr">for</span> <span class="p">(</span>i <span class="kr">in</span> <span class="m">0</span><span class="o">:</span><span class="p">(</span>w<span class="m">-1</span><span class="p">))</span> <span class="p">{</span> |
| <span class="kr">for</span> <span class="p">(</span>j <span class="kr">in</span> <span class="m">1</span><span class="o">:</span>d<span class="p">)</span> <span class="p">{</span> |
| label<span class="p">[</span>i<span class="o">*</span>d<span class="o">+</span>j<span class="p">]</span> <span class="o"><-</span> X<span class="p">[(</span>i<span class="o">*</span>d<span class="o">+</span>j<span class="p">)</span><span class="o">%%</span><span class="p">(</span>w<span class="o">*</span>d<span class="p">)</span><span class="m">+1</span><span class="p">]</span> |
| <span class="p">}</span> |
| <span class="p">}</span> |
| <span class="kr">return</span> <span class="p">(</span>label<span class="p">)</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Get the training data and evaluation data:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> download.data<span class="p">(</span><span class="s">"./data/"</span><span class="p">)</span> |
| ret <span class="o"><-</span> make.data<span class="p">(</span><span class="s">"./data/input.txt"</span><span class="p">,</span> seq.len<span class="o">=</span>seq.len<span class="p">)</span> |
| </pre></div> |
| </div> |
| <div class="highlight-python"><div class="highlight"><pre><span></span> <span class="c1">## Total unique char: 65</span> |
| </pre></div> |
| </div> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> X <span class="o"><-</span> ret<span class="o">$</span>data |
| dic <span class="o"><-</span> ret<span class="o">$</span>dic |
| lookup.table <span class="o"><-</span> ret<span class="o">$</span>lookup.table |
| |
| vocab <span class="o"><-</span> <span class="kp">length</span><span class="p">(</span>dic<span class="p">)</span> |
| |
| shape <span class="o"><-</span> <span class="kp">dim</span><span class="p">(</span>X<span class="p">)</span> |
| train.val.fraction <span class="o"><-</span> <span class="m">0.9</span> |
| size <span class="o"><-</span> shape<span class="p">[</span><span class="m">2</span><span class="p">]</span> |
| |
| X.train.data <span class="o"><-</span> X<span class="p">[,</span> <span class="m">1</span><span class="o">:</span><span class="kp">as.integer</span><span class="p">(</span>size <span class="o">*</span> train.val.fraction<span class="p">)]</span> |
| X.val.data <span class="o"><-</span> X<span class="p">[,</span> <span class="o">-</span><span class="p">(</span><span class="m">1</span><span class="o">:</span><span class="kp">as.integer</span><span class="p">(</span>size <span class="o">*</span> train.val.fraction<span class="p">))]</span> |
| X.train.data <span class="o"><-</span> drop.tail<span class="p">(</span>X.train.data<span class="p">,</span> batch.size<span class="p">)</span> |
| X.val.data <span class="o"><-</span> drop.tail<span class="p">(</span>X.val.data<span class="p">,</span> batch.size<span class="p">)</span> |
| |
| X.train.label <span class="o"><-</span> get.label<span class="p">(</span>X.train.data<span class="p">)</span> |
| X.val.label <span class="o"><-</span> get.label<span class="p">(</span>X.val.data<span class="p">)</span> |
| |
| X.train <span class="o"><-</span> <span class="kt">list</span><span class="p">(</span>data<span class="o">=</span>X.train.data<span class="p">,</span> label<span class="o">=</span>X.train.label<span class="p">)</span> |
| X.val <span class="o"><-</span> <span class="kt">list</span><span class="p">(</span>data<span class="o">=</span>X.val.data<span class="p">,</span> label<span class="o">=</span>X.val.label<span class="p">)</span> |
| </pre></div> |
| </div> |
| </div> |
| <div class="section" id="train-the-model"> |
| <span id="train-the-model"></span><h2>Train the Model<a class="headerlink" href="#train-the-model" title="Permalink to this headline">¶</a></h2> |
| <p>In <code class="docutils literal"><span class="pre">mxnet</span></code>, we have a function called <code class="docutils literal"><span class="pre">mx.lstm</span></code> so that users can build a general LSTM model:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> model <span class="o"><-</span> mx.lstm<span class="p">(</span>X.train<span class="p">,</span> X.val<span class="p">,</span> |
| ctx<span class="o">=</span>mx.cpu<span class="p">(),</span> |
| num.round<span class="o">=</span>num.round<span class="p">,</span> |
| update.period<span class="o">=</span>update.period<span class="p">,</span> |
| num.lstm.layer<span class="o">=</span>num.lstm.layer<span class="p">,</span> |
| seq.len<span class="o">=</span>seq.len<span class="p">,</span> |
| num.hidden<span class="o">=</span>num.hidden<span class="p">,</span> |
| num.embed<span class="o">=</span>num.embed<span class="p">,</span> |
| num.label<span class="o">=</span>vocab<span class="p">,</span> |
| batch.size<span class="o">=</span>batch.size<span class="p">,</span> |
| input.size<span class="o">=</span>vocab<span class="p">,</span> |
| initializer<span class="o">=</span>mx.init.uniform<span class="p">(</span><span class="m">0.1</span><span class="p">),</span> |
| learning.rate<span class="o">=</span>learning.rate<span class="p">,</span> |
| wd<span class="o">=</span>wd<span class="p">,</span> |
| clip_gradient<span class="o">=</span>clip_gradient<span class="p">)</span> |
| </pre></div> |
| </div> |
| <div class="highlight-python"><div class="highlight"><pre><span></span> <span class="c1">## Epoch [31] Train: NLL=3.53787130224343, Perp=34.3936275728271</span> |
| <span class="c1">## Epoch [62] Train: NLL=3.43087958036949, Perp=30.903813186055</span> |
| <span class="c1">## Epoch [93] Train: NLL=3.39771238228587, Perp=29.8956319855751</span> |
| <span class="c1">## Epoch [124] Train: NLL=3.37581711716687, Perp=29.2481732041015</span> |
| <span class="c1">## Epoch [155] Train: NLL=3.34523331338447, Perp=28.3671933405139</span> |
| <span class="c1">## Epoch [186] Train: NLL=3.30756356274787, Perp=27.31848454823</span> |
| <span class="c1">## Epoch [217] Train: NLL=3.25642968403829, Perp=25.9566978956055</span> |
| <span class="c1">## Epoch [248] Train: NLL=3.19825967486207, Perp=24.4898727477925</span> |
| <span class="c1">## Epoch [279] Train: NLL=3.14013971549828, Perp=23.1070950525017</span> |
| <span class="c1">## Epoch [310] Train: NLL=3.08747601837462, Perp=21.9216781782189</span> |
| <span class="c1">## Epoch [341] Train: NLL=3.04015595674863, Perp=20.9085038031042</span> |
| <span class="c1">## Epoch [372] Train: NLL=2.99839339255659, Perp=20.0532932584534</span> |
| <span class="c1">## Epoch [403] Train: NLL=2.95940091012609, Perp=19.2864139984503</span> |
| <span class="c1">## Epoch [434] Train: NLL=2.92603311380224, Perp=18.6534872738302</span> |
| <span class="c1">## Epoch [465] Train: NLL=2.89482756896395, Perp=18.0803835531869</span> |
| <span class="c1">## Epoch [496] Train: NLL=2.86668230478397, Perp=17.5786009078994</span> |
| <span class="c1">## Epoch [527] Train: NLL=2.84089368534943, Perp=17.1310684830416</span> |
| <span class="c1">## Epoch [558] Train: NLL=2.81725862932279, Perp=16.7309220880514</span> |
| <span class="c1">## Epoch [589] Train: NLL=2.79518870141492, Perp=16.3657166956952</span> |
| <span class="c1">## Epoch [620] Train: NLL=2.77445683225304, Perp=16.0299176962855</span> |
| <span class="c1">## Epoch [651] Train: NLL=2.75490970113174, Perp=15.719621374694</span> |
| <span class="c1">## Epoch [682] Train: NLL=2.73697900634351, Perp=15.4402696117257</span> |
| <span class="c1">## Epoch [713] Train: NLL=2.72059739336781, Perp=15.1893935780915</span> |
| <span class="c1">## Epoch [744] Train: NLL=2.70462837571585, Perp=14.948760335793</span> |
| <span class="c1">## Epoch [775] Train: NLL=2.68909904683828, Perp=14.7184093476224</span> |
| <span class="c1">## Epoch [806] Train: NLL=2.67460054451836, Perp=14.5065539595711</span> |
| <span class="c1">## Epoch [837] Train: NLL=2.66078997776751, Perp=14.3075873113043</span> |
| <span class="c1">## Epoch [868] Train: NLL=2.6476781639279, Perp=14.1212134100373</span> |
| <span class="c1">## Epoch [899] Train: NLL=2.63529039846876, Perp=13.9473621677371</span> |
| <span class="c1">## Epoch [930] Train: NLL=2.62367693518974, Perp=13.7863219168709</span> |
| <span class="c1">## Epoch [961] Train: NLL=2.61238282674384, Perp=13.6314936713501</span> |
| <span class="c1">## Iter [1] Train: Time: 10301.6818172932 sec, NLL=2.60536539345356, Perp=13.5361704272949</span> |
| <span class="c1">## Iter [1] Val: NLL=2.26093848746227, Perp=9.59208699731232</span> |
| </pre></div> |
| </div> |
| </div> |
| <div class="section" id="build-inference-from-the-model"> |
| <span id="build-inference-from-the-model"></span><h2>Build Inference from the Model<a class="headerlink" href="#build-inference-from-the-model" title="Permalink to this headline">¶</a></h2> |
| <p>Use the helper function for random sample:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> cdf <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>weights<span class="p">)</span> <span class="p">{</span> |
| total <span class="o"><-</span> <span class="kp">sum</span><span class="p">(</span>weights<span class="p">)</span> |
| result <span class="o"><-</span> <span class="kt">c</span><span class="p">()</span> |
| cumsum <span class="o"><-</span> <span class="m">0</span> |
| <span class="kr">for</span> <span class="p">(</span>w <span class="kr">in</span> weights<span class="p">)</span> <span class="p">{</span> |
| cumsum <span class="o"><-</span> <span class="kp">cumsum</span><span class="o">+</span>w |
| result <span class="o"><-</span> <span class="kt">c</span><span class="p">(</span>result<span class="p">,</span> cumsum <span class="o">/</span> total<span class="p">)</span> |
| <span class="p">}</span> |
| <span class="kr">return</span> <span class="p">(</span>result<span class="p">)</span> |
| <span class="p">}</span> |
| |
| search.val <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>cdf<span class="p">,</span> x<span class="p">)</span> <span class="p">{</span> |
| l <span class="o"><-</span> <span class="m">1</span> |
| r <span class="o"><-</span> <span class="kp">length</span><span class="p">(</span>cdf<span class="p">)</span> |
| <span class="kr">while</span> <span class="p">(</span>l <span class="o"><=</span> r<span class="p">)</span> <span class="p">{</span> |
| m <span class="o"><-</span> <span class="kp">as.integer</span><span class="p">((</span>l<span class="o">+</span>r<span class="p">)</span><span class="o">/</span><span class="m">2</span><span class="p">)</span> |
| <span class="kr">if</span> <span class="p">(</span>cdf<span class="p">[</span>m<span class="p">]</span> <span class="o"><</span> x<span class="p">)</span> <span class="p">{</span> |
| l <span class="o"><-</span> m<span class="m">+1</span> |
| <span class="p">}</span> <span class="kr">else</span> <span class="p">{</span> |
| r <span class="o"><-</span> m<span class="m">-1</span> |
| <span class="p">}</span> |
| <span class="p">}</span> |
| <span class="kr">return</span> <span class="p">(</span>l<span class="p">)</span> |
| <span class="p">}</span> |
| choice <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>weights<span class="p">)</span> <span class="p">{</span> |
| cdf.vals <span class="o"><-</span> cdf<span class="p">(</span><span class="kp">as.array</span><span class="p">(</span>weights<span class="p">))</span> |
| x <span class="o"><-</span> runif<span class="p">(</span><span class="m">1</span><span class="p">)</span> |
| idx <span class="o"><-</span> search.val<span class="p">(</span>cdf.vals<span class="p">,</span> x<span class="p">)</span> |
| <span class="kr">return</span> <span class="p">(</span>idx<span class="p">)</span> |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>Use random output or fixed output by choosing the greatest probability:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> make.output <span class="o"><-</span> <span class="kr">function</span><span class="p">(</span>prob<span class="p">,</span> sample<span class="o">=</span><span class="kc">FALSE</span><span class="p">)</span> <span class="p">{</span> |
| <span class="kr">if</span> <span class="p">(</span><span class="o">!</span><span class="kp">sample</span><span class="p">)</span> <span class="p">{</span> |
| idx <span class="o"><-</span> <span class="kp">which.max</span><span class="p">(</span><span class="kp">as.array</span><span class="p">(</span>prob<span class="p">))</span> |
| <span class="p">}</span> |
| <span class="kr">else</span> <span class="p">{</span> |
| idx <span class="o"><-</span> choice<span class="p">(</span>prob<span class="p">)</span> |
| <span class="p">}</span> |
| <span class="kr">return</span> <span class="p">(</span>idx<span class="p">)</span> |
| |
| <span class="p">}</span> |
| </pre></div> |
| </div> |
| <p>In <code class="docutils literal"><span class="pre">mxnet</span></code>, we have a function called <code class="docutils literal"><span class="pre">mx.lstm.inference</span></code> so that users can build an inference from an LSTM model, and then use the <code class="docutils literal"><span class="pre">mx.lstm.forward</span></code> function to get forward output from the inference.</p> |
| <p>Build an inference from the model:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> infer.model <span class="o"><-</span> mx.lstm.inference<span class="p">(</span>num.lstm.layer<span class="o">=</span>num.lstm.layer<span class="p">,</span> |
| input.size<span class="o">=</span>vocab<span class="p">,</span> |
| num.hidden<span class="o">=</span>num.hidden<span class="p">,</span> |
| num.embed<span class="o">=</span>num.embed<span class="p">,</span> |
| num.label<span class="o">=</span>vocab<span class="p">,</span> |
| arg.params<span class="o">=</span>model<span class="o">$</span>arg.params<span class="p">,</span> |
| ctx<span class="o">=</span>mx.cpu<span class="p">())</span> |
| </pre></div> |
| </div> |
| <p>Generate a sequence of 75 characters using the <code class="docutils literal"><span class="pre">mx.lstm.forward</span></code> function:</p> |
| <div class="highlight-r"><div class="highlight"><pre><span></span> start <span class="o"><-</span> <span class="s">'a'</span> |
| seq.len <span class="o"><-</span> <span class="m">75</span> |
| random.sample <span class="o"><-</span> <span class="kc">TRUE</span> |
| |
| last.id <span class="o"><-</span> dic<span class="p">[[</span>start<span class="p">]]</span> |
| out <span class="o"><-</span> <span class="s">"a"</span> |
| <span class="kr">for</span> <span class="p">(</span>i <span class="kr">in</span> <span class="p">(</span><span class="m">1</span><span class="o">:</span><span class="p">(</span>seq.len<span class="m">-1</span><span class="p">)))</span> <span class="p">{</span> |
| input <span class="o"><-</span> <span class="kt">c</span><span class="p">(</span>last.id<span class="m">-1</span><span class="p">)</span> |
| ret <span class="o"><-</span> mx.lstm.forward<span class="p">(</span>infer.model<span class="p">,</span> input<span class="p">,</span> <span class="kc">FALSE</span><span class="p">)</span> |
| infer.model <span class="o"><-</span> ret<span class="o">$</span>model |
| prob <span class="o"><-</span> ret<span class="o">$</span>prob |
| last.id <span class="o"><-</span> make.output<span class="p">(</span>prob<span class="p">,</span> random.sample<span class="p">)</span> |
| out <span class="o"><-</span> <span class="kp">paste0</span><span class="p">(</span>out<span class="p">,</span> lookup.table<span class="p">[[</span>last.id<span class="p">]])</span> |
| <span class="p">}</span> |
| cat <span class="p">(</span><span class="kp">paste0</span><span class="p">(</span>out<span class="p">,</span> <span class="s">"\n"</span><span class="p">))</span> |
| </pre></div> |
| </div> |
| <p>The result:</p> |
| <div class="highlight-python"><div class="highlight"><pre><span></span> ah not a drobl greens |
| Settled asing lately sistering sounted to their hight |
| </pre></div> |
| </div> |
| </div> |
| <div class="section" id="create-other-rnn-models"> |
| <span id="create-other-rnn-models"></span><h2>Create Other RNN Models<a class="headerlink" href="#create-other-rnn-models" title="Permalink to this headline">¶</a></h2> |
| <p>In <code class="docutils literal"><span class="pre">mxnet</span></code>, other RNN models, like custom RNN and GRU, are also provided:</p> |
| <ul class="simple"> |
| <li>For a custom RNN model, you can replace <code class="docutils literal"><span class="pre">mx.lstm</span></code> with <code class="docutils literal"><span class="pre">mx.rnn</span></code> to train an RNN model. You can replace <code class="docutils literal"><span class="pre">mx.lstm.inference</span></code> and <code class="docutils literal"><span class="pre">mx.lstm.forward</span></code> with <code class="docutils literal"><span class="pre">mx.rnn.inference</span></code> and <code class="docutils literal"><span class="pre">mx.rnn.forward</span></code> to build inference from an RNN model and get the forward result from the inference model.</li> |
| <li>For a GRU model, you can replace <code class="docutils literal"><span class="pre">mx.lstm</span></code> with <code class="docutils literal"><span class="pre">mx.gru</span></code> to train a GRU model. You can replace <code class="docutils literal"><span class="pre">mx.lstm.inference</span></code> and <code class="docutils literal"><span class="pre">mx.lstm.forward</span></code> with <code class="docutils literal"><span class="pre">mx.gru.inference</span></code> and <code class="docutils literal"><span class="pre">mx.gru.forward</span></code> to build inference from a GRU model and get the forward result from the inference model.</li> |
| </ul> |
| </div> |
| <div class="section" id="next-steps"> |
| <span id="next-steps"></span><h2>Next Steps<a class="headerlink" href="#next-steps" title="Permalink to this headline">¶</a></h2> |
| <div class="toctree-wrapper compound"> |
| <ul> |
| <li class="toctree-l1"><a class="reference external" href="http://mxnet.io/tutorials/index.html">MXNet tutorials index</a></li> |
| </ul> |
| </div> |
| </div> |
| </div> |
| <div class="container"> |
| <div class="footer"> |
| <p> © 2015-2017 DMLC. All rights reserved. </p> |
| </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="#">Char RNN Example</a><ul> |
| <li><a class="reference internal" href="#load-the-data">Load the Data</a></li> |
| <li><a class="reference internal" href="#train-the-model">Train the Model</a></li> |
| <li><a class="reference internal" href="#build-inference-from-the-model">Build Inference from the Model</a></li> |
| <li><a class="reference internal" href="#create-other-rnn-models">Create Other RNN Models</a></li> |
| <li><a class="reference internal" href="#next-steps">Next Steps</a></li> |
| </ul> |
| </li> |
| </ul> |
| </div> |
| </div> |
| </div> <!-- pagename != index --> |
| <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 type="text/javascript"> |
| $('body').ready(function () { |
| $('body').css('visibility', 'visible'); |
| }); |
| </script> |
| </div></body> |
| </html> |
