Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git


  
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

	
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

	
95

	
96

	
97

	
98

	
99

	
100

	
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

	
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

	
            <!DOCTYPE html>
<html class="writer-html5" lang="en" >
<head>
  <meta charset="utf-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>super_gradients.training.utils.ema &mdash; SuperGradients 1.0 documentation</title>
      <link rel="stylesheet" href="../../../../_static/pygments.css" type="text/css" />
      <link rel="stylesheet" href="../../../../_static/css/theme.css" type="text/css" />
      <link rel="stylesheet" href="../../../../_static/graphviz.css" type="text/css" />
  <!--[if lt IE 9]>
    <script src="../../../../_static/js/html5shiv.min.js"></script>
  <![endif]-->
  
        <script data-url_root="../../../../" id="documentation_options" src="../../../../_static/documentation_options.js"></script>
        <script src="../../../../_static/jquery.js"></script>
        <script src="../../../../_static/underscore.js"></script>
        <script src="../../../../_static/doctools.js"></script>
    <script src="../../../../_static/js/theme.js"></script>
    <link rel="index" title="Index" href="../../../../genindex.html" />
    <link rel="search" title="Search" href="../../../../search.html" /> 
</head>

<body class="wy-body-for-nav"> 
  <div class="wy-grid-for-nav">
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search" >
            <a href="../../../../index.html" class="icon icon-home"> SuperGradients
          </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" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>
        </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
              <p class="caption"><span class="caption-text">SuperGradients</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html">SuperGradients</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.common.html">Common</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.training.html">Training</a></li>
</ul>

        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
          <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
          <a href="../../../../index.html">SuperGradients</a>
      </nav>

      <div class="wy-nav-content">
        <div class="rst-content">
          <div role="navigation" aria-label="Page navigation">
  <ul class="wy-breadcrumbs">
      <li><a href="../../../../index.html" class="icon icon-home"></a> &raquo;</li>
          <li><a href="../../../index.html">Module code</a> &raquo;</li>
      <li>super_gradients.training.utils.ema</li>
      <li class="wy-breadcrumbs-aside">
      </li>
  </ul>
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
             
  <h1>Source code for super_gradients.training.utils.ema</h1><div class="highlight"><pre>
<span></span><span class="kn">import</span> <span class="nn">math</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="kn">from</span> <span class="nn">copy</span> <span class="kn">import</span> <span class="n">deepcopy</span>
<span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">Union</span>

<span class="kn">import</span> <span class="nn">torch</span>
<span class="kn">from</span> <span class="nn">torch</span> <span class="kn">import</span> <span class="n">nn</span>

<span class="kn">from</span> <span class="nn">super_gradients.training.models</span> <span class="kn">import</span> <span class="n">SgModule</span>


<div class="viewcode-block" id="copy_attr"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.ema.copy_attr">[docs]</a><span class="k">def</span> <span class="nf">copy_attr</span><span class="p">(</span><span class="n">a</span><span class="p">:</span> <span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">,</span> <span class="n">b</span><span class="p">:</span> <span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">,</span> <span class="n">include</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">list</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="p">(),</span> <span class="n">exclude</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">list</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="p">()):</span>
    <span class="c1"># Copy attributes from b to a, options to only include [...] and to exclude [...]</span>
    <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">b</span><span class="o">.</span><span class="vm">__dict__</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
        <span class="k">if</span> <span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">include</span><span class="p">)</span> <span class="ow">and</span> <span class="n">k</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">include</span><span class="p">)</span> <span class="ow">or</span> <span class="n">k</span><span class="o">.</span><span class="n">startswith</span><span class="p">(</span><span class="s1">&#39;_&#39;</span><span class="p">)</span> <span class="ow">or</span> <span class="n">k</span> <span class="ow">in</span> <span class="n">exclude</span><span class="p">:</span>
            <span class="k">continue</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="nb">setattr</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span><span class="p">)</span></div>


<div class="viewcode-block" id="ModelEMA"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.ema.ModelEMA">[docs]</a><span class="k">class</span> <span class="nc">ModelEMA</span><span class="p">:</span>
    <span class="sd">&quot;&quot;&quot; Model Exponential Moving Average from https://github.com/rwightman/pytorch-image-models</span>
<span class="sd">    Keep a moving average of everything in the model state_dict (parameters and buffers).</span>
<span class="sd">    This is intended to allow functionality like</span>
<span class="sd">    https://www.tensorflow.org/api_docs/python/tf/train/ExponentialMovingAverage</span>
<span class="sd">    A smoothed version of the weights is necessary for some training schemes to perform well.</span>
<span class="sd">    This class is sensitive where it is initialized in the sequence of model init,</span>
<span class="sd">    GPU assignment and distributed training wrappers.</span>
<span class="sd">    &quot;&quot;&quot;</span>

    <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">model</span><span class="p">,</span> <span class="n">decay</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.9999</span><span class="p">,</span> <span class="n">beta</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">15</span><span class="p">,</span> <span class="n">exp_activation</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        Init the EMA</span>
<span class="sd">        :param model: Union[SgModule, nn.Module], the training model to construct the EMA model by</span>
<span class="sd">                    IMPORTANT: WHEN THE APPLICATION OF EMA ONLY ON A SUBSET OF ATTRIBUTES IS DESIRED, WRAP THE NN.MODULE</span>
<span class="sd">                    AS SgModule AND OVERWRITE get_include_attributes() AND get_exclude_attributes() AS DESIRED (SEE</span>
<span class="sd">                    YoLoBase IMPLEMENTATION IN super_gradients.trainer.models.yolov5_base.py AS AN EXAMPLE).</span>
<span class="sd">        :param decay: the maximum decay value. as the training process advances, the decay will climb towards this value</span>
<span class="sd">                      until the EMA_t+1 = EMA_t * decay + TRAINING_MODEL * (1- decay)</span>
<span class="sd">        :param beta: the exponent coefficient. The higher the beta, the sooner in the training the decay will saturate to</span>
<span class="sd">                     its final value. beta=15 is ~40% of the training process.</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="c1"># Create EMA</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">ema</span> <span class="o">=</span> <span class="n">deepcopy</span><span class="p">(</span><span class="n">model</span><span class="p">)</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">ema</span><span class="o">.</span><span class="n">eval</span><span class="p">()</span>
        <span class="k">if</span> <span class="n">exp_activation</span><span class="p">:</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">decay_function</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">decay</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">math</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="o">-</span><span class="n">x</span> <span class="o">*</span> <span class="n">beta</span><span class="p">))</span>  <span class="c1"># decay exponential ramp (to help early epochs)</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">decay_function</span> <span class="o">=</span> <span class="k">lambda</span> <span class="n">x</span><span class="p">:</span> <span class="n">decay</span>  <span class="c1"># always return the same decay factor</span>

        <span class="sd">&quot;&quot;&quot;&quot;</span>
<span class="sd">        we hold a list of model attributes (not wights and biases) which we would like to include in each </span>
<span class="sd">        attribute update or exclude from each update. a SgModule declare these attribute using </span>
<span class="sd">        get_include_attributes and get_exclude_attributes functions. for a nn.Module which is not a SgModule</span>
<span class="sd">        all non-private (not starting with &#39;_&#39;) attributes will be updated (and only them).</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">module</span><span class="p">,</span> <span class="n">SgModule</span><span class="p">):</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">include_attributes</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">module</span><span class="o">.</span><span class="n">get_include_attributes</span><span class="p">()</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">exclude_attributes</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">module</span><span class="o">.</span><span class="n">get_exclude_attributes</span><span class="p">()</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="s2">&quot;Warning: EMA should be used with SgModule instance. All attributes of the model will be &quot;</span>
                          <span class="s2">&quot;included in EMA&quot;</span><span class="p">)</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">include_attributes</span> <span class="o">=</span> <span class="p">[]</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">exclude_attributes</span> <span class="o">=</span> <span class="p">[]</span>
        <span class="k">for</span> <span class="n">p</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">ema</span><span class="o">.</span><span class="n">module</span><span class="o">.</span><span class="n">parameters</span><span class="p">():</span>
            <span class="n">p</span><span class="o">.</span><span class="n">requires_grad_</span><span class="p">(</span><span class="kc">False</span><span class="p">)</span>

<div class="viewcode-block" id="ModelEMA.update"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.ema.ModelEMA.update">[docs]</a>    <span class="k">def</span> <span class="nf">update</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">model</span><span class="p">,</span> <span class="n">training_percent</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        Update the state of the EMA model.</span>
<span class="sd">        :param model: current training model</span>
<span class="sd">        :param training_percent: the percentage of the training process [0,1]. i.e 0.4 means 40% of the training have passed</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="c1"># Update EMA parameters</span>
        <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">():</span>
            <span class="n">decay</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">decay_function</span><span class="p">(</span><span class="n">training_percent</span><span class="p">)</span>

            <span class="k">for</span> <span class="n">ema_v</span><span class="p">,</span> <span class="n">model_v</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ema</span><span class="o">.</span><span class="n">module</span><span class="o">.</span><span class="n">state_dict</span><span class="p">()</span><span class="o">.</span><span class="n">values</span><span class="p">(),</span> <span class="n">model</span><span class="o">.</span><span class="n">state_dict</span><span class="p">()</span><span class="o">.</span><span class="n">values</span><span class="p">()):</span>
                <span class="k">if</span> <span class="n">ema_v</span><span class="o">.</span><span class="n">dtype</span><span class="o">.</span><span class="n">is_floating_point</span><span class="p">:</span>
                    <span class="n">ema_v</span><span class="o">.</span><span class="n">copy_</span><span class="p">(</span><span class="n">ema_v</span> <span class="o">*</span> <span class="n">decay</span> <span class="o">+</span> <span class="p">(</span><span class="mf">1.</span> <span class="o">-</span> <span class="n">decay</span><span class="p">)</span> <span class="o">*</span> <span class="n">model_v</span><span class="o">.</span><span class="n">detach</span><span class="p">())</span></div>

<div class="viewcode-block" id="ModelEMA.update_attr"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.ema.ModelEMA.update_attr">[docs]</a>    <span class="k">def</span> <span class="nf">update_attr</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">model</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        This function updates model attributes (not weight and biases) from original model to the ema model.</span>
<span class="sd">        attributes of the original model, such as anchors and grids (of detection models), may be crucial to the</span>
<span class="sd">        model operation and need to be updated.</span>
<span class="sd">        If include_attributes and exclude_attributes lists were not defined, all non-private (not starting with &#39;_&#39;)</span>
<span class="sd">        attributes will be updated (and only them).</span>
<span class="sd">        :param model: the source model</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">copy_attr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ema</span><span class="o">.</span><span class="n">module</span><span class="p">,</span> <span class="n">model</span><span class="o">.</span><span class="n">module</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">include_attributes</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">exclude_attributes</span><span class="p">)</span></div></div>
</pre></div>

           </div>
          </div>
          <footer>

  <hr/>

  <div role="contentinfo">
    <p>&#169; Copyright 2021, SuperGradients team.</p>
  </div>

  Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    provided by <a href="https://readthedocs.org">Read the Docs</a>.
   

</footer>
        </div>
      </div>
    </section>
  </div>
  <script>
      jQuery(function () {
          SphinxRtdTheme.Navigation.enable(true);
      });
  </script> 

</body>
</html>