cluster_train_en.html



<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  <title>PaddlePaddle Distributed Training &mdash; PaddlePaddle  documentation</title>
  

    <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"/>
        <link rel="up" title="HOW TO" href="../../index_en.html"/>
        <link rel="next" title="Paddle On Kubernetes" href="../k8s/k8s_en.html"/>
        <link rel="prev" title="Detail Description" href="../cmd_parameter/detail_introduction_en.html"/> 

  <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">
        <a class="fork-on-github" href="https://github.com/PaddlePaddle/Paddle" target="_blank"><i class="fa fa-github"></i>Fork me on Github</a>
        <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">
          <li><a href="/">Home</a></li>
        </ul>
      </div>
      <div class="doc-module">
        
        <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../../../getstarted/index_en.html">GET STARTED</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="../../../api/index_en.html">API</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../mobile/index_en.html">MOBILE</a></li>
</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">
<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/build_and_install/index_en.html">Install and Build</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../../getstarted/build_and_install/docker_install_en.html">PaddlePaddle in Docker Containers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../getstarted/build_and_install/build_from_source_en.html">Installing from Sources</a></li>
</ul>
</li>
</ul>
</li>
<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>
</ul>
</li>
<li class="toctree-l2 current"><a class="current reference internal" href="#">PaddlePaddle Distributed Training</a></li>
<li class="toctree-l2"><a class="reference internal" href="../k8s/k8s_en.html">Paddle On Kubernetes</a></li>
<li class="toctree-l2"><a class="reference internal" href="../k8s/k8s_aws_en.html">Distributed PaddlePaddle Training on AWS with Kubernetes</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../dev/build_en.html">Build PaddlePaddle from Source Code and Run Unit Test</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../dev/new_layer_en.html">Write New Layers</a></li>
<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="../../deep_model/rnn/index_en.html">RNN Models</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../deep_model/rnn/rnn_config_en.html">RNN Configuration</a></li>
</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="../../../api/index_en.html">API</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../../api/v2/model_configs.html">Model Configuration</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/activation.html">Activation</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/layer.html">Layers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/evaluators.html">Evaluators</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/optimizer.html">Optimizer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/pooling.html">Pooling</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/networks.html">Networks</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../../api/v2/config/attr.html">Parameter Attribute</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../../api/v2/data.html">Data Reader Interface and DataSets</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../../api/v2/run_logic.html">Training and Inference</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../../mobile/index_en.html">MOBILE</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../../mobile/cross_compiling_for_android_en.html">Build PaddlePaddle for Android</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../../mobile/cross_compiling_for_raspberry_en.html">Build PaddlePaddle for Raspberry Pi</a></li>
</ul>
</li>
</ul>

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

      
<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
        <li><a href="../../index_en.html">HOW TO</a> > </li>
      
    <li>PaddlePaddle Distributed Training</li>
  </ul>
</div>
      
      <div class="wy-nav-content" id="doc-content">
        <div class="rst-content">
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="paddlepaddle-distributed-training">
<span id="paddlepaddle-distributed-training"></span><h1>PaddlePaddle Distributed Training<a class="headerlink" href="#paddlepaddle-distributed-training" title="Permalink to this headline">¶</a></h1>
<ul class="simple">
<li><a class="reference external" href="#introduction">Introduction</a></li>
<li><a class="reference external" href="#preparations">Preparations</a></li>
<li><a class="reference external" href="#command-line-arguments">Command-line arguments</a><ul>
<li><a class="reference external" href="#starting-parameter-server">Starting parameter server</a></li>
<li><a class="reference external" href="#starting-trainer">Starting trainer</a></li>
<li><a class="reference external" href="#prepare-training-dataset">Prepare Training Dataset</a></li>
<li><a class="reference external" href="#prepare-training-program">Prepare Training program</a></li>
</ul>
</li>
<li><a class="reference external" href="#use-cluster-platforms-or-cluster-management-tools">Use cluster platforms or cluster management tools</a><ul>
<li><a class="reference external" href="#cluster-training-using-fabric">Cluster Training Using Fabric</a><ul>
<li><a class="reference external" href="#prepare-a-linux-cluster">Prepare a Linux cluster</a></li>
<li><a class="reference external" href="#launching-cluster-job">Launching Cluster Job</a></li>
<li><a class="reference external" href="#kill-cluster-job">Kill Cluster Job</a></li>
<li><a class="reference external" href="#check-cluster-training-result">Check Cluster Training Result</a></li>
<li><a class="reference external" href="#check-model-output">Check Model Output</a></li>
</ul>
</li>
<li><a class="reference external" href="#cluster-training-using-openmpi">Cluster Training Using OpenMPI</a><ul>
<li><a class="reference external" href="#prepare-an-openmpi-cluster">Prepare an OpenMPI cluster</a></li>
<li><a class="reference external" href="#launching-cluster-job-1">Launching Cluster Job</a></li>
</ul>
</li>
<li><a class="reference external" href="#cluster-training-using-kubernetes">Cluster Training Using Kubernetes</a></li>
</ul>
</li>
</ul>
<div class="section" id="introduction">
<span id="introduction"></span><h2>Introduction<a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
<p>In this article, 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>
<p><img src="https://user-images.githubusercontent.com/13348433/31772146-41523d84-b511-11e7-8a12-a69fd136c283.png" width="500"></p>
<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>
</div>
<div class="section" id="preparations">
<span id="preparations"></span><h2>Preparations<a class="headerlink" href="#preparations" title="Permalink to this headline">¶</a></h2>
<ol class="simple">
<li>Prepare your computer cluster. It&#8217;s normally a bunch of Linux servers connected by LAN. Each server will be assigned a unique IP address. The computers in the cluster can be called &#8220;nodes&#8221;.</li>
<li>Install PaddlePaddle on every node. If you are going to take advantage of GPU cards, you&#8217;ll also need to install proper driver and CUDA libraries. To install PaddlePaddle please read <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/tree/develop/doc/getstarted/build_and_install">this build and install</a> document. We strongly recommend using <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_en.rst">Docker installation</a>.</li>
</ol>
<p>After installation, you can check the version by typing the below command (run a docker container  if using docker: <code class="docutils literal"><span class="pre">docker</span> <span class="pre">run</span> <span class="pre">-it</span> <span class="pre">paddlepaddle/paddle:[tag]</span> <span class="pre">/bin/bash</span></code>):</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span>$ paddle version
PaddlePaddle <span class="m">0</span>.10.0rc, compiled with
    with_avx: ON
    with_gpu: OFF
    with_double: OFF
    with_python: ON
    with_rdma: OFF
    with_timer: OFF
</pre></div>
</div>
<p>We&#8217;ll take <code class="docutils literal"><span class="pre">doc/howto/usage/cluster/src/word2vec</span></code> as an example to introduce distributed training using PaddlePaddle v2 API.</p>
</div>
<div class="section" id="command-line-arguments">
<span id="command-line-arguments"></span><h2>Command-line arguments<a class="headerlink" href="#command-line-arguments" title="Permalink to this headline">¶</a></h2>
<div class="section" id="starting-parameter-server">
<span id="starting-parameter-server"></span><h3>Starting parameter server<a class="headerlink" href="#starting-parameter-server" title="Permalink to this headline">¶</a></h3>
<p>Type the below command to start a parameter server which will wait for trainers to connect:</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span>$ paddle pserver --port<span class="o">=</span><span class="m">7164</span> --ports_num<span class="o">=</span><span class="m">1</span> --ports_num_for_sparse<span class="o">=</span><span class="m">1</span> --num_gradient_servers<span class="o">=</span><span class="m">1</span>
</pre></div>
</div>
<p>If you wish to run parameter servers in background, and save a log file, you can type:</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span>$ stdbuf -oL /usr/bin/nohup paddle pserver --port<span class="o">=</span><span class="m">7164</span> --ports_num<span class="o">=</span><span class="m">1</span> --ports_num_for_sparse<span class="o">=</span><span class="m">1</span> --num_gradient_servers<span class="o">=</span><span class="m">1</span> <span class="p">&amp;</span>&gt; pserver.log
</pre></div>
</div>
<p>| param  | required | default | description |
| &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- |
| port  | required | 7164 | port which parameter server will listen on. If ports_num greater than 1, parameter server will listen on multiple ports for more network throughput |
| ports_num  | required | 1 | total number of ports will listen on  |
| ports_num_for_sparse  | required | 1 | number of ports which serves sparse parameter update  |
| num_gradient_servers  | required | 1 | total number of gradient servers |</p>
</div>
<div class="section" id="starting-trainer">
<span id="starting-trainer"></span><h3>Starting trainer<a class="headerlink" href="#starting-trainer" title="Permalink to this headline">¶</a></h3>
<p>Type the command below to start the trainer(name the file whatever you want, like &#8220;train.py&#8221;)</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span>$ python train.py
</pre></div>
</div>
<p>Trainers&#8217; network need to be connected with parameter servers&#8217; network to finish the job. Trainers need to know port and IPs to locate parameter servers. You can pass arguments to trainers through <a class="reference external" href="https://en.wikipedia.org/wiki/Environment_variable">environment variables</a> or pass to <code class="docutils literal"><span class="pre">paddle.init()</span></code> function. Arguments passed to the <code class="docutils literal"><span class="pre">paddle.init()</span></code> function will overwrite environment variables.</p>
<p>Use environment viriables:</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nb">export</span> <span class="nv">PADDLE_INIT_USE_GPU</span><span class="o">=</span>False
<span class="nb">export</span> <span class="nv">PADDLE_INIT_TRAINER_COUNT</span><span class="o">=</span><span class="m">1</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_PORT</span><span class="o">=</span><span class="m">7164</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_PORTS_NUM</span><span class="o">=</span><span class="m">1</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_PORTS_NUM_FOR_SPARSE</span><span class="o">=</span><span class="m">1</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_NUM_GRADIENT_SERVERS</span><span class="o">=</span><span class="m">1</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_TRAINER_ID</span><span class="o">=</span><span class="m">0</span>
<span class="nb">export</span> <span class="nv">PADDLE_INIT_PSERVERS</span><span class="o">=</span><span class="m">127</span>.0.0.1
python train.py
</pre></div>
</div>
<p>Pass arguments:</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">paddle</span><span class="o">.</span><span class="n">init</span><span class="p">(</span>
        <span class="n">use_gpu</span><span class="o">=</span><span class="bp">False</span><span class="p">,</span>
        <span class="n">trainer_count</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
        <span class="n">port</span><span class="o">=</span><span class="mi">7164</span><span class="p">,</span>
        <span class="n">ports_num</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
        <span class="n">ports_num_for_sparse</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
        <span class="n">num_gradient_servers</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
        <span class="n">trainer_id</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span>
        <span class="n">pservers</span><span class="o">=</span><span class="s2">&quot;127.0.0.1&quot;</span><span class="p">)</span>
</pre></div>
</div>
<p>| param  | required | default | description |
| &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- | &#8212;&#8212;&#8212;&#8212;- |
| use_gpu  | optional | False | set to &#8220;True&#8221; to enable GPU training |
| trainer_count  | required | 1 | total count of trainers in the training job |
| port  | required | 7164 | port to connect to parameter server  |
| ports_num  | required | 1 | number of ports for communication |
| ports_num_for_sparse  | required | 1 | number of ports for sparse type caculation |
| num_gradient_servers  | required | 1 | total number of gradient server |
| trainer_id  | required | 0 | ID for every trainer, start from 0 |
| pservers  | required | 127.0.0.1 | list of IPs of parameter servers, separated by &#8221;,&#8221; |</p>
</div>
<div class="section" id="prepare-training-dataset">
<span id="prepare-training-dataset"></span><h3>Prepare Training Dataset<a class="headerlink" href="#prepare-training-dataset" title="Permalink to this headline">¶</a></h3>
<p>Here&#8217;s some example code <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py">prepare.py</a>, it will download public <code class="docutils literal"><span class="pre">imikolov</span></code> dataset and split it into multiple files according to job parallelism(trainers count). Modify <code class="docutils literal"><span class="pre">SPLIT_COUNT</span></code> at the begining of <code class="docutils literal"><span class="pre">prepare.py</span></code> to change the count of output files.</p>
<p>In the real world, we often use <code class="docutils literal"><span class="pre">MapReduce</span></code> job&#8217;s output as training data, so there will be lots of files. You can use <code class="docutils literal"><span class="pre">mod</span></code> to assign training file to trainers:</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">os</span>
<span class="n">train_list</span> <span class="o">=</span> <span class="p">[]</span>
<span class="n">flist</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">listdir</span><span class="p">(</span><span class="s2">&quot;/train_data/&quot;</span><span class="p">)</span>
<span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">flist</span><span class="p">:</span>
  <span class="n">suffix</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">f</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s2">&quot;-&quot;</span><span class="p">)[</span><span class="mi">1</span><span class="p">])</span>
  <span class="k">if</span> <span class="n">suffix</span> <span class="o">%</span> <span class="n">TRAINER_COUNT</span> <span class="o">==</span> <span class="n">TRAINER_ID</span><span class="p">:</span>
    <span class="n">train_list</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">f</span><span class="p">)</span>
</pre></div>
</div>
<p>Example code <code class="docutils literal"><span class="pre">prepare.py</span></code> will split training data and testing data into 3 files with digital suffix like <code class="docutils literal"><span class="pre">-00000</span></code>, <code class="docutils literal"><span class="pre">-00001</span></code> and<code class="docutils literal"><span class="pre">-00002</span></code>:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">train</span><span class="o">.</span><span class="n">txt</span>
<span class="n">train</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00000</span>
<span class="n">train</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00001</span>
<span class="n">train</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00002</span>
<span class="n">test</span><span class="o">.</span><span class="n">txt</span>
<span class="n">test</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00000</span>
<span class="n">test</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00001</span>
<span class="n">test</span><span class="o">.</span><span class="n">txt</span><span class="o">-</span><span class="mi">00002</span>
</pre></div>
</div>
<p>When job started, every trainer needs to get it&#8217;s own part of data. In some distributed systems a storage service will be provided, so the date under that path can be accessed by all the trainer nodes. Without the storage service, you must copy the training data to each trainer node.</p>
<p>Different training jobs may have different data format and <code class="docutils literal"><span class="pre">reader()</span></code> function, developers may need to write different data prepare scripts and <code class="docutils literal"><span class="pre">reader()</span></code> functions for their job.</p>
</div>
<div class="section" id="prepare-training-program">
<span id="prepare-training-program"></span><h3>Prepare Training program<a class="headerlink" href="#prepare-training-program" title="Permalink to this headline">¶</a></h3>
<p>We&#8217;ll create a <em>workspace</em> directory on each node, storing your training program, dependencies, mounted or downloaded dataset directory.</p>
<p>Your workspace may looks like:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span>.
|-- my_lib.py
|-- word_dict.pickle
|-- train.py
|-- train_data_dir/
|   |-- train.txt-00000
|   |-- train.txt-00001
|   |-- train.txt-00002
`-- test_data_dir/
    |-- test.txt-00000
    |-- test.txt-00001
    `-- test.txt-00002
</pre></div>
</div>
<ul>
<li><p class="first"><code class="docutils literal"><span class="pre">my_lib.py</span></code>: user defined libraries, like PIL libs. This is optional.</p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">word_dict.pickle</span></code>: dict file for training word embeding.</p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">train.py</span></code>: training program. Sample code: <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py">api_train_v2_cluster.py</a>. <strong><em>NOTE:</em></strong> You may need to modify the head part of <code class="docutils literal"><span class="pre">train.py</span></code> when using different cluster platform to retrive configuration environment variables:</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">cluster_train_file</span> <span class="o">=</span> <span class="s2">&quot;./train_data_dir/train/train.txt&quot;</span>
<span class="n">cluster_test_file</span> <span class="o">=</span> <span class="s2">&quot;./test_data_dir/test/test.txt&quot;</span>
<span class="n">node_id</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">getenv</span><span class="p">(</span><span class="s2">&quot;OMPI_COMM_WORLD_RANK&quot;</span><span class="p">)</span>
<span class="k">if</span> <span class="ow">not</span> <span class="n">node_id</span><span class="p">:</span>
    <span class="k">raise</span> <span class="ne">EnvironmentError</span><span class="p">(</span><span class="s2">&quot;must provied OMPI_COMM_WORLD_RANK&quot;</span><span class="p">)</span>
</pre></div>
</div>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">train_data_dir</span></code>: containing training data. Mount from storage service or copy trainning data to here.</p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">test_data_dir</span></code>: containing testing data.</p>
</li>
</ul>
</div>
</div>
<div class="section" id="use-cluster-platforms-or-cluster-management-tools">
<span id="use-cluster-platforms-or-cluster-management-tools"></span><h2>Use cluster platforms or cluster management tools<a class="headerlink" href="#use-cluster-platforms-or-cluster-management-tools" title="Permalink to this headline">¶</a></h2>
<p>PaddlePaddle supports running jobs on several platforms including:</p>
<ul class="simple">
<li><a class="reference external" href="http://kubernetes.io">Kubernetes</a> open-source system for automating deployment, scaling, and management of containerized applications from Google.</li>
<li><a class="reference external" href="https://www.open-mpi.org">OpenMPI</a> Mature high performance parallel computing framework.</li>
<li><a class="reference external" href="http://www.fabfile.org">Fabric</a> A cluster management tool. Write scripts to submit jobs or manage the cluster.</li>
</ul>
<p>We&#8217;ll introduce cluster job management on these platforms. The examples can be found under <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/scripts/cluster_train_v2">cluster_train_v2</a>.</p>
<p>These cluster platforms provide API or environment variables for training processes, when the job is dispatched to different nodes. Like node ID, IP or total number of nodes etc.</p>
<div class="section" id="cluster-training-using-fabric">
<span id="cluster-training-using-fabric"></span><h3>Cluster Training Using Fabric<a class="headerlink" href="#cluster-training-using-fabric" title="Permalink to this headline">¶</a></h3>
<div class="section" id="prepare-a-linux-cluster">
<span id="prepare-a-linux-cluster"></span><h4>Prepare a Linux cluster<a class="headerlink" href="#prepare-a-linux-cluster" title="Permalink to this headline">¶</a></h4>
<p>Run <code class="docutils literal"><span class="pre">kubectl</span> <span class="pre">-f</span> <span class="pre">ssh_servers.yaml</span></code> under the directory:  <code class="docutils literal"><span class="pre">paddle/scripts/cluster_train_v2/fabric/docker_cluster</span></code> will launch a demo cluster. Run <code class="docutils literal"><span class="pre">kubectl</span> <span class="pre">get</span> <span class="pre">po</span> <span class="pre">-o</span> <span class="pre">wide</span></code> to get IP addresses of these nodes.</p>
</div>
<div class="section" id="launching-cluster-job">
<span id="launching-cluster-job"></span><h4>Launching Cluster Job<a class="headerlink" href="#launching-cluster-job" title="Permalink to this headline">¶</a></h4>
<p><code class="docutils literal"><span class="pre">paddle.py</span></code> provides automatical scripts to start all PaddlePaddle cluster processes in different nodes. By default, all command line options can be set as <code class="docutils literal"><span class="pre">paddle.py</span></code> command options and <code class="docutils literal"><span class="pre">paddle.py</span></code> will transparently and automatically set these options to PaddlePaddle lower level processes.</p>
<p><code class="docutils literal"><span class="pre">paddle.py</span></code>provides two distinguished command option for easy job launching.</p>
<ul class="simple">
<li><code class="docutils literal"><span class="pre">job_dispatch_package</span></code> set it with local <code class="docutils literal"><span class="pre">workspace</span></code> directory, it will be dispatched to all nodes which is set in <code class="docutils literal"><span class="pre">conf.py</span></code>. It could be helpful for frequently manipulating workspace files. otherwise, frequent multi-nodes workspace deployment is very annoying.</li>
<li><code class="docutils literal"><span class="pre">job_workspace</span></code>  set it with already deployed workspace directory, <code class="docutils literal"><span class="pre">paddle.py</span></code> will skip dispatch stage to directly launch cluster job with all nodes. It could help to reduce heavy
dispatch latency.</li>
</ul>
<p><code class="docutils literal"><span class="pre">cluster_train/run.sh</span></code> provides command line sample to run <code class="docutils literal"><span class="pre">demo/recommendation</span></code> cluster job, just modify <code class="docutils literal"><span class="pre">job_dispatch_package</span></code> and <code class="docutils literal"><span class="pre">job_workspace</span></code> with your defined directory, then:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">sh</span> <span class="n">run</span><span class="o">.</span><span class="n">sh</span>
</pre></div>
</div>
<p>The cluster Job will start in several seconds.</p>
</div>
<div class="section" id="kill-cluster-job">
<span id="kill-cluster-job"></span><h4>Kill Cluster Job<a class="headerlink" href="#kill-cluster-job" title="Permalink to this headline">¶</a></h4>
<p><code class="docutils literal"><span class="pre">paddle.py</span></code> can capture <code class="docutils literal"><span class="pre">Ctrl</span> <span class="pre">+</span> <span class="pre">C</span></code> SIGINT signal to automatically kill all processes launched by it. So just stop <code class="docutils literal"><span class="pre">paddle.py</span></code> to kill cluster job. You should manually kill the job if the program crashed.</p>
</div>
<div class="section" id="check-cluster-training-result">
<span id="check-cluster-training-result"></span><h4>Check Cluster Training Result<a class="headerlink" href="#check-cluster-training-result" title="Permalink to this headline">¶</a></h4>
<p>Check log in $workspace/log for details, each node owns same log structure.</p>
<p><code class="docutils literal"><span class="pre">paddle_trainer.INFO</span></code>
It provides almost all internal output log for training,  same as local training. Check runtime model convergence here.</p>
<p><code class="docutils literal"><span class="pre">paddle_pserver2.INFO</span></code>
It provides parameter server running log, which could help to diagnose distributed error.</p>
<p><code class="docutils literal"><span class="pre">server.log</span></code>
It provides stderr and stdout of parameter server process. Check error log if training crashes.</p>
<p><code class="docutils literal"><span class="pre">train.log</span></code>
It provides stderr and stdout of trainer process. Check error log if training crashes.</p>
</div>
<div class="section" id="check-model-output">
<span id="check-model-output"></span><h4>Check Model Output<a class="headerlink" href="#check-model-output" title="Permalink to this headline">¶</a></h4>
<p>After one pass finished, model files will be written in <code class="docutils literal"><span class="pre">output</span></code> directory in node 0.
<code class="docutils literal"><span class="pre">nodefile</span></code> in workspace indicates the node id of current cluster job.</p>
</div>
</div>
<div class="section" id="cluster-training-using-openmpi">
<span id="cluster-training-using-openmpi"></span><h3>Cluster Training Using OpenMPI<a class="headerlink" href="#cluster-training-using-openmpi" title="Permalink to this headline">¶</a></h3>
<div class="section" id="prepare-an-openmpi-cluster">
<span id="prepare-an-openmpi-cluster"></span><h4>Prepare an OpenMPI cluster<a class="headerlink" href="#prepare-an-openmpi-cluster" title="Permalink to this headline">¶</a></h4>
<p>Run the following command to start a 3-node MPI cluster and one &#8220;head&#8221; node.</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nb">cd</span> paddle/scripts/cluster_train_v2/openmpi/docker_cluster
kubectl create -f head.yaml
kubectl create -f mpi-nodes.yaml
</pre></div>
</div>
<p>Then you can log in to every OpenMPI node using ssh without input any passwords.</p>
</div>
<div class="section" id="launching-cluster-job">
<span id="id1"></span><h4>Launching Cluster Job<a class="headerlink" href="#launching-cluster-job" title="Permalink to this headline">¶</a></h4>
<p>Follow the steps to launch a PaddlePaddle training job in OpenMPI cluster:\</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="c1"># find out node IP addresses</span>
kubectl get po -o wide
<span class="c1"># generate a &quot;machines&quot; file containing node IP addresses</span>
kubectl get po -o wide <span class="p">|</span> grep nodes <span class="p">|</span> awk <span class="s1">&#39;{print $6}&#39;</span> &gt; machines
<span class="c1"># copy necessary files onto &quot;head&quot; node</span>
scp -i ssh/id_rsa.mpi.pub machines prepare.py train.py start_mpi_train.sh tutorial@<span class="o">[</span>headIP<span class="o">]</span>:~
<span class="c1"># login to head node using ssh</span>
ssh -i ssh/id_rsa.mpi.pub tutorial@<span class="o">[</span>headIP<span class="o">]</span>
<span class="c1"># --------------- in head node ---------------</span>
<span class="c1"># prepare training data</span>
python prepare.py
<span class="c1"># copy training data and dict file to MPI nodes</span>
cat machines <span class="p">|</span> xargs -i scp word_dict.pickle train.py start_mpi_train.sh machines <span class="o">{}</span>:/home/tutorial
<span class="c1"># creat a directory for storing log files</span>
mpirun -hostfile machines -n <span class="m">3</span> mkdir /home/tutorial/logs
<span class="c1"># copy training data to every node</span>
scp train.txt-00000 test.txt-00000 <span class="o">[</span>node1IP<span class="o">]</span>:/home/tutorial
scp train.txt-00001 test.txt-00001 <span class="o">[</span>node2IP<span class="o">]</span>:/home/tutorial
scp train.txt-00002 test.txt-00002 <span class="o">[</span>node3IP<span class="o">]</span>:/home/tutorial
<span class="c1"># start the job</span>
mpirun -hostfile machines -n <span class="m">3</span>  /home/tutorial/start_mpi_train.sh
</pre></div>
</div>
</div>
</div>
<div class="section" id="cluster-training-using-kubernetes">
<span id="cluster-training-using-kubernetes"></span><h3>Cluster Training Using Kubernetes<a class="headerlink" href="#cluster-training-using-kubernetes" title="Permalink to this headline">¶</a></h3>
<p>The details can be found <span class="xref doc">here</span></p>
</div>
</div>
</div>


           </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
        <a href="../k8s/k8s_en.html" class="btn btn-neutral float-right" title="Paddle On Kubernetes" accesskey="n">Next <span class="fa fa-arrow-circle-right"></span></a>
      
      
        <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>
      
    </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',
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
        };
    </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://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
       
  
    <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>
</html>