Skip to content

P
PaddleSlim

PaddlePaddle / PaddleSlim
1 年多 前同步成功

通知 51
Star 1434
Fork 344
P
PaddleSlim
切换分支/标签
查找文件 普通视图历史永久链接Permalink
index.html 12.9 KB
Newer Older
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 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 
<!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 http-equiv="X-UA-Compatible" content="IE=edge">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  
  <link rel="shortcut icon" href="../../img/favicon.ico">
  <title>Sensitivity demo - PaddleSlim Docs</title>
  <link href='https://fonts.googleapis.com/css?family=Lato:400,700|Roboto+Slab:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>

  <link rel="stylesheet" href="../../css/theme.css" type="text/css" />
  <link rel="stylesheet" href="../../css/theme_extra.css" type="text/css" />
  <link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
  
  <script>
    // Current page data
    var mkdocs_page_name = "Sensitivity demo";
    var mkdocs_page_input_path = "tutorials/sensitivity_demo.md";
    var mkdocs_page_url = null;
  </script>
  
  <script src="../../js/jquery-2.1.1.min.js" defer></script>
  <script src="../../js/modernizr-2.8.3.min.js" defer></script>
  <script src="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
  <script>hljs.initHighlightingOnLoad();</script> 
  
</head>

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

  <div class="wy-grid-for-nav">

    
    <nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
      <div class="wy-side-nav-search">
        <a href="../.." class="icon icon-home"> PaddleSlim Docs</a>
        <div role="search">
  <form id ="rtd-search-form" class="wy-form" action="../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" title="Type search term here" />
  </form>
</div>
      </div>

      <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
	<ul class="current">
	  
          
            <li class="toctree-l1">
		
    <a class="" href="../..">Home</a>
	    </li>
          
            <li class="toctree-l1">
B
Deployed 482d368 with MkDocs version: 1.0.4  
baiyfbupt 已提交
58 59 60 61 62 
    <a class="" href="../../model_zoo/">模型库</a>
	    </li>
          
            <li class="toctree-l1">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
63 
    <span class="caption-text">教程</span>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
64 65 66 
    <ul class="subnav">
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
67 
    <a class="" href="../quant_post_demo/">离线量化</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
68 69 70 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
71 
    <a class="" href="../quant_aware_demo/">量化训练</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
72 73 74 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
75 
    <a class="" href="../quant_embedding_demo/">Embedding量化</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
76 77 78 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
79 
    <a class="" href="../nas_demo/">SA搜索</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
80 81 82 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
83 
    <a class="" href="../distillation_demo/">知识蒸馏</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
84 
                </li>
B
Deployed 9bcff00 with MkDocs version: 1.0.4  
baiyfbupt 已提交
85 86 87 88 89 
    </ul>
	    </li>
          
            <li class="toctree-l1">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
90 
    <span class="caption-text">API</span>
B
Deployed 9bcff00 with MkDocs version: 1.0.4  
baiyfbupt 已提交
91 
    <ul class="subnav">
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
92 93 
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
94 
    <a class="" href="../../api/quantization_api/">量化</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
95 96 97 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
98 
    <a class="" href="../../api/prune_api/">剪枝与敏感度</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
99 100 101 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
102 
    <a class="" href="../../api/analysis_api/">模型分析</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
103 104 105 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
106 
    <a class="" href="../../api/single_distiller_api/">知识蒸馏</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
107 108 109 
                </li>
                <li class="">
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
110 111 112 113 114 115 116 117 118 
    <a class="" href="../../api/nas_api/">SA搜索</a>
                </li>
                <li class="">
                    
    <a class="" href="../../api/search_space/">搜索空间</a>
                </li>
                <li class="">
                    
    <a class="" href="../../table_latency/">硬件延时评估表</a>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 
                </li>
    </ul>
	    </li>
          
            <li class="toctree-l1">
		
    <a class="" href="../../algo/algo/">算法原理</a>
	    </li>
          
        </ul>
      </div>
      &nbsp;
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">

      
      <nav class="wy-nav-top" role="navigation" aria-label="top navigation">
        <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
        <a href="../..">PaddleSlim Docs</a>
      </nav>

      
      <div class="wy-nav-content">
        <div class="rst-content">
          <div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
    <li><a href="../..">Docs</a> &raquo;</li>
    
      
    
    <li>Sensitivity demo</li>
    <li class="wy-breadcrumbs-aside">
      
        <a href="https://github.com/PaddlePaddle/PaddleSlim/edit/master/docs/tutorials/sensitivity_demo.md"
          class="icon icon-github"> Edit on GitHub</a>
      
    </li>
  </ul>
  <hr/>
</div>
          <div role="main">
            <div class="section">
              
                <p>该示例介绍如何分析卷积网络中各卷积层的敏感度，以及如何根据计算出的敏感度选择一组合适的剪裁率。
该示例默认会自动下载并使用MNIST数据。支持以下模型：</p>
<ul>
<li>MobileNetV1</li>
<li>MobileNetV2</li>
<li>ResNet50</li>
</ul>
<h2 id="1">1. 接口介绍<a class="headerlink" href="#1" title="Permanent link">#</a></h2>
<p>该示例涉及以下接口：</p>
<ul>
<li><a href="https://paddlepaddle.github.io/PaddleSlim/api/prune_api/#sensitivity">paddleslim.prune.sensitivity</a></li>
<li><a href="https://paddlepaddle.github.io/PaddleSlim/api/prune_api/#merge_sensitive">paddleslim.prune.merge_sensitive</a></li>
<li><a href="https://paddlepaddle.github.io/PaddleSlim/api/prune_api/#get_ratios_by_losssensitivities-loss">paddleslim.prune.get_ratios_by_loss</a></li>
</ul>
<h2 id="2">2. 运行示例<a class="headerlink" href="#2" title="Permanent link">#</a></h2>
<p>在路径<code>PaddleSlim/demo/sensitive</code>下执行以下代码运行示例：</p>
<table class="codehilitetable"><tr><td class="linenos"><div class="linenodiv"><pre><span></span>1
2</pre></div></td><td class="code"><div class="codehilite"><pre><span></span><span class="n">export</span> <span class="n">CUDA_VISIBLE_DEVICES</span><span class="o">=</span><span class="mi">0</span>
<span class="n">python</span> <span class="n">train</span><span class="p">.</span><span class="n">py</span> <span class="c1">--model &quot;MobileNetV1&quot;</span>
</pre></div>
</td></tr></table>

<p>通过<code>python train.py --help</code>查看更多选项。</p>
<h2 id="3">3. 重要步骤说明<a class="headerlink" href="#3" title="Permanent link">#</a></h2>
<h3 id="31">3.1 计算敏感度<a class="headerlink" href="#31" title="Permanent link">#</a></h3>
<p>计算敏感度之前，用户需要搭建好用于测试的网络，以及实现评估模型精度的回调函数。</p>
<p>调用<code>paddleslim.prune.sensitivity</code>接口计算敏感度。敏感度信息会追加到<code>sensitivities_file</code>选项所指定的文件中，如果需要重新计算敏感度，需要先删除<code>sensitivities_file</code>文件。</p>
<p>如果模型评估速度较慢，可以通过多进程的方式加速敏感度计算过程。比如在进程1中设置<code>pruned_ratios=[0.1, 0.2, 0.3, 0.4]</code>，并将敏感度信息存放在文件<code>sensitivities_0.data</code>中，然后在进程2中设置<code>pruned_ratios=[0.5, 0.6, 0.7]</code>，并将敏感度信息存储在文件<code>sensitivities_1.data</code>中。这样每个进程只会计算指定剪切率下的敏感度信息。多进程可以运行在单机多卡，或多机多卡。</p>
<p>代码如下：</p>
<table class="codehilitetable"><tr><td class="linenos"><div class="linenodiv"><pre><span></span>1
2
3
4
5
6
7
8</pre></div></td><td class="code"><div class="codehilite"><pre><span></span><span class="o">#</span> <span class="err">进程</span><span class="mi">1</span>
<span class="n">sensitivity</span><span class="p">(</span>
    <span class="n">val_program</span><span class="p">,</span>
    <span class="n">place</span><span class="p">,</span>
    <span class="n">params</span><span class="p">,</span>
    <span class="n">test</span><span class="p">,</span>
    <span class="n">sensitivities_file</span><span class="o">=</span><span class="ss">&quot;sensitivities_0.data&quot;</span><span class="p">,</span>
    <span class="n">pruned_ratios</span><span class="o">=</span><span class="p">[</span><span class="mi">0</span><span class="p">.</span><span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">3</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">4</span><span class="p">])</span>
</pre></div>
</td></tr></table>

<table class="codehilitetable"><tr><td class="linenos"><div class="linenodiv"><pre><span></span>1
2
3
4
5
6
7
8</pre></div></td><td class="code"><div class="codehilite"><pre><span></span><span class="o">#</span> <span class="err">进程</span><span class="mi">2</span>
<span class="n">sensitivity</span><span class="p">(</span>
    <span class="n">val_program</span><span class="p">,</span>
    <span class="n">place</span><span class="p">,</span>
    <span class="n">params</span><span class="p">,</span>
    <span class="n">test</span><span class="p">,</span>
    <span class="n">sensitivities_file</span><span class="o">=</span><span class="ss">&quot;sensitivities_1.data&quot;</span><span class="p">,</span>
    <span class="n">pruned_ratios</span><span class="o">=</span><span class="p">[</span><span class="mi">0</span><span class="p">.</span><span class="mi">5</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">6</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">7</span><span class="p">])</span>
</pre></div>
</td></tr></table>

<h3 id="32">3.2 合并敏感度<a class="headerlink" href="#32" title="Permanent link">#</a></h3>
<p>如果用户通过上一节多进程的方式生成了多个存储敏感度信息的文件，可以通过<code>paddleslim.prune.merge_sensitive</code>将其合并，合并后的敏感度信息存储在一个<code>dict</code>中。代码如下：</p>
<table class="codehilitetable"><tr><td class="linenos"><div class="linenodiv"><pre><span></span>1</pre></div></td><td class="code"><div class="codehilite"><pre><span></span><span class="n">sens</span> <span class="o">=</span> <span class="n">merge_sensitive</span><span class="p">([</span><span class="ss">&quot;./sensitivities_0.data&quot;</span><span class="p">,</span> <span class="ss">&quot;./sensitivities_1.data&quot;</span><span class="p">])</span>
</pre></div>
</td></tr></table>

<h3 id="33">3.3 计算剪裁率<a class="headerlink" href="#33" title="Permanent link">#</a></h3>
<p>调用<code>paddleslim.prune.get_ratios_by_loss</code>接口计算一组剪裁率。</p>
<table class="codehilitetable"><tr><td class="linenos"><div class="linenodiv"><pre><span></span>1</pre></div></td><td class="code"><div class="codehilite"><pre><span></span><span class="n">ratios</span> <span class="o">=</span> <span class="n">get_ratios_by_loss</span><span class="p">(</span><span class="n">sens</span><span class="p">,</span> <span class="mi">0</span><span class="p">.</span><span class="mi">01</span><span class="p">)</span>
</pre></div>
</td></tr></table>

<p>其中，<code>0.01</code>为一个阈值，对于任意卷积层，其剪裁率为使精度损失低于阈值<code>0.01</code>的最大剪裁率。</p>
<p>用户在计算出一组剪裁率之后可以通过接口<code>paddleslim.prune.Pruner</code>剪裁网络，并用接口<code>paddleslim.analysis.flops</code>计算<code>FLOPs</code>。如果<code>FLOPs</code>不满足要求，调整阈值重新计算出一组剪裁率。</p>
              
            </div>
          </div>
          <footer>
  

  <hr/>

  <div role="contentinfo">
    <!-- Copyright etc -->
    
  </div>

  Built with <a href="http://www.mkdocs.org">MkDocs</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>

  <div class="rst-versions" role="note" style="cursor: pointer">
    <span class="rst-current-version" data-toggle="rst-current-version">
      
          <a href="https://github.com/PaddlePaddle/PaddleSlim/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
      
      
      
    </span>
</div>
    <script>var base_url = '../..';</script>
    <script src="../../js/theme.js" defer></script>
      <script src="../../mathjax-config.js" defer></script>
      <script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML" defer></script>
      <script src="../../search/main.js" defer></script>

</body>
</html>