index_en.html 12.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10


<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
11
  <title>Distributed Training &mdash; PaddlePaddle  documentation</title>
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
  

  
  

  

  
  
    

  

  
  
    <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
  

  
  
        <link rel="index" title="Index"
              href="../../genindex.html"/>
        <link rel="search" title="Search" href="../../search.html"/>
    <link rel="top" title="PaddlePaddle  documentation" href="../../index.html"/>
36
        <link rel="up" title="HOW TO" href="../index_en.html"/>
37
        <link rel="next" title="Preparations" href="preparations_en.html"/>
38
        <link rel="prev" title="Detail Description" href="../cmd_parameter/detail_introduction_en.html"/> 
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67

  <link rel="stylesheet" href="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/css/perfect-scrollbar.min.css" type="text/css" />
  <link rel="stylesheet" href="../../_static/css/override.css" type="text/css" />
  <script>
  var _hmt = _hmt || [];
  (function() {
    var hm = document.createElement("script");
    hm.src = "//hm.baidu.com/hm.js?b9a314ab40d04d805655aab1deee08ba";
    var s = document.getElementsByTagName("script")[0]; 
    s.parentNode.insertBefore(hm, s);
  })();
  </script>

  

  
  <script src="../../_static/js/modernizr.min.js"></script>

</head>

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

  
  <header class="site-header">
    <div class="site-logo">
      <a href="/"><img src="../../_static/images/PP_w.png"></a>
    </div>
    <div class="site-nav-links">
      <div class="site-menu">
68
        <a class="fork-on-github" href="https://github.com/PaddlePaddle/Paddle" target="_blank"><i class="fa fa-github"></i>Fork me on Github</a>
69 70 71 72 73 74 75 76 77 78 79 80
        <div class="language-switcher dropdown">
          <a type="button" data-toggle="dropdown">
            <span>English</span>
            <i class="fa fa-angle-up"></i>
            <i class="fa fa-angle-down"></i>
          </a>
          <ul class="dropdown-menu">
            <li><a href="/doc_cn">中文</a></li>
            <li><a href="/doc">English</a></li>
          </ul>
        </div>
        <ul class="site-page-links">
81
          <li><a href="/">Home</a></li>
82 83 84 85 86
        </ul>
      </div>
      <div class="doc-module">
        
        <ul class="current">
87 88 89 90
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_en.html">GET STARTED</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_en.html">Install and Build</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="../index_en.html">HOW TO</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_en.html">Development</a></li>
91
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_en.html">FAQ</a></li>
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
</ul>

        
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>        
      </div>
    </div>
  </header>
  
  <div class="main-content-wrap">

    
    <nav class="doc-menu-vertical" role="navigation">
        
          
          <ul class="current">
113 114
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_en.html">GET STARTED</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/quickstart_en.html">Quick Start</a></li>
115
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/concepts/use_concepts_en.html">Basic Concept</a></li>
116 117
</ul>
</li>
118
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_en.html">Install and Build</a><ul>
119
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/pip_install_en.html">Install using pip</a></li>
120 121
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/docker_install_en.html">Run in Docker Containers</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/build_from_source_en.html">Build from Sources</a></li>
122 123
</ul>
</li>
124 125 126 127 128
<li class="toctree-l1 current"><a class="reference internal" href="../index_en.html">HOW TO</a><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="../cmd_parameter/index_en.html">Set Command-line Parameters</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/use_case_en.html">Use Case</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/arguments_en.html">Argument Outline</a></li>
<li class="toctree-l3"><a class="reference internal" href="../cmd_parameter/detail_introduction_en.html">Detail Description</a></li>
129 130
</ul>
</li>
131 132 133 134
<li class="toctree-l2 current"><a class="current reference internal" href="#">Distributed Training</a><ul>
<li class="toctree-l3"><a class="reference internal" href="preparations_en.html">Preparations</a></li>
<li class="toctree-l3"><a class="reference internal" href="cmd_argument_en.html">Command-line arguments</a></li>
<li class="toctree-l3"><a class="reference internal" href="multi_cluster/index_en.html">Use different clusters</a><ul>
135 136 137 138
<li class="toctree-l4"><a class="reference internal" href="multi_cluster/fabric_en.html">Fabric</a></li>
<li class="toctree-l4"><a class="reference internal" href="multi_cluster/openmpi_en.html">OpenMPI</a></li>
<li class="toctree-l4"><a class="reference internal" href="multi_cluster/k8s_en.html">Kubernetes</a></li>
<li class="toctree-l4"><a class="reference internal" href="multi_cluster/k8s_aws_en.html">Kubernetes on AWS</a></li>
139 140
</ul>
</li>
141 142
</ul>
</li>
143 144
<li class="toctree-l2"><a class="reference internal" href="../rnn/index_en.html">RNN Models</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../rnn/rnn_config_en.html">RNN Configuration</a></li>
145 146 147
<li class="toctree-l3"><a class="reference internal" href="../rnn/recurrent_group_en.html">Recurrent Group Tutorial</a></li>
<li class="toctree-l3"><a class="reference internal" href="../rnn/hierarchical_layer_en.html">Layers supporting hierarchical sequence as input</a></li>
<li class="toctree-l3"><a class="reference internal" href="../rnn/hrnn_rnn_api_compare_en.html">API comparision between RNN and hierarchical RNN</a></li>
148 149 150 151 152 153 154 155
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../optimization/gpu_profiling_en.html">Tune GPU Performance</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_en.html">Development</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../dev/contribute_to_paddle_en.html">Contribute Code</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../dev/write_docs_en.html">Contribute Documentation</a></li>
156 157
</ul>
</li>
158 159 160 161 162 163 164 165
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_en.html">FAQ</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../faq/build_and_install/index_en.html">Install, Build and Unit test</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/model/index_en.html">Model Configuration</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/parameter/index_en.html">Parameter Setting</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/local/index_en.html">Local Training and Prediction</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/cluster/index_en.html">Cluster Training and Prediction</a></li>
</ul>
</li>
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
</ul>

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

      

 







<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
186
        <li><a href="../index_en.html">HOW TO</a> > </li>
187
      
188
    <li>Distributed Training</li>
189 190 191 192 193 194 195 196
  </ul>
</div>
      
      <div class="wy-nav-content" id="doc-content">
        <div class="rst-content">
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
197 198
  <div class="section" id="distributed-training">
<h1>Distributed Training<a class="headerlink" href="#distributed-training" title="Permalink to this headline"></a></h1>
199 200 201 202 203 204 205 206 207
<p>In this section, we&#8217;ll explain how to run distributed training jobs with PaddlePaddle on different types of clusters. The diagram below shows the main architecture of a distributed trainning job:</p>
<a class="reference internal image-reference" href="../../_images/ps_en.png"><img alt="../../_images/ps_en.png" src="../../_images/ps_en.png" style="width: 500px;" /></a>
<ul class="simple">
<li>Data shard: training data will be split into multiple partitions, trainers use the partitions of the whole dataset to do the training job.</li>
<li>Trainer: each trainer reads the data shard, and train the neural network. Then the trainer will upload calculated &#8220;gradients&#8221; to parameter servers, and wait for parameters to be optimized on the parameter server side. When that finishes, the trainer download optimized parameters and continues its training.</li>
<li>Parameter server: every parameter server stores part of the whole neural network model data. They will do optimization calculations when gradients are uploaded from trainers, and then send updated parameters to trainers.</li>
</ul>
<p>PaddlePaddle can support both synchronize stochastic gradient descent (SGD) and asynchronous SGD.</p>
<p>When training with synchronize SGD, PaddlePaddle uses an internal &#8220;synchronize barrier&#8221; which makes gradients update and parameter download in strict order. On the other hand, asynchronous SGD won&#8217;t wait for all trainers to finish upload at a single step, this will increase the parallelism of distributed training: parameter servers do not depend on each other, they&#8217;ll do parameter optimization concurrently. Parameter servers will not wait for trainers, so trainers will also do their work concurrently. But asynchronous SGD will introduce more randomness and noises in the gradient.</p>
208 209
<div class="toctree-wrapper compound">
<ul>
210 211 212
<li class="toctree-l1"><a class="reference internal" href="preparations_en.html">Preparations</a></li>
<li class="toctree-l1"><a class="reference internal" href="cmd_argument_en.html">Command-line arguments</a></li>
<li class="toctree-l1"><a class="reference internal" href="multi_cluster/index_en.html">Use different clusters</a></li>
213 214 215 216 217 218 219 220 221 222 223
</ul>
</div>
</div>


           </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
224
        <a href="preparations_en.html" class="btn btn-neutral float-right" title="Preparations" accesskey="n">Next <span class="fa fa-arrow-circle-right"></span></a>
225 226
      
      
227
        <a href="../cmd_parameter/detail_introduction_en.html" class="btn btn-neutral" title="Detail Description" accesskey="p"><span class="fa fa-arrow-circle-left"></span> Previous</a>
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260
      
    </div>
  

  <hr/>

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

    </p>
  </div>
  Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>. 

</footer>

        </div>
      </div>

    </section>

  </div>
  


  

    <script type="text/javascript">
        var DOCUMENTATION_OPTIONS = {
            URL_ROOT:'../../',
            VERSION:'',
            COLLAPSE_INDEX:false,
            FILE_SUFFIX:'.html',
261 262
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
263 264 265 266 267
        };
    </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>
268
      <script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
269 270 271 272 273 274 275 276 277 278 279 280 281
       
  

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

</body>
282
</html>