Skip to content

P
Paddle

机器未来 / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
切换分支/标签
查找文件 普通视图历史永久链接Permalink
api.html 35.1 KB
Newer Older
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 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 


<!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 Design Doc &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="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">
T
Deploy to GitHub Pages: 7b81055395daa0ce3d02ccb060476c410ee4eb0f  
Travis CI 已提交
65 
        <a class="fork-on-github" href="https://github.com/PaddlePaddle/Paddle" target="_blank"><i class="fa fa-github"></i>Fork me on Github</a>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
66 67 68 69 70 71 72 73 74 75 76 77 
        <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">
T
Deploy to GitHub Pages: 18e5edf6a5b09868dbb083c312fe852cb8c29e55  
Travis CI 已提交
78 
          <li><a href="/">Home</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
79 80 81 82 83 84 85 86 
        </ul>
      </div>
      <div class="doc-module">
        
        <ul>
<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="../howto/index_en.html">HOW TO</a></li>
<li class="toctree-l1"><a class="reference internal" href="../api/index_en.html">API</a></li>
T
Deploy to GitHub Pages: 664691c89af182de3919fc72a7c875311e03893d  
Travis CI 已提交
87 
<li class="toctree-l1"><a class="reference internal" href="../mobile/index_en.html">MOBILE</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 
</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>
<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>
T
Deploy to GitHub Pages: 0c0ff2828ccedb51db23290d6df9e4c83839d6af  
Travis CI 已提交
111 112 
<li class="toctree-l3"><a class="reference internal" href="../getstarted/build_and_install/pip_install_en.html">Install Using pip</a></li>
<li class="toctree-l3"><a class="reference internal" href="../getstarted/build_and_install/docker_install_en.html">Run in Docker Containers</a></li>
T
Deploy to GitHub Pages: 458ffbf4fe7fe6da5e2ea133c027d560815cc3cc  
Travis CI 已提交
113 
<li class="toctree-l3"><a class="reference internal" href="../howto/dev/build_en.html">Build using Docker</a></li>
T
Deploy to GitHub Pages: 0c0ff2828ccedb51db23290d6df9e4c83839d6af  
Travis CI 已提交
114 
<li class="toctree-l3"><a class="reference internal" href="../getstarted/build_and_install/build_from_source_en.html">Build from Sources</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
115 116 117 118 119 120 121 122 123 124 125 
</ul>
</li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../howto/index_en.html">HOW TO</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../howto/usage/cmd_parameter/index_en.html">Set Command-line Parameters</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../howto/usage/cmd_parameter/use_case_en.html">Use Case</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/usage/cmd_parameter/arguments_en.html">Argument Outline</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/usage/cmd_parameter/detail_introduction_en.html">Detail Description</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: 63ffe5250a120ff430469b8d000deb2b031c4881  
Travis CI 已提交
126 
<li class="toctree-l2"><a class="reference internal" href="../howto/usage/cluster/cluster_train_en.html">PaddlePaddle Distributed Training</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
127 
<li class="toctree-l2"><a class="reference internal" href="../howto/dev/new_layer_en.html">Write New Layers</a></li>
T
Deploy to GitHub Pages: f122a5da2f27038b48f6ed607e296d762050e920  
Travis CI 已提交
128 
<li class="toctree-l2"><a class="reference internal" href="../howto/dev/contribute_to_paddle_en.html">Contribute Code</a></li>
T
Deploy to GitHub Pages: 70ee79c2567d5128a9d527b444fcee30f83c02a5  
Travis CI 已提交
129 
<li class="toctree-l2"><a class="reference internal" href="../howto/dev/write_docs_en.html">Contribute Documentation</a></li>
T
Deploy to GitHub Pages: 7c088b529c97f6dc3c28d60dd0510a64bce61312  
Travis CI 已提交
130 131 132 133 
<li class="toctree-l2"><a class="reference internal" href="../howto/deep_model/rnn/index_en.html">RNN Models</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../howto/deep_model/rnn/rnn_config_en.html">RNN Configuration</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
134 135 136 137 
<li class="toctree-l2"><a class="reference internal" href="../howto/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>
T
Deploy to GitHub Pages: 362a14d777ec795f1f6b397c1039ebb594050a10  
Travis CI 已提交
138 139 140 
<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>
T
Deploy to GitHub Pages: 0beda8736ffcac80a95bea20f873d8972d2b5b5e  
Travis CI 已提交
141 
<li class="toctree-l3"><a class="reference internal" href="../api/v2/config/evaluators.html">Evaluators</a></li>
T
Deploy to GitHub Pages: 362a14d777ec795f1f6b397c1039ebb594050a10  
Travis CI 已提交
142 143 144 145 146 147 
<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>
T
Deploy to GitHub Pages: d4d8f742a3bc4366bf410bc9ff29c0c32666bf60  
Travis CI 已提交
148 149 150 151 152 153 
<li class="toctree-l2"><a class="reference internal" href="../api/v2/data.html">Data Reader Interface and DataSets</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/data/data_reader.html">Data Reader Interface</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/data/image.html">Image Interface</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/data/dataset.html">Dataset</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: 362a14d777ec795f1f6b397c1039ebb594050a10  
Travis CI 已提交
154 
<li class="toctree-l2"><a class="reference internal" href="../api/v2/run_logic.html">Training and Inference</a></li>
T
Deploy to GitHub Pages: 1fe05c458fa1d7ff1949759c2a06ed6d19ab8048  
Travis CI 已提交
155 156 157 158 159 160 161 162 163 164 165 166 167 
<li class="toctree-l2"><a class="reference internal" href="../api/v2/fluid.html">Fluid</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/layers.html">Layers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/data_feeder.html">DataFeeder</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/executor.html">Executor</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/initializer.html">Initializer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/evaluator.html">Evaluator</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/nets.html">Nets</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/optimizer.html">Optimizer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/param_attr.html">ParamAttr</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/profiler.html">Profiler</a></li>
<li class="toctree-l3"><a class="reference internal" href="../api/v2/fluid/regularizer.html">Regularizer</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
168 169 
</ul>
</li>
T
Deploy to GitHub Pages: 664691c89af182de3919fc72a7c875311e03893d  
Travis CI 已提交
170 171 
<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>
T
Deploy to GitHub Pages: 78c20e3ed3ca9c03239292ec1c3769697dc976c4  
Travis CI 已提交
172 
<li class="toctree-l2"><a class="reference internal" href="../mobile/cross_compiling_for_ios_en.html">PaddlePaddle Compiling Guide for iOS</a></li>
T
Deploy to GitHub Pages: 664691c89af182de3919fc72a7c875311e03893d  
Travis CI 已提交
173 174 175 
<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>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 
</ul>

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

      

 







<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
    <li>PaddlePaddle Design Doc</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-design-doc">
<span id="paddlepaddle-design-doc"></span><h1>PaddlePaddle Design Doc<a class="headerlink" href="#paddlepaddle-design-doc" title="Permalink to this headline"></a></h1>
<div class="section" id="ingredients">
<span id="ingredients"></span><h2>Ingredients<a class="headerlink" href="#ingredients" title="Permalink to this headline"></a></h2>
<p>As our design principle is starting from the essence: how could we
T
Deploy to GitHub Pages: 63bc2ff876724016062223bc3c966a4db43ae9c5  
Travis CI 已提交
210 
allow users to express and solve their problems as neural networks.
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 
Some essential concepts that our API have to provide include:</p>
<ol class="simple">
<li>A <em>topology</em> is an expression of <em>layers</em>.</li>
<li>A layer could be any kind of computation, including <em>cost</em>.</li>
<li>Some layers have parameters, some don&#8217;t. Most costs don&#8217;t have
parameters.</li>
<li>In some topologies, layers share parameters.  For
example,
<a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/1311#issuecomment-279121850">the network for training a ranking model</a>.</li>
<li>At programming time, users specify topologies and possible sharing
of parameters.  PaddlePaddle can figure out and create parameters
required (and possibly shared) by one or more topologies.</li>
</ol>
</div>
<div class="section" id="starting-from-examples">
<span id="starting-from-examples"></span><h2>Starting from Examples<a class="headerlink" href="#starting-from-examples" title="Permalink to this headline"></a></h2>
<p>As a summarization
of
<a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/1315">our disucssion</a>,
let us present two examples here:</p>
<div class="section" id="example-1-sharing-parameters-between-layers">
<span id="example-1-sharing-parameters-between-layers"></span><h3>Example 1. Sharing Parameters between Layers<a class="headerlink" href="#example-1-sharing-parameters-between-layers" title="Permalink to this headline"></a></h3>
<p>We use
the
<a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/1311#issuecomment-279121850">3-branch ranking</a> model
in this example.  For your convenience, I copy-a-paste the model&#8217;s
topology as follows:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">A</span> <span class="o">-&gt;</span> <span class="n">f</span> <span class="o">-</span>\
<span class="n">Q</span> <span class="o">-&gt;</span> <span class="n">f</span> <span class="o">--&gt;</span> <span class="n">cost</span>
<span class="n">B</span> <span class="o">-&gt;</span> <span class="n">f</span> <span class="o">-/</span>
</pre></div>
</div>
<p>The following program trains the topology including the cost, and then
use the sub-network in the trained topology in inference:</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">f</span><span class="p">(</span><span class="ow">in</span><span class="p">):</span>
    <span class="n">e</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">embedding</span><span class="p">(</span><span class="ow">in</span><span class="p">,</span> <span class="n">parameter_name</span><span class="o">=</span><span class="s2">&quot;embedding&quot;</span><span class="p">)</span>
    <span class="n">o</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">softmax</span><span class="p">(</span><span class="n">e</span><span class="p">,</span> <span class="n">parameter_name</span><span class="o">=</span><span class="s2">&quot;semantic&quot;</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">o</span>

<span class="c1"># Create 3 topologies (subnets), they share parameters because all</span>
<span class="c1"># correspoinding layers have the same parameter names.</span>
<span class="n">fA</span> <span class="o">=</span> <span class="n">f</span><span class="p">(</span><span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">data</span><span class="p">(</span><span class="n">input_name</span><span class="o">=</span><span class="s2">&quot;A&quot;</span><span class="p">))</span>
<span class="n">fB</span> <span class="o">=</span> <span class="n">f</span><span class="p">(</span><span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">data</span><span class="p">(</span><span class="n">input_name</span><span class="o">=</span><span class="s2">&quot;B&quot;</span><span class="p">))</span>
<span class="n">fQ</span> <span class="o">=</span> <span class="n">f</span><span class="p">(</span><span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">data</span><span class="p">(</span><span class="n">input_name</span><span class="o">=</span><span class="s2">&quot;Q&quot;</span><span class="p">))</span>

<span class="n">topology</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">less_than</span><span class="p">(</span>
               <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">cross_entropy</span><span class="p">(</span><span class="n">fA</span><span class="p">,</span> <span class="n">fQ</span><span class="p">),</span>
               <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">corss_entropy</span><span class="p">(</span><span class="n">fB</span><span class="p">,</span> <span class="n">fQ</span><span class="p">))</span>

<span class="c1"># Derive parameters required in topology and create them in model.</span>
<span class="n">parameters</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">parameters</span><span class="o">.</span><span class="n">create</span><span class="p">(</span><span class="n">topology</span><span class="p">)</span>

<span class="c1"># Estimate parameters used in topology from data.</span>
<span class="n">paddle</span><span class="o">.</span><span class="n">train</span><span class="p">(</span><span class="n">topology</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">reader</span><span class="o">=</span><span class="n">read_ranking_model_data</span><span class="p">)</span>

<span class="c1"># Inference using fA (or fB or fC, as they share their parameters).</span>
<span class="p">[</span><span class="n">testA</span><span class="p">,</span> <span class="n">testB</span><span class="p">,</span> <span class="n">testQ</span><span class="p">]</span> <span class="o">=</span> <span class="n">read_ranking_model_data</span><span class="p">()</span>
<span class="k">print</span> <span class="s2">&quot;The sematic-vector of testA: &quot;</span><span class="p">,</span> <span class="n">paddle</span><span class="o">.</span><span class="n">infer</span><span class="p">(</span><span class="n">fA</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">testA</span><span class="p">)</span>
</pre></div>
</div>
</div>
<div class="section" id="example-2-sharing-parameters-between-models">
<span id="example-2-sharing-parameters-between-models"></span><h3>Example 2. Sharing Parameters between &#8220;Models&#8221;<a class="headerlink" href="#example-2-sharing-parameters-between-models" title="Permalink to this headline"></a></h3>
<p>We use <a class="reference external" href="https://github.com/PaddlePaddle/book/tree/develop/gan">GAN</a> in
this example.  In the following example program, <code class="docutils literal"><span class="pre">d0</span></code> and <code class="docutils literal"><span class="pre">d1</span></code>
correspond to the two networks in the following figure:</p>
<p><img src="https://github.com/wangyang59/book/raw/00036f4b0da5225041a6824587c1a01cf20159b1/gan/image/gan_ig.png" width=400 /></p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">G</span><span class="p">(</span><span class="ow">in</span><span class="p">):</span>
    <span class="c1"># over-simplified example as G has only one layers:</span>
    <span class="k">return</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">fc</span><span class="p">(</span><span class="ow">in</span><span class="p">,</span> <span class="n">parameter_name</span><span class="o">=</span><span class="s2">&quot;G&quot;</span><span class="p">)</span>

<span class="k">def</span> <span class="nf">D</span><span class="p">(</span><span class="ow">in</span><span class="p">);</span>
    <span class="c1"># again, over-simplified:</span>
    <span class="k">return</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">fc</span><span class="p">(</span><span class="ow">in</span><span class="p">,</span> <span class="n">parameter_name</span><span class="o">=</span><span class="s2">&quot;D&quot;</span><span class="p">)</span>

<span class="c1"># Construct the first topology, which contains both D and G.</span>
<span class="c1"># By learning this topology, we update parameters of G.</span>
<span class="n">d0</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">should_be_false</span><span class="p">(</span><span class="n">D</span><span class="p">(</span><span class="n">G</span><span class="p">(</span><span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">data</span><span class="p">())))</span>

<span class="c1"># Construct a second topology d1, which contains only D. By</span>
<span class="c1"># training this topology, we update parameters of D.  Note</span>
<span class="c1"># that d1 share parameters with d0.</span>
<span class="n">d1</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">should_be_true</span><span class="p">(</span><span class="n">D</span><span class="p">(</span><span class="n">paddle</span><span class="o">.</span><span class="n">layer</span><span class="o">.</span><span class="n">data</span><span class="p">()))</span>

<span class="c1"># Create parameters from a list of multiple topologies (models) for</span>
<span class="c1"># the chance to share parameters between these topologies.</span>
<span class="n">parameters</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">parameters</span><span class="o">.</span><span class="n">create</span><span class="p">([</span><span class="n">d0</span><span class="p">,</span> <span class="n">d1</span><span class="p">])</span>

<span class="c1"># Iterative training of GAN.</span>
<span class="k">for</span> <span class="o">...</span><span class="p">:</span>
    <span class="n">train</span><span class="p">(</span><span class="n">d0</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">reader</span><span class="o">=</span><span class="n">read_from_rng</span><span class="p">,</span> <span class="n">immutable_parameters</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;D&quot;</span><span class="p">})</span>
    <span class="n">train</span><span class="p">(</span><span class="n">d1</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">reader</span><span class="o">=</span><span class="n">read_from_realistic_images</span><span class="p">)</span>

<span class="c1"># Use d1 for inference:</span>
<span class="k">print</span> <span class="s2">&quot;D thinks a batch of images are realistic &quot;</span><span class="p">,</span> <span class="n">infer</span><span class="p">(</span><span class="n">d1</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">read_mnist_images</span><span class="p">)</span>
</pre></div>
</div>
</div>
<div class="section" id="summarization">
<span id="summarization"></span><h3>Summarization<a class="headerlink" href="#summarization" title="Permalink to this headline"></a></h3>
<p>Above two programs reveal some important design concerns:</p>
<ol class="simple">
<li>Users describe a topology as an expression of layers.  Every layer
has a <em>parameter name</em>.  If the users don&#8217;t specify it explicitly, it&#8217;s automatically generated as a unique name.  By
specifying the parameter name, users can specify the sharing of
parameters between layers and even between topologies.</li>
<li><code class="docutils literal"><span class="pre">paddle.parameters.create</span></code> figures out parameters required by one
or more topologies from parameter names of layers.  It creates these
parameters and returns a <code class="docutils literal"><span class="pre">ParameterSet</span></code> object, which is in essence
a map from <em>parameter names</em> to <em>parameters</em>.</li>
<li>At training and inference time, <code class="docutils literal"><span class="pre">paddle.train</span></code> and <code class="docutils literal"><span class="pre">paddle.infer</span></code>
requires both a topology and the parameter set that holds the parameters of that topology.  There are some reasons:<ol>
<li>This prevents users from forgetting to call
<code class="docutils literal"><span class="pre">paddle.parameters.create</span></code>.</li>
<li><code class="docutils literal"><span class="pre">paddle.train</span></code> needs to know which parameter set to update.</li>
<li>Users could load another (pre-trained) parameter set and use it
with a topology in <code class="docutils literal"><span class="pre">train.infer</span></code>.</li>
</ol>
</li>
<li>By specifying the <code class="docutils literal"><span class="pre">immutable_parameters</span></code> parameter of
<code class="docutils literal"><span class="pre">paddle.train</span></code>, we can forbid the update of these parameters.</li>
</ol>
</div>
</div>
<div class="section" id="reader">
<span id="reader"></span><h2>Reader<a class="headerlink" href="#reader" title="Permalink to this headline"></a></h2>
<p>Not all programming frameworks allow users to define I/O functions.
An example is Google MapReduce, which can only read from text,
SSTable, and RecordIO files.  Hadoop MapReduce allows users to define
readers and writers by deriving from base classes <code class="docutils literal"><span class="pre">Reader</span></code> and
<code class="docutils literal"><span class="pre">Writer</span></code>.  The former is less flexible but also less error-prone.  We
decide to provide the flexibility to users to define their readers.</p>
<p>There are some open questions here:</p>
<ol class="simple">
<li><strong>Should a reader return a Python dictionary?</strong></li>
<li><strong>How to map multiple outputs from a reader to multiple data layers?</strong></li>
<li><strong>How to easily compose some existing readers to read more data and
feed a topology with more data layers?</strong></li>
</ol>
</div>
<div class="section" id="training">
<span id="training"></span><h2>Training<a class="headerlink" href="#training" title="Permalink to this headline"></a></h2>
<p>The recommended way to training a model is to call <code class="docutils literal"><span class="pre">paddle.train</span></code>,
which simply calls <code class="docutils literal"><span class="pre">paddle.trainer.Default</span></code>, a global variable of
type <code class="docutils literal"><span class="pre">paddle.trainer.SGD</span></code>.  Equivalently, we can do</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">opt</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">.</span><span class="n">trainer</span><span class="o">.</span><span class="n">SGD</span><span class="p">(</span><span class="o">...</span><span class="p">,</span> <span class="n">paddle</span><span class="o">.</span><span class="n">updater</span><span class="o">.</span><span class="n">Adam</span><span class="p">(</span><span class="o">...</span><span class="p">))</span>
<span class="n">opt</span><span class="o">.</span><span class="n">train</span><span class="p">(</span><span class="n">topology</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">reader</span><span class="o">=</span><span class="n">read</span><span class="p">,</span> <span class="o">...</span><span class="p">)</span>
</pre></div>
</div>
<div class="section" id="updater">
<span id="updater"></span><h3>Updater<a class="headerlink" href="#updater" title="Permalink to this headline"></a></h3>
<p>Please be aware that a trainer can accept an updater as its data
member, where an updater is a class derived from
<code class="docutils literal"><span class="pre">paddle.trainer.Updater</span></code>.  This is to make it easier to customize
trainers, as discussed
<a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/1319">here</a>.</p>
</div>
<div class="section" id="event-handler">
<span id="event-handler"></span><h3>Event Handler<a class="headerlink" href="#event-handler" title="Permalink to this headline"></a></h3>
<p><code class="docutils literal"><span class="pre">paddle.train</span></code> and <code class="docutils literal"><span class="pre">paddle.trainer.XXX.train</span></code> take an optional
parameter <code class="docutils literal"><span class="pre">event_handler</span></code>, which should be either <code class="docutils literal"><span class="pre">None</span></code> or a function
that handle some events:</p>
<ol class="simple">
<li>BeginTraining</li>
<li>EndTraining</li>
<li>BeginIteration</li>
<li>EndIteration</li>
<li>BeginPass</li>
<li>EndPass</li>
</ol>
<p>where EndPass is sent if and only if the reader yields
<code class="docutils literal"><span class="pre">end_pass=True</span></code>.</p>
<p>An example as follows:</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">event_handler</span><span class="p">(</span><span class="n">event</span><span class="p">):</span>
    <span class="k">if</span> <span class="n">ininstance</span><span class="p">(</span><span class="n">event</span><span class="p">,</span> <span class="n">paddle</span><span class="o">.</span><span class="n">event</span><span class="o">.</span><span class="n">EndIteration</span><span class="p">):</span>
        <span class="k">print</span> <span class="n">paddle</span><span class="o">.</span><span class="n">test</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>

<span class="n">paddle</span><span class="o">.</span><span class="n">train</span><span class="p">(</span><span class="n">topology</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">reader</span><span class="p">,</span> <span class="n">event_handler</span><span class="p">)</span>
</pre></div>
</div>
<p>If we are writing a PaddlePaddle program in and for iPython/Jypyter,
we can use metaplotlib in the event handler to plot a curve of
cost/error versus iterations, as shown
<a class="reference external" href="https://blog.dominodatalab.com/interactive-dashboards-in-jupyter/">here</a>.</p>
</div>
<div class="section" id="distributed-training">
<span id="distributed-training"></span><h3>Distributed Training<a class="headerlink" href="#distributed-training" title="Permalink to this headline"></a></h3>
<p>If users want to do distributed training on a cluster, s/he should
call <code class="docutils literal"><span class="pre">paddle.dist_train</span></code> and provides access tokens to the cluster as
a parameter.</p>
<p>For example, if the user has a TLS certificate that allows him to
access a Kubernetes cluster, s/he should be able to call</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">paddle</span><span class="o">.</span><span class="n">dist_train</span><span class="p">(</span><span class="n">model</span><span class="p">,</span>
                  <span class="n">trainer</span><span class="o">=</span><span class="n">paddle</span><span class="o">.</span><span class="n">trainer</span><span class="o">.</span><span class="n">SGD</span><span class="p">(</span><span class="o">...</span><span class="p">,</span>
                                             <span class="n">paddle</span><span class="o">.</span><span class="n">updater</span><span class="o">.</span><span class="n">Adam</span><span class="p">(</span><span class="o">...</span><span class="p">)),</span>
                  <span class="n">reader</span><span class="o">=</span><span class="n">read</span><span class="p">,</span>
                  <span class="n">k8s_user</span><span class="o">=</span><span class="s2">&quot;yi&quot;</span><span class="p">,</span>
                  <span class="n">k8s_token</span><span class="o">=</span><span class="s2">&quot;kube_cluster_tls.pem&quot;</span><span class="p">,</span>
                  <span class="n">k8s_job</span><span class="o">=</span><span class="s2">&quot;hello&quot;</span><span class="p">,</span>
                  <span class="n">num_parameter_servers</span><span class="o">=</span><span class="mi">15</span><span class="p">)</span>
</pre></div>
</div>
T
Deploy to GitHub Pages: 63bc2ff876724016062223bc3c966a4db43ae9c5  
Travis CI 已提交
413 
<p>The pseudo code of <code class="docutils literal"><span class="pre">paddle.dist_train</span></code> is as follows:</p>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">dist_train</span><span class="p">(</span><span class="n">topology</span><span class="p">,</span> <span class="n">parameters</span><span class="p">,</span> <span class="n">trainer</span><span class="p">,</span> <span class="n">reader</span><span class="p">,</span> <span class="o">...</span><span class="p">):</span>
    <span class="k">if</span> <span class="n">os</span><span class="o">.</span><span class="n">getenv</span><span class="p">(</span><span class="s2">&quot;KUBERNETES_SERVICE_HOST&quot;</span><span class="p">)</span> <span class="o">==</span> <span class="bp">None</span><span class="p">:</span>
        <span class="n">image_name</span> <span class="o">=</span> <span class="n">k8s_user</span> <span class="o">+</span> <span class="s1">&#39;/&#39;</span> <span class="o">+</span> <span class="n">k8s_job</span>
        <span class="n">docker_build</span><span class="p">(</span><span class="n">image_name</span><span class="p">)</span>
        <span class="n">docker_push</span><span class="p">()</span>
        <span class="n">kube_ctrl_start_job</span><span class="p">(</span><span class="n">image_name</span><span class="p">,</span> <span class="n">k8s_user</span><span class="p">,</span> <span class="n">k8s_token</span><span class="p">)</span>
    <span class="k">else</span><span class="p">:</span>
        <span class="n">rank</span> <span class="o">=</span> <span class="n">kube_list_containers_in_job_and_return_current_containers_rank</span><span class="p">()</span>
        <span class="k">if</span> <span class="n">rank</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
            <span class="n">master</span><span class="p">()</span>
        <span class="k">elif</span> <span class="n">rank</span> <span class="o">&lt;</span> <span class="mi">15</span><span class="p">:</span>
            <span class="n">parameter_server</span><span class="p">()</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="n">trainer</span><span class="o">.</span><span class="n">train</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">reader</span><span class="o">=</span><span class="n">read</span><span class="p">)</span>
</pre></div>
</div>
<p>Please be aware that if a process is running on the Kubernetes
cluster, it will have some environment variables pre-defined.</p>
<p>If <code class="docutils literal"><span class="pre">dist_train</span></code> doesn&#8217;t see these environment variables, it knows
that it&#8217;s running on users&#8217; personal computer, and it should work as a
<em>launcher</em>.  Otherwise, it knows that it&#8217;s running on the cluster and
need to figure out its role as either the master, or a trainer, or a
parameter server.</p>
</div>
</div>
</div>


           </div>
          </div>
          <footer>
  

  <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',
T
Deploy to GitHub Pages: 18e5edf6a5b09868dbb083c312fe852cb8c29e55  
Travis CI 已提交
476 477 
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
478 479 480 481 482 
        };
    </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>
T
Deploy to GitHub Pages: 02a8795fcfdc3a8712b24cd0ca775c8d9920d5f4  
Travis CI 已提交
483 
      <script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
484 485 486 487 488 489 490 491 492 493 494 495 496 
       
  

  
  
    <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>
T
Deploy to GitHub Pages: 299e959a190017689d18af70d300f681745f4c67  
Travis CI 已提交
497 
</html>