mirror of
https://github.com/labmlai/annotated_deep_learning_paper_implementations.git
synced 2025-10-31 02:39:16 +08:00
549 lines
36 KiB
HTML
549 lines
36 KiB
HTML
<!DOCTYPE html>
|
|
<html>
|
|
<head>
|
|
<meta http-equiv="content-type" content="text/html;charset=utf-8"/>
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0"/>
|
|
<meta name="description" content="A set of PyTorch implementations/tutorials of popular gradient descent based optimizers. Currently includes Adam, AMSGrad and RAdam optimizers."/>
|
|
|
|
<meta name="twitter:card" content="summary"/>
|
|
<meta name="twitter:image:src" content="https://avatars1.githubusercontent.com/u/64068543?s=400&v=4"/>
|
|
<meta name="twitter:title" content="Optimizers"/>
|
|
<meta name="twitter:description" content="A set of PyTorch implementations/tutorials of popular gradient descent based optimizers. Currently includes Adam, AMSGrad and RAdam optimizers."/>
|
|
<meta name="twitter:site" content="@labmlai"/>
|
|
<meta name="twitter:creator" content="@labmlai"/>
|
|
|
|
<meta property="og:url" content="https://nn.labml.ai/optimizers/index.html"/>
|
|
<meta property="og:title" content="Optimizers"/>
|
|
<meta property="og:image" content="https://avatars1.githubusercontent.com/u/64068543?s=400&v=4"/>
|
|
<meta property="og:site_name" content="LabML Neural Networks"/>
|
|
<meta property="og:type" content="object"/>
|
|
<meta property="og:title" content="Optimizers"/>
|
|
<meta property="og:description" content="A set of PyTorch implementations/tutorials of popular gradient descent based optimizers. Currently includes Adam, AMSGrad and RAdam optimizers."/>
|
|
|
|
<title>Optimizers</title>
|
|
<link rel="shortcut icon" href="/icon.png"/>
|
|
<link rel="stylesheet" href="../pylit.css">
|
|
<link rel="canonical" href="https://nn.labml.ai/optimizers/index.html"/>
|
|
<!-- Global site tag (gtag.js) - Google Analytics -->
|
|
<script async src="https://www.googletagmanager.com/gtag/js?id=G-4V3HC8HBLH"></script>
|
|
<script>
|
|
window.dataLayer = window.dataLayer || [];
|
|
|
|
function gtag() {
|
|
dataLayer.push(arguments);
|
|
}
|
|
|
|
gtag('js', new Date());
|
|
|
|
gtag('config', 'G-4V3HC8HBLH');
|
|
</script>
|
|
</head>
|
|
<body>
|
|
<div id='container'>
|
|
<div id="background"></div>
|
|
<div class='section'>
|
|
<div class='docs'>
|
|
<p>
|
|
<a class="parent" href="/">home</a>
|
|
<a class="parent" href="index.html">optimizers</a>
|
|
</p>
|
|
<p>
|
|
|
|
<a href="https://github.com/lab-ml/labml_nn/tree/master/labml_nn/optimizers/__init__.py">
|
|
<img alt="Github"
|
|
src="https://img.shields.io/github/stars/lab-ml/nn?style=social"
|
|
style="max-width:100%;"/></a>
|
|
<a href="https://join.slack.com/t/labforml/shared_invite/zt-egj9zvq9-Dl3hhZqobexgT7aVKnD14g/"
|
|
rel="nofollow">
|
|
<img alt="Join Slact"
|
|
src="https://img.shields.io/badge/slack-chat-green.svg?logo=slack"
|
|
style="max-width:100%;"/></a>
|
|
<a href="https://twitter.com/labmlai"
|
|
rel="nofollow">
|
|
<img alt="Twitter"
|
|
src="https://img.shields.io/twitter/follow/labmlai?style=social"
|
|
style="max-width:100%;"/></a>
|
|
</p>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-0'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-0'>#</a>
|
|
</div>
|
|
<h1>Optimizers</h1>
|
|
<h2>Optimizer Implementations</h2>
|
|
<ul>
|
|
<li><a href="adam.html">Adam Optimizer</a></li>
|
|
<li><a href="amsgrad.html">AMSGrad Optimizer</a></li>
|
|
<li><a href="adam_warmup.html">Adam Optimizer with warmup</a></li>
|
|
<li><a href="noam.html">Noam Optimizer</a></li>
|
|
<li><a href="radam.html">Rectified Adam Optimizer</a></li>
|
|
<li><a href="ada_belief.html">AdaBelief Optimizer</a></li>
|
|
</ul>
|
|
<p>This <a href="mnist_experiment.html">MNIST example</a> uses these optimizers.</p>
|
|
<h2>Generic Adaptive Optimizer Base class and Weight Decay</h2>
|
|
<p>This file defines a common base class for <em>Adam</em> and extensions of it.
|
|
The base class helps use implement other optimizers with minimal code
|
|
because of re-usability.</p>
|
|
<p>We also define a special class for L2 weight decay, so that we don’t
|
|
have to implement it inside each of the optimizers,
|
|
and can easily extend to other weight decays like L1 without
|
|
changing the optimizers.</p>
|
|
<p>Here are some concepts on PyTorch optimizers:</p>
|
|
<h3>Parameter groups</h3>
|
|
<p>PyTorch optimizers group parameters into sets called groups.
|
|
Each group can have it’s own hyper-parameters like learning rates.</p>
|
|
<p>In most common cases there will be only one group.
|
|
This is when you initialize your optimizer with,</p>
|
|
<pre><code class="python">Optimizer(model.parameters())
|
|
</code></pre>
|
|
|
|
<p>You can define multiple parameter groups when initializing the optimizer:</p>
|
|
<pre><code class="python">Optimizer([{'params': model1.parameters()}, {'params': model2.parameters(), 'lr': 2}])
|
|
</code></pre>
|
|
|
|
<p>Here we pass a list of groups. Each group is a dictionary with it’s parameters under the key ‘params’.
|
|
You specify any hyper-parameters as well. If the hyper parameters are not defined they will default
|
|
to the optimizer level defaults.</p>
|
|
<p>You can access (and even change) these groups, and their hyper-parameters with <code>optimizer.param_groups</code>.
|
|
Most learning rate schedule implementations I’ve come across do access this and change ‘lr’.</p>
|
|
<h3>States</h3>
|
|
<p>Optimizer maintains states (a dictionary) for each parameter (a tensor), in a dictionary <code>optimizer.state</code>.
|
|
This is where the optimizer maintains things like exponential averages.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">59</span><span></span><span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">Dict</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">Any</span>
|
|
<span class="lineno">60</span>
|
|
<span class="lineno">61</span><span class="kn">import</span> <span class="nn">torch</span>
|
|
<span class="lineno">62</span><span class="kn">from</span> <span class="nn">torch</span> <span class="kn">import</span> <span class="n">nn</span>
|
|
<span class="lineno">63</span><span class="kn">from</span> <span class="nn">torch.optim.optimizer</span> <span class="kn">import</span> <span class="n">Optimizer</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-1'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-1'>#</a>
|
|
</div>
|
|
<h2>Base class for <em>Adam</em> and extensions</h2>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">66</span><span class="k">class</span> <span class="nc">GenericAdaptiveOptimizer</span><span class="p">(</span><span class="n">Optimizer</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-2'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-2'>#</a>
|
|
</div>
|
|
<h3>Initialize</h3>
|
|
<ul>
|
|
<li><code>params</code> is the collection of parameters or set of parameter groups.</li>
|
|
<li><code>defaults</code> a dictionary of default hyper-parameters</li>
|
|
<li>‘lr` is the learning rate, $\alpha$</li>
|
|
<li><code>betas</code> is the tuple $(\beta_1, \beta_2)$</li>
|
|
<li><code>eps</code> is $\epsilon$</li>
|
|
</ul>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">71</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">params</span><span class="p">,</span> <span class="n">defaults</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="n">Any</span><span class="p">],</span> <span class="n">lr</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">betas</span><span class="p">:</span> <span class="n">Tuple</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="nb">float</span><span class="p">],</span> <span class="n">eps</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-3'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-3'>#</a>
|
|
</div>
|
|
<p>Check the hyper-parameters</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">83</span> <span class="k">if</span> <span class="ow">not</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">lr</span><span class="p">:</span>
|
|
<span class="lineno">84</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Invalid learning rate: </span><span class="si">{</span><span class="n">lr</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
|
|
<span class="lineno">85</span> <span class="k">if</span> <span class="ow">not</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">eps</span><span class="p">:</span>
|
|
<span class="lineno">86</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Invalid epsilon value: </span><span class="si">{</span><span class="n">eps</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
|
|
<span class="lineno">87</span> <span class="k">if</span> <span class="ow">not</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">betas</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o"><</span> <span class="mf">1.0</span><span class="p">:</span>
|
|
<span class="lineno">88</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Invalid beta parameter at index 0: </span><span class="si">{</span><span class="n">betas</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
|
|
<span class="lineno">89</span> <span class="k">if</span> <span class="ow">not</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">betas</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o"><</span> <span class="mf">1.0</span><span class="p">:</span>
|
|
<span class="lineno">90</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Invalid beta parameter at index 1: </span><span class="si">{</span><span class="n">betas</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-4'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-4'>#</a>
|
|
</div>
|
|
<p>Add the hyper-parameters to the defaults</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">93</span> <span class="n">defaults</span><span class="o">.</span><span class="n">update</span><span class="p">(</span><span class="nb">dict</span><span class="p">(</span><span class="n">lr</span><span class="o">=</span><span class="n">lr</span><span class="p">,</span> <span class="n">betas</span><span class="o">=</span><span class="n">betas</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="n">eps</span><span class="p">))</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-5'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-5'>#</a>
|
|
</div>
|
|
<p>Initialize the PyTorch optimizer.
|
|
This will create parameter groups with the default hyper-parameters</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">96</span> <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">params</span><span class="p">,</span> <span class="n">defaults</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-6'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-6'>#</a>
|
|
</div>
|
|
<h3>Initialize state for a given parameter tensor</h3>
|
|
<p>This should be overridden with code to initialize <code>state</code> for parameters <code>param</code>.
|
|
<code>group</code> is the parameter group dictionary to which <code>param</code> belongs.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">98</span> <span class="k">def</span> <span class="nf">init_state</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">state</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="nb">any</span><span class="p">],</span> <span class="n">group</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="nb">any</span><span class="p">],</span> <span class="n">param</span><span class="p">:</span> <span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-7'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-7'>#</a>
|
|
</div>
|
|
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">105</span> <span class="k">pass</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-8'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-8'>#</a>
|
|
</div>
|
|
<h3>Take optimizer step on a parameter tensor</h3>
|
|
<p>This should be overridden and take the optimization step on <code>param</code> tensor $\theta$,
|
|
where <code>grad</code> is the gradient for that parameter, $g_t$,
|
|
<code>state</code> is the optimizer state dictionary for that parameter, and
|
|
<code>group</code> is the parameter group dictionary <code>param</code> belongs to.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">107</span> <span class="k">def</span> <span class="nf">step_param</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">state</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="nb">any</span><span class="p">],</span> <span class="n">group</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="nb">any</span><span class="p">],</span> <span class="n">grad</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">param</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-9'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-9'>#</a>
|
|
</div>
|
|
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">116</span> <span class="k">pass</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-10'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-10'>#</a>
|
|
</div>
|
|
<h3>Optimizer step</h3>
|
|
<p>We have created a template method that does the common stuff every <em>Adam</em> based optimizer needs.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">118</span> <span class="nd">@torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">()</span>
|
|
<span class="lineno">119</span> <span class="k">def</span> <span class="nf">step</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">closure</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-11'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-11'>#</a>
|
|
</div>
|
|
<p>Calculate loss.</p>
|
|
<p>🤔 I’m not sure when you need this. I guess it’s if you define a function that
|
|
calculates the loss, does <code>loss.backward</code> and return the loss, instead of calling
|
|
it on your own you could pass it to <code>optimizer.step</code>. 🤷♂️</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">130</span> <span class="n">loss</span> <span class="o">=</span> <span class="kc">None</span>
|
|
<span class="lineno">131</span> <span class="k">if</span> <span class="n">closure</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="lineno">132</span> <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">enable_grad</span><span class="p">():</span>
|
|
<span class="lineno">133</span> <span class="n">loss</span> <span class="o">=</span> <span class="n">closure</span><span class="p">()</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-12'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-12'>#</a>
|
|
</div>
|
|
<p>Iterate through the parameter groups</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">136</span> <span class="k">for</span> <span class="n">group</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">param_groups</span><span class="p">:</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-13'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-13'>#</a>
|
|
</div>
|
|
<p>Iterate through the parameters in the parameter group</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">138</span> <span class="k">for</span> <span class="n">param</span> <span class="ow">in</span> <span class="n">group</span><span class="p">[</span><span class="s1">'params'</span><span class="p">]:</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-14'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-14'>#</a>
|
|
</div>
|
|
<p>Skip if the parameter has no gradient</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">140</span> <span class="k">if</span> <span class="n">param</span><span class="o">.</span><span class="n">grad</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="lineno">141</span> <span class="k">continue</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-15'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-15'>#</a>
|
|
</div>
|
|
<p>Get the gradient tensor</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">143</span> <span class="n">grad</span> <span class="o">=</span> <span class="n">param</span><span class="o">.</span><span class="n">grad</span><span class="o">.</span><span class="n">data</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-16'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-16'>#</a>
|
|
</div>
|
|
<p>We don’t handle sparse gradients</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">145</span> <span class="k">if</span> <span class="n">grad</span><span class="o">.</span><span class="n">is_sparse</span><span class="p">:</span>
|
|
<span class="lineno">146</span> <span class="k">raise</span> <span class="ne">RuntimeError</span><span class="p">(</span><span class="s1">'GenericAdaptiveOptimizer does not support sparse gradients,'</span>
|
|
<span class="lineno">147</span> <span class="s1">' please consider SparseAdam instead'</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-17'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-17'>#</a>
|
|
</div>
|
|
<p>Get the state for the parameter</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">150</span> <span class="n">state</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">state</span><span class="p">[</span><span class="n">param</span><span class="p">]</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-18'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-18'>#</a>
|
|
</div>
|
|
<p>Initialize the state if state is uninitialized</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">153</span> <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">state</span><span class="p">)</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
|
|
<span class="lineno">154</span> <span class="bp">self</span><span class="o">.</span><span class="n">init_state</span><span class="p">(</span><span class="n">state</span><span class="p">,</span> <span class="n">group</span><span class="p">,</span> <span class="n">param</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-19'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-19'>#</a>
|
|
</div>
|
|
<p>Take the optimization step on the parameter</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">157</span> <span class="bp">self</span><span class="o">.</span><span class="n">step_param</span><span class="p">(</span><span class="n">state</span><span class="p">,</span> <span class="n">group</span><span class="p">,</span> <span class="n">grad</span><span class="p">,</span> <span class="n">param</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-20'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-20'>#</a>
|
|
</div>
|
|
<p>Return the loss, calculated from closure</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">160</span> <span class="k">return</span> <span class="n">loss</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-21'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-21'>#</a>
|
|
</div>
|
|
<h2>L2 Weight decay</h2>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">163</span><span class="k">class</span> <span class="nc">WeightDecay</span><span class="p">:</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-22'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-22'>#</a>
|
|
</div>
|
|
<h3>Initialize weight decay</h3>
|
|
<ul>
|
|
<li><code>weight_decay</code> is the decay coefficient</li>
|
|
<li><code>weight_decouple</code> is a flag indicating whether to add the weight decay to the gradient or directly
|
|
decay from the parameter. If added to the gradient it will go through the normal optimizer update.</li>
|
|
<li><code>absolute</code> this flag indicates whether the weight decay coefficient is absolute. This is applicable
|
|
when the decay is performed directly on the parameter. If this is false the actual decay is
|
|
<code>weight_decay</code> * <code>learning_rate</code>.</li>
|
|
</ul>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">168</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">weight_decay</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.</span><span class="p">,</span> <span class="n">weight_decouple</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="n">absolute</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-23'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-23'>#</a>
|
|
</div>
|
|
<p>Check hyper-parameters</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">180</span> <span class="k">if</span> <span class="ow">not</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">weight_decay</span><span class="p">:</span>
|
|
<span class="lineno">181</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Invalid weight_decay value: </span><span class="si">{</span><span class="n">weight_decay</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
|
|
<span class="lineno">182</span>
|
|
<span class="lineno">183</span> <span class="bp">self</span><span class="o">.</span><span class="n">absolute</span> <span class="o">=</span> <span class="n">absolute</span>
|
|
<span class="lineno">184</span> <span class="bp">self</span><span class="o">.</span><span class="n">weight_decouple</span> <span class="o">=</span> <span class="n">weight_decouple</span>
|
|
<span class="lineno">185</span> <span class="bp">self</span><span class="o">.</span><span class="n">weight_decay</span> <span class="o">=</span> <span class="n">weight_decay</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-24'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-24'>#</a>
|
|
</div>
|
|
<p>Return defaults for parameter groups</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">187</span> <span class="k">def</span> <span class="nf">defaults</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-25'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-25'>#</a>
|
|
</div>
|
|
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">191</span> <span class="k">return</span> <span class="nb">dict</span><span class="p">(</span><span class="n">weight_decay</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">weight_decay</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-26'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-26'>#</a>
|
|
</div>
|
|
<h3>Perform weight decay and return the gradient</h3>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">193</span> <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">param</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">,</span> <span class="n">grad</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">group</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="nb">any</span><span class="p">]):</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-27'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-27'>#</a>
|
|
</div>
|
|
<p>If we are doing the decay on the parameter directly</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">199</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">weight_decouple</span><span class="p">:</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-28'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-28'>#</a>
|
|
</div>
|
|
<p>If the weight decay coefficient is absolute</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">201</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">absolute</span><span class="p">:</span>
|
|
<span class="lineno">202</span> <span class="n">param</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">mul_</span><span class="p">(</span><span class="mf">1.0</span> <span class="o">-</span> <span class="n">group</span><span class="p">[</span><span class="s1">'weight_decay'</span><span class="p">])</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-29'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-29'>#</a>
|
|
</div>
|
|
<p>Otherwise,</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">204</span> <span class="k">else</span><span class="p">:</span>
|
|
<span class="lineno">205</span> <span class="n">param</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">mul_</span><span class="p">(</span><span class="mf">1.0</span> <span class="o">-</span> <span class="n">group</span><span class="p">[</span><span class="s1">'lr'</span><span class="p">]</span> <span class="o">*</span> <span class="n">group</span><span class="p">[</span><span class="s1">'weight_decay'</span><span class="p">])</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-30'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-30'>#</a>
|
|
</div>
|
|
<p>Return the unmodified gradient</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">207</span> <span class="k">return</span> <span class="n">grad</span>
|
|
<span class="lineno">208</span> <span class="k">else</span><span class="p">:</span>
|
|
<span class="lineno">209</span> <span class="k">if</span> <span class="n">group</span><span class="p">[</span><span class="s1">'weight_decay'</span><span class="p">]</span> <span class="o">!=</span> <span class="mi">0</span><span class="p">:</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-31'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-31'>#</a>
|
|
</div>
|
|
<p>Add the weight decay to the gradient and return the modified gradient</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">211</span> <span class="k">return</span> <span class="n">grad</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">param</span><span class="o">.</span><span class="n">data</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="n">group</span><span class="p">[</span><span class="s1">'weight_decay'</span><span class="p">])</span>
|
|
<span class="lineno">212</span> <span class="k">else</span><span class="p">:</span>
|
|
<span class="lineno">213</span> <span class="k">return</span> <span class="n">grad</span></pre></div>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
<script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.4/MathJax.js?config=TeX-AMS_HTML">
|
|
</script>
|
|
<!-- MathJax configuration -->
|
|
<script type="text/x-mathjax-config">
|
|
MathJax.Hub.Config({
|
|
tex2jax: {
|
|
inlineMath: [ ['$','$'] ],
|
|
displayMath: [ ['$$','$$'] ],
|
|
processEscapes: true,
|
|
processEnvironments: true
|
|
},
|
|
// Center justify equations in code and markdown cells. Elsewhere
|
|
// we use CSS to left justify single line equations in code cells.
|
|
displayAlign: 'center',
|
|
"HTML-CSS": { fonts: ["TeX"] }
|
|
});
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
</script>
|
|
</body>
|
|
</html> |