Skip to content

P
PaddleSlim

PaddlePaddle / PaddleSlim
大约 1 年 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 29.9 KB
Newer Older
B
Deployed 09f99c9 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>Embedding量化 - 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 = "Embedding\u91cf\u5316";
    var mkdocs_page_input_path = "tutorials/quant_embedding_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 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 
    <span class="caption-text">教程</span>
    <ul class="subnav">
                <li class="">
                    
    <a class="" href="../quant_post_demo/">离线量化</a>
                </li>
                <li class="">
                    
    <a class="" href="../quant_aware_demo/">量化训练</a>
                </li>
                <li class=" current">
                    
    <a class="current" href="./">Embedding量化</a>
    <ul class="subnav">
            
    <li class="toctree-l3"><a href="#embedding">Embedding量化示例</a></li>
    
        <ul>
        
            <li><a class="toctree-l4" href="#skip-gramword2vector">基于skip-gram的word2vector模型</a></li>
        
            <li><a class="toctree-l4" href="#skip-gramword2vector_1">量化基于skip-gram的word2vector模型</a></li>
        
        </ul>
    

    </ul>
                </li>
                <li class="">
                    
    <a class="" href="../nas_demo/">SA搜索</a>
                </li>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
95 96 97 98 
                <li class="">
                    
    <a class="" href="../distillation_demo/">知识蒸馏</a>
                </li>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
99 100 
    </ul>
	    </li>
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 
          
            <li class="toctree-l1">
		
    <span class="caption-text">API</span>
    <ul class="subnav">
                <li class="">
                    
    <a class="" href="../../api/quantization_api/">量化</a>
                </li>
                <li class="">
                    
    <a class="" href="../../api/prune_api/">剪枝与敏感度</a>
                </li>
                <li class="">
                    
    <a class="" href="../../api/analysis_api/">模型分析</a>
                </li>
                <li class="">
                    
    <a class="" href="../../api/single_distiller_api/">知识蒸馏</a>
                </li>
                <li class="">
                    
    <a class="" href="../../api/nas_api/">SA搜索</a>
                </li>
                <li class="">
B
Deployed ab60054 with MkDocs version: 1.0.4  
baiyfbupt 已提交
128 
    <a class="" href="../../search_space/">搜索空间</a>
B
Deployed 55210fe with MkDocs version: 1.0.4  
baiyfbupt 已提交
129 130 131 132 133 134 135 
                </li>
                <li class="">
                    
    <a class="" href="../../table_latency/">硬件延时评估表</a>
                </li>
    </ul>
	    </li>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
136 137 138 
          
            <li class="toctree-l1">
B
Deployed 9bcff00 with MkDocs version: 1.0.4  
baiyfbupt 已提交
139 
    <a class="" href="../../algo/algo/">算法原理</a>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
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 
	    </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>教程 &raquo;</li>
        
      
    
    <li>Embedding量化</li>
    <li class="wy-breadcrumbs-aside">
B
Deployed 0b29482 with MkDocs version: 1.0.4  
baiyfbupt 已提交
171 
        <a href="https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/docs/tutorials/quant_embedding_demo.md"
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
172 173 174 175 176 177 178 179 180 181 
          class="icon icon-github"> Edit on GitHub</a>
      
    </li>
  </ul>
  <hr/>
</div>
          <div role="main">
            <div class="section">
              
                <h1 id="embedding">Embedding量化示例<a class="headerlink" href="#embedding" title="Permanent link">#</a></h1>
B
Deployed ab60054 with MkDocs version: 1.0.4  
baiyfbupt 已提交
182 183 
<p>本示例介绍如何使用Embedding量化的接口 <a href="https://paddlepaddle.github.io/PaddleSlim/api/quantization_api/">paddleslim.quant.quant_embedding</a> 。<code>quant_embedding</code>接口将网络中的Embedding参数从<code>float32</code>类型量化到 <code>8-bit</code>整数类型，在几乎不损失模型精度的情况下减少模型的存储空间和显存占用。</p>
<p>接口介绍请参考 <a href="https://paddlepaddle.github.io/PaddleSlim/api/quantization_api/">量化API文档</a>。</p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
184 185 186 
<p>该接口对program的修改：</p>
<p>量化前:</p>
<p align="center">
B
Deployed ab60054 with MkDocs version: 1.0.4  
baiyfbupt 已提交
187 
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleSlim/develop/demo/quant/quant_embedding/image/before.png" height=200 width=100 hspace='10'/> <br />
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
188 189 190 191 192 
<strong>图1：量化前的模型结构</strong>
</p>

<p>量化后：</p>
<p align="center">
B
Deployed ab60054 with MkDocs version: 1.0.4  
baiyfbupt 已提交
193 
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleSlim/develop/demo/quant/quant_embedding/image/after.png" height=300 width=300 hspace='10'/> <br />
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
194 195 196 197 198 199 
<strong>图2: 量化后的模型结构</strong>
</p>

<p>以下将以 <code>基于skip-gram的word2vector模型</code> 为例来说明如何使用<code>quant_embedding</code>接口。首先介绍 <code>基于skip-gram的word2vector模型</code> 的正常训练和测试流程。</p>
<h2 id="skip-gramword2vector">基于skip-gram的word2vector模型<a class="headerlink" href="#skip-gramword2vector" title="Permanent link">#</a></h2>
<p>以下是本例的简要目录结构及说明：</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
200 
<div class="codehilite"><pre><span></span>.
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 
├── cluster_train.py    # 分布式训练函数
├── cluster_train.sh    # 本地模拟多机脚本
├── train.py            # 训练函数
├── infer.py            # 预测脚本
├── net.py              # 网络结构
├── preprocess.py       # 预处理脚本，包括构建词典和预处理文本
├── reader.py           # 训练阶段的文本读写
├── train.py            # 训练函数
└── utils.py            # 通用函数
</pre></div>

<h3 id="_1">介绍<a class="headerlink" href="#_1" title="Permanent link">#</a></h3>
<p>本例实现了skip-gram模式的word2vector模型。</p>
<p>同时推荐用户参考<a href="https://aistudio.baidu.com/aistudio/projectDetail/124377"> IPython Notebook demo</a></p>
<h3 id="_2">数据下载<a class="headerlink" href="#_2" title="Permanent link">#</a></h3>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
216 
<p>全量数据集使用的是来自1 Billion Word Language Model Benchmark的(<a href="http://www.statmt.org/lm-benchmark">http://www.statmt.org/lm-benchmark</a>) 的数据集.</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
217 
<div class="codehilite"><pre><span></span>mkdir data
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
218 219 220 221 222 223 
wget http://www.statmt.org/lm-benchmark/1-billion-word-language-modeling-benchmark-r13output.tar.gz
tar xzvf <span class="m">1</span>-billion-word-language-modeling-benchmark-r13output.tar.gz
mv <span class="m">1</span>-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled/ data/
</pre></div>

<p>备用数据地址下载命令如下</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
224 
<div class="codehilite"><pre><span></span>mkdir data
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
225 226 227 228 229 230 
wget https://paddlerec.bj.bcebos.com/word2vec/1-billion-word-language-modeling-benchmark-r13output.tar
tar xvf <span class="m">1</span>-billion-word-language-modeling-benchmark-r13output.tar
mv <span class="m">1</span>-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled/ data/
</pre></div>

<p>为了方便快速验证，我们也提供了经典的text8样例数据集，包含1700w个词。 下载命令如下</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
231 
<div class="codehilite"><pre><span></span>mkdir data
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
232 233 234 235 236 237 238 239 
wget https://paddlerec.bj.bcebos.com/word2vec/text.tar
tar xvf text.tar
mv text data/
</pre></div>

<h3 id="_3">数据预处理<a class="headerlink" href="#_3" title="Permanent link">#</a></h3>
<p>以样例数据集为例进行预处理。全量数据集注意解压后以training-monolingual.tokenized.shuffled 目录为预处理目录，和样例数据集的text目录并列。</p>
<p>词典格式: 词&lt;空格&gt;词频。注意低频词用'UNK'表示</p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
240 
<p>可以按格式自建词典，如果自建词典跳过第一步。
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
241 
<div class="codehilite"><pre><span></span><span class="n">the</span> <span class="mi">1061396</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
242 243 244 245 246 247 248 249 250 
<span class="k">of</span> <span class="mi">593677</span>
<span class="k">and</span> <span class="mi">416629</span>
<span class="n">one</span> <span class="mi">411764</span>
<span class="k">in</span> <span class="mi">372201</span>
<span class="n">a</span> <span class="mi">325873</span>
<span class="o">&lt;</span><span class="n">UNK</span><span class="o">&gt;</span> <span class="mi">324608</span>
<span class="k">to</span> <span class="mi">316376</span>
<span class="n">zero</span> <span class="mi">264975</span>
<span class="n">nine</span> <span class="mi">250430</span>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
251 
</pre></div></p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
252 
<p>第一步根据英文语料生成词典，中文语料可以通过修改text_strip方法自定义处理方法。</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
253 
<div class="codehilite"><pre><span></span>python preprocess.py --build_dict --build_dict_corpus_dir data/text/ --dict_path data/test_build_dict
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
254 255 256 
</pre></div>

<p>第二步根据词典将文本转成id, 同时进行downsample，按照概率过滤常见词, 同时生成word和id映射的文件，文件名为词典+"<em>word_to_id</em>"。</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
257 
<div class="codehilite"><pre><span></span>python preprocess.py --filter_corpus --dict_path data/test_build_dict --input_corpus_dir data/text --output_corpus_dir data/convert_text8 --min_count <span class="m">5</span> --downsample <span class="m">0</span>.001
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
258 259 260 261 
</pre></div>

<h3 id="_4">训练<a class="headerlink" href="#_4" title="Permanent link">#</a></h3>
<p>具体的参数配置可运行</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
262 
<div class="codehilite"><pre><span></span>python train.py -h
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
263 264 
</pre></div>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
265 
<p>单机多线程训练
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
266 267 
<div class="codehilite"><pre><span></span><span class="nv">OPENBLAS_NUM_THREADS</span><span class="o">=</span><span class="m">1</span> <span class="nv">CPU_NUM</span><span class="o">=</span><span class="m">5</span> python train.py --train_data_dir data/convert_text8 --dict_path data/test_build_dict --num_passes <span class="m">10</span> --batch_size <span class="m">100</span> --model_output_dir v1_cpu5_b100_lr1dir --base_lr <span class="m">1</span>.0 --print_batch <span class="m">1000</span> --with_speed --is_sparse
</pre></div></p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
268 
<p>本地单机模拟多机训练</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
269 
<div class="codehilite"><pre><span></span>sh cluster_train.sh
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
270 271 
</pre></div>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
272 
<p>本示例中按照单机多线程训练的命令进行训练，训练完毕后，可看到在当前文件夹下保存模型的路径为:     <code>v1_cpu5_b100_lr1dir</code>, 运行 <code>ls v1_cpu5_b100_lr1dir</code>可看到该文件夹下保存了训练的10个epoch的模型文件。
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
273 274 
<div class="codehilite"><pre><span></span><span class="n">pass</span><span class="o">-</span><span class="mi">0</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">1</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">2</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">3</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">4</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">5</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">6</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">7</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">8</span>  <span class="n">pass</span><span class="o">-</span><span class="mi">9</span>
</pre></div></p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
275 276 
<h3 id="_5">预测<a class="headerlink" href="#_5" title="Permanent link">#</a></h3>
<p>测试集下载命令如下</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
277 
<div class="codehilite"><pre><span></span><span class="c1">#全量数据集测试集</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
278 279 280 281 282 
wget https://paddlerec.bj.bcebos.com/word2vec/test_dir.tar
<span class="c1">#样本数据集测试集</span>
wget https://paddlerec.bj.bcebos.com/word2vec/test_mid_dir.tar
</pre></div>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
283 
<p>预测命令，注意词典名称需要加后缀"<em>word_to_id</em>", 此文件是预处理阶段生成的。
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
284 
<div class="codehilite"><pre><span></span>python infer.py --infer_epoch --test_dir data/test_mid_dir --dict_path data/test_build_dict_word_to_id_ --batch_size <span class="m">20000</span> --model_dir v1_cpu5_b100_lr1dir/  --start_index <span class="m">0</span> --last_index <span class="m">9</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
285 
</pre></div>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
286 
运行该预测命令, 可看到如下输出
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
287 
<div class="codehilite"><pre><span></span><span class="p">(</span><span class="s1">&#39;start index: &#39;</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="s1">&#39; last_index:&#39;</span><span class="p">,</span> <span class="mi">9</span><span class="p">)</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 
<span class="p">(</span><span class="s1">&#39;vocab_size:&#39;</span><span class="p">,</span> <span class="mi">63642</span><span class="p">)</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">249</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">0</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">014</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">590</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">1</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">033</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">982</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">2</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">055</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1338</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">3</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">075</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1653</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">4</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">093</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1914</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">5</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">107</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2204</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">6</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">124</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2416</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">7</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">136</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2606</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">8</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">146</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2722</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">9</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">153</span>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
309 
</pre></div></p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
310 
<h2 id="skip-gramword2vector_1">量化<code>基于skip-gram的word2vector模型</code><a class="headerlink" href="#skip-gramword2vector_1" title="Permanent link">#</a></h2>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
311 
<p>量化配置为:
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
312 
<div class="codehilite"><pre><span></span><span class="n">config</span> <span class="o">=</span> <span class="err">{</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
313 314 315 
        <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span>
        <span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span>
        <span class="err">}</span>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
316 
</pre></div></p>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
317 
<p>运行命令为：</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
318 
<div class="codehilite"><pre><span></span>python infer.py --infer_epoch --test_dir data/test_mid_dir --dict_path data/test_build_dict_word_to_id_ --batch_size <span class="m">20000</span> --model_dir v1_cpu5_b100_lr1dir/  --start_index <span class="m">0</span> --last_index <span class="m">9</span> --emb_quant True
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
319 320 321 
</pre></div>

<p>运行输出为:</p>
B
Deployed 2c115a8 with MkDocs version: 1.0.4  
baiyfbupt 已提交
322 
<div class="codehilite"><pre><span></span><span class="p">(</span><span class="s1">&#39;start index: &#39;</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="s1">&#39; last_index:&#39;</span><span class="p">,</span> <span class="mi">9</span><span class="p">)</span>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
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 
<span class="p">(</span><span class="s1">&#39;vocab_size:&#39;</span><span class="p">,</span> <span class="mi">63642</span><span class="p">)</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">253</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">0</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">014</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">586</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">1</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">033</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">970</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">2</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">054</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1364</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">3</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">077</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1642</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">4</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">092</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">1936</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">5</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">109</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2216</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">6</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">124</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2419</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">7</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">136</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2603</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">8</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">146</span>
<span class="n">quant_embedding</span> <span class="n">config</span> <span class="err">{</span><span class="s1">&#39;quantize_type&#39;</span><span class="p">:</span> <span class="s1">&#39;abs_max&#39;</span><span class="p">,</span> <span class="s1">&#39;params_name&#39;</span><span class="p">:</span> <span class="s1">&#39;emb&#39;</span><span class="p">,</span> <span class="s1">&#39;quantize_bits&#39;</span><span class="p">:</span> <span class="mi">8</span><span class="p">,</span> <span class="s1">&#39;dtype&#39;</span><span class="p">:</span> <span class="s1">&#39;int8&#39;</span><span class="err">}</span>
<span class="n">step</span><span class="p">:</span><span class="mi">1</span> <span class="mi">2719</span>
<span class="n">epoch</span><span class="p">:</span><span class="mi">9</span>          <span class="n">acc</span><span class="p">:</span><span class="mi">0</span><span class="p">.</span><span class="mi">153</span>
</pre></div>

<p>量化后的模型保存在<code>./output_quant</code>中，可看到量化后的参数<code>'emb.int8'</code>的大小为3.9M, 在<code>./v1_cpu5_b100_lr1dir</code>中可看到量化前的参数<code>'emb'</code>的大小为16M。</p>
              
            </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
        <a href="../nas_demo/" class="btn btn-neutral float-right" title="SA搜索">Next <span class="icon icon-circle-arrow-right"></span></a>
      
      
        <a href="../quant_aware_demo/" class="btn btn-neutral" title="量化训练"><span class="icon icon-circle-arrow-left"></span> Previous</a>
      
    </div>
  

  <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><a href="../quant_aware_demo/" style="color: #fcfcfc;">&laquo; Previous</a></span>
      
      
        <span style="margin-left: 15px"><a href="../nas_demo/" style="color: #fcfcfc">Next &raquo;</a></span>
      
    </span>
</div>
    <script>var base_url = '../..';</script>
    <script src="../../js/theme.js" defer></script>
      <script src="../../mathjax-config.js" defer></script>
B
Deployed 0f9c3ce with MkDocs version: 1.0.4  
baiyfbupt 已提交
405 
      <script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML" defer></script>
B
Deployed 09f99c9 with MkDocs version: 1.0.4  
baiyfbupt 已提交
406 407 408 409 
      <script src="../../search/main.js" defer></script>

</body>
</html>