Skip to content

P
Paddle

Crayon鑫 / 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
index_cn.html 28.2 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 


<!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>图像分类教程 &mdash; PaddlePaddle  文档</title>
  

  
  

  

  
  
    

  

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

  
  
        <link rel="index" title="索引"
              href="../../genindex.html"/>
        <link rel="search" title="搜索" href="../../search.html"/>
T
Deploy to GitHub Pages: 18e5edf6a5b09868dbb083c312fe852cb8c29e55  
Travis CI 已提交
35 
    <link rel="top" title="PaddlePaddle  文档" href="../../index.html"/>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
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 

  <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 
        </ul>
      </div>
      <div class="doc-module">
T
Deploy to GitHub Pages: 18e5edf6a5b09868dbb083c312fe852cb8c29e55  
Travis CI 已提交
83 
        <ul>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../howto/index_cn.html">进阶指南</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../api/index_cn.html">API</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</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">
T
Deploy to GitHub Pages: 18e5edf6a5b09868dbb083c312fe852cb8c29e55  
Travis CI 已提交
108 
          <ul>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
109 110 
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/build_and_install/index_cn.html">安装与编译</a><ul>
T
Deploy to GitHub Pages: 5914c5c7488d48a412c1503ef96de2c1a6b06ed6  
Travis CI 已提交
111 
<li class="toctree-l3"><a class="reference internal" href="../../getstarted/build_and_install/docker_install_cn.html">PaddlePaddle的Docker容器使用方式</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
112 113 114 115 
<li class="toctree-l3"><a class="reference internal" href="../../getstarted/build_and_install/ubuntu_install_cn.html">Ubuntu部署PaddlePaddle</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../getstarted/build_and_install/cmake/build_from_source_cn.html">PaddlePaddle的编译选项</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: ab5ad2215d037ed6985f0830669c4bde8944a944  
Travis CI 已提交
116 
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/concepts/use_concepts_cn.html">基本使用概念</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
117 118 119 120 121 122 123 124 125 126 127 128 129 
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../howto/index_cn.html">进阶指南</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../howto/usage/cmd_parameter/index_cn.html">设置命令行参数</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../howto/usage/cmd_parameter/use_case_cn.html">使用案例</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/usage/cmd_parameter/arguments_cn.html">参数概述</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/usage/cmd_parameter/detail_introduction_cn.html">细节描述</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/usage/cluster/cluster_train_cn.html">运行分布式训练</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/usage/k8s/k8s_basis_cn.html">Kubernetes 简介</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/usage/k8s/k8s_cn.html">Kubernetes单机训练</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/usage/k8s/k8s_distributed_cn.html">Kubernetes分布式训练</a></li>
T
Deploy to GitHub Pages: 111812fa3b41797a3940bbcd51e51ff21b799ffe  
Travis CI 已提交
130 
<li class="toctree-l2"><a class="reference internal" href="../../howto/dev/build_cn.html">编译PaddlePaddle和运行单元测试</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
131 132 133 
<li class="toctree-l2"><a class="reference internal" href="../../howto/dev/write_docs_cn.html">如何贡献/修改文档</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/dev/contribute_to_paddle_cn.html">如何贡献代码</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/deep_model/rnn/index_cn.html">RNN相关模型</a><ul>
T
Deploy to GitHub Pages: 7c088b529c97f6dc3c28d60dd0510a64bce61312  
Travis CI 已提交
134 
<li class="toctree-l3"><a class="reference internal" href="../../howto/deep_model/rnn/rnn_config_cn.html">RNN配置</a></li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
135 136 137 138 139 140 141 142 
<li class="toctree-l3"><a class="reference internal" href="../../howto/deep_model/rnn/recurrent_group_cn.html">Recurrent Group教程</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/deep_model/rnn/hierarchical_layer_cn.html">支持双层序列作为输入的Layer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/deep_model/rnn/hrnn_rnn_api_compare_cn.html">单双层RNN API对比介绍</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/optimization/gpu_profiling_cn.html">GPU性能分析与调优</a></li>
</ul>
</li>
T
Deploy to GitHub Pages: 5f2cbce4190fb83883fed45668b5e4d87a577589  
Travis CI 已提交
143 
<li class="toctree-l1"><a class="reference internal" href="../../api/index_cn.html">API</a><ul>
T
Deploy to GitHub Pages: 362a14d777ec795f1f6b397c1039ebb594050a10  
Travis CI 已提交
144 145 146 
<li class="toctree-l2"><a class="reference internal" href="../../api/v2/model_configs.html">模型配置</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
147 
<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
148 149 150 151 152 153 154 155 
<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">数据访问</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../api/v2/run_logic.html">训练与应用</a></li>
T
Deploy to GitHub Pages: 5f2cbce4190fb83883fed45668b5e4d87a577589  
Travis CI 已提交
156 157 
</ul>
</li>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</a></li>
</ul>

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

      

 







<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
    <li>图像分类教程</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="">
<span id="id1"></span><h1>图像分类教程<a class="headerlink" href="#" title="永久链接至标题"></a></h1>
<p>在本教程中，我们将使用CIFAR-10数据集训练一个卷积神经网络，并使用这个神经网络来对图片进行分类。如下图所示，卷积神经网络可以辨识图片中的主体，并给出分类结果。
<center><img alt="Image Classification" src="../../_images/image_classification.png" /></center></p>
<div class="section" id="">
<span id="id2"></span><h2>数据准备<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<p>首先下载CIFAR-10数据集。下面是CIFAR-10数据集的官方网址：</p>
<p><a class="reference external" href="https://www.cs.toronto.edu/~kriz/cifar.html">https://www.cs.toronto.edu/~kriz/cifar.html</a></p>
<p>我们准备了一个脚本，可以用于从官方网站上下载CIFAR-10数据集，转为jpeg文件并存入特定的目录。使用这个脚本前请确认已经安装了pillow及相关依赖模块。可以参照下面的命令进行安装：</p>
<ol class="simple">
<li>安装pillow</li>
</ol>
<div class="highlight-bash"><div class="highlight"><pre><span></span>sudo apt-get install libjpeg-dev
pip install pillow
</pre></div>
</div>
<ol class="simple">
<li>下载数据集</li>
</ol>
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nb">cd</span> demo/image_classification/data/
sh download_cifar.sh
</pre></div>
</div>
<p>CIFAR-10数据集包含60000张32x32的彩色图片。图片分为10类，每个类包含6000张。其中50000张图片作为训练集，10000张作为测试集。</p>
<p>下图展示了所有的图片类别，每个类别中随机抽取了10张图片。
<center><img alt="Image Classification" src="../../_images/cifar.png" /></center></p>
<p>脚本运行完成后，我们应当会得到一个名为cifar-out的文件夹，其下子文件夹的结构如下</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">train</span>
<span class="o">---</span><span class="n">airplane</span>
<span class="o">---</span><span class="n">automobile</span>
<span class="o">---</span><span class="n">bird</span>
<span class="o">---</span><span class="n">cat</span>
<span class="o">---</span><span class="n">deer</span>
<span class="o">---</span><span class="n">dog</span>
<span class="o">---</span><span class="n">frog</span>
<span class="o">---</span><span class="n">horse</span>
<span class="o">---</span><span class="n">ship</span>
<span class="o">---</span><span class="n">truck</span>
<span class="n">test</span>
<span class="o">---</span><span class="n">airplane</span>
<span class="o">---</span><span class="n">automobile</span>
<span class="o">---</span><span class="n">bird</span>
<span class="o">---</span><span class="n">cat</span>
<span class="o">---</span><span class="n">deer</span>
<span class="o">---</span><span class="n">dog</span>
<span class="o">---</span><span class="n">frog</span>
<span class="o">---</span><span class="n">horse</span>
<span class="o">---</span><span class="n">ship</span>
<span class="o">---</span><span class="n">truck</span>
</pre></div>
</div>
<p>cifar-out下包含<code class="docutils literal"><span class="pre">train</span></code>和<code class="docutils literal"><span class="pre">test</span></code>两个文件夹，其中分别包含了CIFAR-10中的训练集和测试集。这两个文件夹下各自有10个子文件夹，每个子文件夹下存储相应分类的图片。将图片按照上述结构存储好之后，我们就可以着手对分类模型进行训练了。</p>
</div>
<div class="section" id="">
<span id="id3"></span><h2>预处理<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<p>数据下载之后，还需要进行预处理，将数据转换为Paddle的格式。我们可以通过如下命令进行预处理工作：</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">cd</span> <span class="n">demo</span><span class="o">/</span><span class="n">image_classification</span><span class="o">/</span>
<span class="n">sh</span> <span class="n">preprocess</span><span class="o">.</span><span class="n">sh</span>
</pre></div>
</div>
<p>其中<code class="docutils literal"><span class="pre">preprocess.sh</span></code> 调用 <code class="docutils literal"><span class="pre">./demo/image_classification/preprocess.py</span></code> 对图片进行预处理</p>
<div class="highlight-sh"><div class="highlight"><pre><span></span><span class="nb">export</span> <span class="nv">PYTHONPATH</span><span class="o">=</span><span class="nv">$PYTHONPATH</span>:../../
<span class="nv">data_dir</span><span class="o">=</span>./data/cifar-out
python preprocess.py -i <span class="nv">$data_dir</span> -s <span class="m">32</span> -c <span class="m">1</span>
</pre></div>
</div>
<p><code class="docutils literal"><span class="pre">./demo/image_classification/preprocess.py</span></code> 使用如下参数：</p>
<ul class="simple">
<li><code class="docutils literal"><span class="pre">-i</span></code> 或 <code class="docutils literal"><span class="pre">--input</span></code> 给出输入数据所在路径；</li>
<li><code class="docutils literal"><span class="pre">-s</span></code> 或 <code class="docutils literal"><span class="pre">--size</span></code> 给出图片尺寸；</li>
<li><code class="docutils literal"><span class="pre">-c</span></code> 或 <code class="docutils literal"><span class="pre">--color</span></code> 标示图片是彩色图或灰度图</li>
</ul>
</div>
<div class="section" id="">
<span id="id4"></span><h2>模型训练<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<p>在开始训练之前，我们需要先创建一个模型配置文件。下面我们给出了一个配置示例。<strong>注意</strong>，这里的列出的和<code class="docutils literal"><span class="pre">vgg_16_cifar.py</span></code>文件稍有差别，因为该文件可适用于预测。</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">paddle.trainer_config_helpers</span> <span class="kn">import</span> <span class="o">*</span>
<span class="n">data_dir</span><span class="o">=</span><span class="s1">&#39;data/cifar-out/batches/&#39;</span>
<span class="n">meta_path</span><span class="o">=</span><span class="n">data_dir</span><span class="o">+</span><span class="s1">&#39;batches.meta&#39;</span>
<span class="n">args</span> <span class="o">=</span> <span class="p">{</span><span class="s1">&#39;meta&#39;</span><span class="p">:</span><span class="n">meta_path</span><span class="p">,</span> <span class="s1">&#39;mean_img_size&#39;</span><span class="p">:</span> <span class="mi">32</span><span class="p">,</span>
        <span class="s1">&#39;img_size&#39;</span><span class="p">:</span> <span class="mi">32</span><span class="p">,</span> <span class="s1">&#39;num_classes&#39;</span><span class="p">:</span> <span class="mi">10</span><span class="p">,</span>
        <span class="s1">&#39;use_jpeg&#39;</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">&#39;color&#39;</span><span class="p">:</span> <span class="s2">&quot;color&quot;</span><span class="p">}</span>
<span class="n">define_py_data_sources2</span><span class="p">(</span><span class="n">train_list</span><span class="o">=</span><span class="n">data_dir</span><span class="o">+</span><span class="s2">&quot;train.list&quot;</span><span class="p">,</span>
                        <span class="n">test_list</span><span class="o">=</span><span class="n">data_dir</span><span class="o">+</span><span class="s1">&#39;test.list&#39;</span><span class="p">,</span>
                        <span class="n">module</span><span class="o">=</span><span class="s1">&#39;image_provider&#39;</span><span class="p">,</span>
                        <span class="n">obj</span><span class="o">=</span><span class="s1">&#39;processData&#39;</span><span class="p">,</span>
                        <span class="n">args</span><span class="o">=</span><span class="n">args</span><span class="p">)</span>
<span class="n">settings</span><span class="p">(</span>
    <span class="n">batch_size</span> <span class="o">=</span> <span class="mi">128</span><span class="p">,</span>
    <span class="n">learning_rate</span> <span class="o">=</span> <span class="mf">0.1</span> <span class="o">/</span> <span class="mf">128.0</span><span class="p">,</span>
    <span class="n">learning_method</span> <span class="o">=</span> <span class="n">MomentumOptimizer</span><span class="p">(</span><span class="mf">0.9</span><span class="p">),</span>
    <span class="n">regularization</span> <span class="o">=</span> <span class="n">L2Regularization</span><span class="p">(</span><span class="mf">0.0005</span> <span class="o">*</span> <span class="mi">128</span><span class="p">))</span>

<span class="n">img</span> <span class="o">=</span> <span class="n">data_layer</span><span class="p">(</span><span class="n">name</span><span class="o">=</span><span class="s1">&#39;image&#39;</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="mi">3</span><span class="o">*</span><span class="mi">32</span><span class="o">*</span><span class="mi">32</span><span class="p">)</span>
<span class="n">lbl</span> <span class="o">=</span> <span class="n">data_layer</span><span class="p">(</span><span class="n">name</span><span class="o">=</span><span class="s2">&quot;label&quot;</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="mi">10</span><span class="p">)</span>
<span class="c1"># small_vgg is predined in trainer_config_helpers.network</span>
<span class="n">predict</span> <span class="o">=</span> <span class="n">small_vgg</span><span class="p">(</span><span class="n">input_image</span><span class="o">=</span><span class="n">img</span><span class="p">,</span> <span class="n">num_channels</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
<span class="n">outputs</span><span class="p">(</span><span class="n">classification_cost</span><span class="p">(</span><span class="nb">input</span><span class="o">=</span><span class="n">predict</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="n">lbl</span><span class="p">))</span>
</pre></div>
</div>
<p>在第一行中我们载入用于定义网络的函数。</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">paddle.trainer_config_helpers</span> <span class="kn">import</span> <span class="o">*</span>
</pre></div>
</div>
<p>之后定义的<code class="docutils literal"><span class="pre">define_py_data_sources2</span></code>使用Python数据提供器，其中 <code class="docutils literal"><span class="pre">args</span></code>将在<code class="docutils literal"><span class="pre">image_provider.py</span></code>进行使用，该文件负责产生图片数据并传递给Paddle系统</p>
<ul class="simple">
<li><code class="docutils literal"><span class="pre">meta</span></code>: 训练集平均值。</li>
<li><code class="docutils literal"><span class="pre">mean_img_size</span></code>: 平均特征图的高度及宽度。</li>
<li><code class="docutils literal"><span class="pre">img_size</span></code>：输入图片的高度及宽度。</li>
<li><code class="docutils literal"><span class="pre">num_classes</span></code>：类别个数。</li>
<li><code class="docutils literal"><span class="pre">use_jpeg</span></code>：处理过程中数据存储格式。</li>
<li><code class="docutils literal"><span class="pre">color</span></code>：标示是否为彩色图片。</li>
</ul>
<p><code class="docutils literal"><span class="pre">settings</span></code>用于设置训练算法。在下面的例子中，learning rate被设置为0.1除以batch size，而weight decay则为0.0005乘以batch size。</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="n">settings</span><span class="p">(</span>
   <span class="n">batch_size</span> <span class="o">=</span> <span class="mi">128</span><span class="p">,</span>
   <span class="n">learning_rate</span> <span class="o">=</span> <span class="mf">0.1</span> <span class="o">/</span> <span class="mf">128.0</span><span class="p">,</span>
   <span class="n">learning_method</span> <span class="o">=</span> <span class="n">MomentumOptimizer</span><span class="p">(</span><span class="mf">0.9</span><span class="p">),</span>
   <span class="n">regularization</span> <span class="o">=</span> <span class="n">L2Regularization</span><span class="p">(</span><span class="mf">0.0005</span> <span class="o">*</span> <span class="mi">128</span><span class="p">)</span>
<span class="p">)</span>
</pre></div>
</div>
<p><code class="docutils literal"><span class="pre">small_vgg</span></code>定义了网络结构。这里我们使用的是一个小的VGG网络。关于VGG卷积神经网络的描述可以参考：<a class="reference external" href="http://www.robots.ox.ac.uk/~vgg/research/very_deep/">http://www.robots.ox.ac.uk/~vgg/research/very_deep/</a>。</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="c1"># small_vgg is predined in trainer_config_helpers.network</span>
<span class="n">predict</span> <span class="o">=</span> <span class="n">small_vgg</span><span class="p">(</span><span class="n">input_image</span><span class="o">=</span><span class="n">img</span><span class="p">,</span> <span class="n">num_channels</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
</pre></div>
</div>
<p>配置创建完毕后，可以运行脚本train.sh来训练模型。</p>
<div class="highlight-bash"><div class="highlight"><pre><span></span><span class="nv">config</span><span class="o">=</span>vgg_16_cifar.py
<span class="nv">output</span><span class="o">=</span>./cifar_vgg_model
<span class="nv">log</span><span class="o">=</span>train.log

paddle train <span class="se">\</span>
--config<span class="o">=</span><span class="nv">$config</span> <span class="se">\</span>
--dot_period<span class="o">=</span><span class="m">10</span> <span class="se">\</span>
--log_period<span class="o">=</span><span class="m">100</span> <span class="se">\</span>
--test_all_data_in_one_period<span class="o">=</span><span class="m">1</span> <span class="se">\</span>
--use_gpu<span class="o">=</span><span class="m">1</span> <span class="se">\</span>
--save_dir<span class="o">=</span><span class="nv">$output</span> <span class="se">\</span>
<span class="m">2</span>&gt;<span class="p">&amp;</span><span class="m">1</span> <span class="p">|</span> tee <span class="nv">$log</span>

python -m paddle.utils.plotcurve -i <span class="nv">$log</span> &gt; plot.png
</pre></div>
</div>
<ul class="simple">
<li>这里我们使用的是GPU模式进行训练。如果你没有GPU环境，可以设置<code class="docutils literal"><span class="pre">use_gpu=0</span></code>。</li>
<li><code class="docutils literal"><span class="pre">./demo/image_classification/vgg_16_cifar.py</span></code>是网络和数据配置文件。各项参数的详细说明可以在命令行参数相关文档中找到。</li>
<li>脚本<code class="docutils literal"><span class="pre">plotcurve.py</span></code>依赖于python的<code class="docutils literal"><span class="pre">matplotlib</span></code>模块。因此如果这个脚本运行失败，也许是因为需要安装<code class="docutils literal"><span class="pre">matplotlib</span></code>。
在训练完成后，训练及测试误差曲线图会被<code class="docutils literal"><span class="pre">plotcurve.py</span></code>脚本保存在 <code class="docutils literal"><span class="pre">plot.png</span></code>中。下面是一个误差曲线图的示例：</li>
</ul>
<p><center><img alt="Training and testing curves." src="../../_images/plot.png" /></center></p>
</div>
<div class="section" id="">
<span id="id5"></span><h2>预测<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<p>在训练完成后，模型及参数会被保存在路径<code class="docutils literal"><span class="pre">./cifar_vgg_model/pass-%05d</span></code>下。例如第300个pass的模型会被保存在<code class="docutils literal"><span class="pre">./cifar_vgg_model/pass-00299</span></code>。</p>
<p>要对一个图片的进行分类预测，我们可以使用<code class="docutils literal"><span class="pre">predict.sh</span></code>，该脚本将输出预测分类的标签：</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">sh</span> <span class="n">predict</span><span class="o">.</span><span class="n">sh</span>
</pre></div>
</div>
<p>predict.sh:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span>model=cifar_vgg_model/pass-00299/
image=data/cifar-out/test/airplane/seaplane_s_000978.png
use_gpu=1
python prediction.py $model $image $use_gpu
</pre></div>
</div>
</div>
<div class="section" id="">
<span id="id6"></span><h2>练习<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<p>在CUB-200数据集上使用VGG模型训练一个鸟类图片分类模型。相关的鸟类数据集可以从如下地址下载，其中包含了200种鸟类的照片（主要来自北美洲）。</p>
<p><a class="reference external" href="http://www.vision.caltech.edu/visipedia/CUB-200.html">http://www.vision.caltech.edu/visipedia/CUB-200.html</a></p>
</div>
<div class="section" id="">
<span id="id7"></span><h2>细节探究<a class="headerlink" href="#" title="永久链接至标题"></a></h2>
<div class="section" id="">
<span id="id8"></span><h3>卷积神经网络<a class="headerlink" href="#" title="永久链接至标题"></a></h3>
<p>卷积神经网络是一种使用卷积层的前向神经网络，很适合构建用于理解图片内容的模型。一个典型的神经网络如下图所示：</p>
<p><img alt="Convolutional Neural Network" src="../../_images/lenet.png" /></p>
<p>一个卷积神经网络包含如下层：</p>
<ul class="simple">
<li>卷积层：通过卷积操作从图片或特征图中提取特征</li>
<li>池化层：使用max-pooling对特征图下采样</li>
<li>全连接层：使输入层到隐藏层的神经元是全部连接的。</li>
</ul>
<p>卷积神经网络在图片分类上有着惊人的性能，这是因为它发掘出了图片的两类重要信息：局部关联性质和空间不变性质。通过交替使用卷积和池化处理， 卷积神经网络能够很好的表示这两类信息。</p>
<p>关于如何定义网络中的层，以及如何在层之间进行连接，请参考Layer文档。</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
413 414 
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
415 416 417 418 419 420 
        };
    </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="../../_static/translations.js"></script>
T
Deploy to GitHub Pages: 93007fea7f57cbb3d1297a7628d5604374393096  
Travis CI 已提交
421 
      <script type="text/javascript" src="https://cdn.bootcss.com/mathjax/2.7.0/MathJax.js"></script>
T
Deploy to GitHub Pages: 2c54905c3e3154327a8eb57ad96d0770dcc5f435  
Travis CI 已提交
422 423 424 425 426 427 428 429 430 431 432 433 434 
       
  

  
  
    <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
435 
</html>