mirror of
https://github.com/labmlai/annotated_deep_learning_paper_implementations.git
synced 2025-08-14 01:13:00 +08:00
384 lines
24 KiB
HTML
384 lines
24 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="This is an annotated implementation/tutorial of the Masked Language Model in PyTorch."/>
|
|
|
|
<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="Masked Language Model"/>
|
|
<meta name="twitter:description" content="This is an annotated implementation/tutorial of the Masked Language Model in PyTorch."/>
|
|
<meta name="twitter:site" content="@labmlai"/>
|
|
<meta name="twitter:creator" content="@labmlai"/>
|
|
|
|
<meta property="og:url" content="https://nn.labml.ai/transformers/mlm/index.html"/>
|
|
<meta property="og:title" content="Masked Language Model"/>
|
|
<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="Masked Language Model"/>
|
|
<meta property="og:description" content="This is an annotated implementation/tutorial of the Masked Language Model in PyTorch."/>
|
|
|
|
<title>Masked Language Model</title>
|
|
<link rel="shortcut icon" href="/icon.png"/>
|
|
<link rel="stylesheet" href="../../pylit.css">
|
|
<link rel="canonical" href="https://nn.labml.ai/transformers/mlm/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">transformers</a>
|
|
<a class="parent" href="index.html">mlm</a>
|
|
</p>
|
|
<p>
|
|
|
|
<a href="https://github.com/lab-ml/labml_nn/tree/master/labml_nn/transformers/mlm/__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://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>Masked Language Model (MLM)</h1>
|
|
<p>This is a <a href="https://pytorch.org">PyTorch</a> implementation of the Masked Language Model (MLM)
|
|
used to pre-train the BERT model introduced in the paper
|
|
<a href="https://arxiv.org/abs/1810.04805">BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding</a>.</p>
|
|
<h2>BERT Pretraining</h2>
|
|
<p>BERT model is a transformer model.
|
|
The paper pre-trains the model using MLM and with next sentence prediction.
|
|
We have only implemented MLM here.</p>
|
|
<h3>Next sentence prediction</h3>
|
|
<p>In <em>next sentence prediction</em>, the model is given two sentences <code>A</code> and <code>B</code> and the model
|
|
makes a binary prediction whether <code>B</code> is the sentence that follows <code>A</code> in the actual text.
|
|
The model is fed with actual sentence pairs 50% of the time and random pairs 50% of the time.
|
|
This classification is done while applying MLM. <em>We haven’t implemented this here.</em></p>
|
|
<h2>Masked LM</h2>
|
|
<p>This masks a percentage of tokens at random and trains the model to predict
|
|
the masked tokens.
|
|
They <strong>mask 15% of the tokens</strong> by replacing them with a special <code>[MASK]</code> token.</p>
|
|
<p>The loss is computed on predicting the masked tokens only.
|
|
This causes a problem during fine-tuning and actual usage since there are no <code>[MASK]</code> tokens
|
|
at that time.
|
|
Therefore we might not get any meaningful representations.</p>
|
|
<p>To overcome this <strong>10% of the masked tokens are replaced with the original token</strong>,
|
|
and another <strong>10% of the masked tokens are replaced with a random token</strong>.
|
|
This trains the model to give representations about the actual token whether or not the
|
|
input token at that position is a <code>[MASK]</code>.
|
|
And replacing with a random token causes it to
|
|
give a representation that has information from the context as well;
|
|
because it has to use the context to fix randomly replaced tokens.</p>
|
|
<h2>Training</h2>
|
|
<p>MLMs are harder to train than autoregressive models because they have a smaller training signal.
|
|
i.e. only a small percentage of predictions are trained per sample.</p>
|
|
<p>Another problem is since the model is bidirectional, any token can see any other token.
|
|
This makes the “credit assignment” harder.
|
|
Let’s say you have the character level model trying to predict <code>home *s where i want to be</code>.
|
|
At least during the early stages of the training, it’ll be super hard to figure out why the
|
|
replacement for <code>*</code> should be <code>i</code>, it could be anything from the whole sentence.
|
|
Whilst, in an autoregressive setting the model will only have to use <code>h</code> to predict <code>o</code> and
|
|
<code>hom</code> to predict <code>e</code> and so on. So the model will initially start predicting with a shorter context first
|
|
and then learn to use longer contexts later.
|
|
Since MLMs have this problem it’s a lot faster to train if you start with a smaller sequence length
|
|
initially and then use a longer sequence length later.</p>
|
|
<p>Here is <a href="experiment.html">the training code</a> for a simple MLM model.</p>
|
|
<p><a href="https://app.labml.ai/run/3a6d22b6c67111ebb03d6764d13a38d1"><img alt="View Run" src="https://img.shields.io/badge/labml-experiment-brightgreen" /></a></p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">67</span><span></span><span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">List</span>
|
|
<span class="lineno">68</span>
|
|
<span class="lineno">69</span><span class="kn">import</span> <span class="nn">torch</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>Masked LM (MLM)</h2>
|
|
<p>This class implements the masking procedure for a given batch of token sequences.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">72</span><span class="k">class</span> <span class="nc">MLM</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>
|
|
<ul>
|
|
<li><code>padding_token</code> is the padding token `[PAD].
|
|
We will use this to mark the labels that shouldn’t be used for loss calculation.</li>
|
|
<li><code>mask_token</code> is the masking token <code>[MASK]</code>.</li>
|
|
<li><code>no_mask_tokens</code> is a list of tokens that should not be masked.
|
|
This is useful if we are training the MLM with another task like classification at the same time,
|
|
and we have tokens such as <code>[CLS]</code> that shouldn’t be masked.</li>
|
|
<li><code>n_tokens</code> total number of tokens (used for generating random tokens)</li>
|
|
<li><code>masking_prob</code> is the masking probability</li>
|
|
<li><code>randomize_prob</code> is the probability of replacing with a random token</li>
|
|
<li><code>no_change_prob</code> is the probability of replacing with original token</li>
|
|
</ul>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">79</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="o">*</span><span class="p">,</span>
|
|
<span class="lineno">80</span> <span class="n">padding_token</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span> <span class="n">mask_token</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span> <span class="n">no_mask_tokens</span><span class="p">:</span> <span class="n">List</span><span class="p">[</span><span class="nb">int</span><span class="p">],</span> <span class="n">n_tokens</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
|
|
<span class="lineno">81</span> <span class="n">masking_prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.15</span><span class="p">,</span> <span class="n">randomize_prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.1</span><span class="p">,</span> <span class="n">no_change_prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.1</span><span class="p">,</span>
|
|
<span class="lineno">82</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>
|
|
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">95</span> <span class="bp">self</span><span class="o">.</span><span class="n">n_tokens</span> <span class="o">=</span> <span class="n">n_tokens</span>
|
|
<span class="lineno">96</span> <span class="bp">self</span><span class="o">.</span><span class="n">no_change_prob</span> <span class="o">=</span> <span class="n">no_change_prob</span>
|
|
<span class="lineno">97</span> <span class="bp">self</span><span class="o">.</span><span class="n">randomize_prob</span> <span class="o">=</span> <span class="n">randomize_prob</span>
|
|
<span class="lineno">98</span> <span class="bp">self</span><span class="o">.</span><span class="n">masking_prob</span> <span class="o">=</span> <span class="n">masking_prob</span>
|
|
<span class="lineno">99</span> <span class="bp">self</span><span class="o">.</span><span class="n">no_mask_tokens</span> <span class="o">=</span> <span class="n">no_mask_tokens</span> <span class="o">+</span> <span class="p">[</span><span class="n">padding_token</span><span class="p">,</span> <span class="n">mask_token</span><span class="p">]</span>
|
|
<span class="lineno">100</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding_token</span> <span class="o">=</span> <span class="n">padding_token</span>
|
|
<span class="lineno">101</span> <span class="bp">self</span><span class="o">.</span><span class="n">mask_token</span> <span class="o">=</span> <span class="n">mask_token</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-4'>
|
|
<div class='docs doc-strings'>
|
|
<div class='section-link'>
|
|
<a href='#section-4'>#</a>
|
|
</div>
|
|
<ul>
|
|
<li><code>x</code> is the batch of input token sequences.
|
|
It’s a tensor of type <code>long</code> with shape <code>[seq_len, batch_size]</code>.</li>
|
|
</ul>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">103</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">x</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-5'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-5'>#</a>
|
|
</div>
|
|
<p>Mask <code>masking_prob</code> of tokens</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">110</span> <span class="n">full_mask</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">x</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">device</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">masking_prob</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-6'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-6'>#</a>
|
|
</div>
|
|
<p>Unmask <code>no_mask_tokens</code></p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">112</span> <span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">no_mask_tokens</span><span class="p">:</span>
|
|
<span class="lineno">113</span> <span class="n">full_mask</span> <span class="o">&=</span> <span class="n">x</span> <span class="o">!=</span> <span class="n">t</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-7'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-7'>#</a>
|
|
</div>
|
|
<p>A mask for tokens to be replaced with original tokens</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">116</span> <span class="n">unchanged</span> <span class="o">=</span> <span class="n">full_mask</span> <span class="o">&</span> <span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">x</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">device</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">no_change_prob</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-8'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-8'>#</a>
|
|
</div>
|
|
<p>A mask for tokens to be replaced with a random token</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">118</span> <span class="n">random_token_mask</span> <span class="o">=</span> <span class="n">full_mask</span> <span class="o">&</span> <span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">x</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">device</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">randomize_prob</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>
|
|
<p>Indexes of tokens to be replaced with random tokens</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">120</span> <span class="n">random_token_idx</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">nonzero</span><span class="p">(</span><span class="n">random_token_mask</span><span class="p">,</span> <span class="n">as_tuple</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span></pre></div>
|
|
</div>
|
|
</div>
|
|
<div class='section' id='section-10'>
|
|
<div class='docs'>
|
|
<div class='section-link'>
|
|
<a href='#section-10'>#</a>
|
|
</div>
|
|
<p>Random tokens for each of the locations</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">122</span> <span class="n">random_tokens</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">n_tokens</span><span class="p">,</span> <span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">random_token_idx</span><span class="p">[</span><span class="mi">0</span><span class="p">]),),</span> <span class="n">device</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">device</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>The final set of tokens that are going to be replaced by <code>[MASK]</code></p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">124</span> <span class="n">mask</span> <span class="o">=</span> <span class="n">full_mask</span> <span class="o">&</span> <span class="o">~</span><span class="n">random_token_mask</span> <span class="o">&</span> <span class="o">~</span><span class="n">unchanged</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>Make a clone of the input for the labels</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">127</span> <span class="n">y</span> <span class="o">=</span> <span class="n">x</span><span class="o">.</span><span class="n">clone</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>Replace with <code>[MASK]</code> tokens;
|
|
note that this doesn’t include the tokens that will have the original token unchanged and
|
|
those that get replace with a random token.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">132</span> <span class="n">x</span><span class="o">.</span><span class="n">masked_fill_</span><span class="p">(</span><span class="n">mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mask_token</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>Assign random tokens</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">134</span> <span class="n">x</span><span class="p">[</span><span class="n">random_token_idx</span><span class="p">]</span> <span class="o">=</span> <span class="n">random_tokens</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>Assign token <code>[PAD]</code> to all the other locations in the labels.
|
|
The labels equal to <code>[PAD]</code> will not be used in the loss.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">138</span> <span class="n">y</span><span class="o">.</span><span class="n">masked_fill_</span><span class="p">(</span><span class="o">~</span><span class="n">full_mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding_token</span><span class="p">)</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>Return the masked input and the labels</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre><span class="lineno">141</span> <span class="k">return</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</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>
|
|
<script>
|
|
function handleImages() {
|
|
var images = document.querySelectorAll('p>img')
|
|
|
|
console.log(images);
|
|
for (var i = 0; i < images.length; ++i) {
|
|
handleImage(images[i])
|
|
}
|
|
}
|
|
|
|
function handleImage(img) {
|
|
img.parentElement.style.textAlign = 'center'
|
|
|
|
var modal = document.createElement('div')
|
|
modal.id = 'modal'
|
|
|
|
var modalContent = document.createElement('div')
|
|
modal.appendChild(modalContent)
|
|
|
|
var modalImage = document.createElement('img')
|
|
modalContent.appendChild(modalImage)
|
|
|
|
var span = document.createElement('span')
|
|
span.classList.add('close')
|
|
span.textContent = 'x'
|
|
modal.appendChild(span)
|
|
|
|
img.onclick = function () {
|
|
console.log('clicked')
|
|
document.body.appendChild(modal)
|
|
modalImage.src = img.src
|
|
}
|
|
|
|
span.onclick = function () {
|
|
document.body.removeChild(modal)
|
|
}
|
|
}
|
|
|
|
handleImages()
|
|
</script>
|
|
</body>
|
|
</html> |