mirror of
https://github.com/labmlai/annotated_deep_learning_paper_implementations.git
synced 2025-08-15 10:11:39 +08:00
137 lines
6.6 KiB
HTML
137 lines
6.6 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 a simple PyTorch implementation/tutorial of the paper Generalization through Memorization: Nearest Neighbor Language Models using FAISS. It runs a kNN model on the final transformer layer embeddings to improve the loss of transformer based language models. It's also great for domain adaptation without pre-training."/>
|
|
|
|
<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="k-Nearest Neighbor Language Models"/>
|
|
<meta name="twitter:description" content="This is a simple PyTorch implementation/tutorial of the paper Generalization through Memorization: Nearest Neighbor Language Models using FAISS. It runs a kNN model on the final transformer layer embeddings to improve the loss of transformer based language models. It's also great for domain adaptation without pre-training."/>
|
|
<meta name="twitter:site" content="@labmlai"/>
|
|
<meta name="twitter:creator" content="@labmlai"/>
|
|
|
|
<meta property="og:url" content="https://nn.labml.ai/transformers/knn/index.html"/>
|
|
<meta property="og:title" content="k-Nearest Neighbor Language Models"/>
|
|
<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="k-Nearest Neighbor Language Models"/>
|
|
<meta property="og:description" content="This is a simple PyTorch implementation/tutorial of the paper Generalization through Memorization: Nearest Neighbor Language Models using FAISS. It runs a kNN model on the final transformer layer embeddings to improve the loss of transformer based language models. It's also great for domain adaptation without pre-training."/>
|
|
|
|
<title>k-Nearest Neighbor Language Models</title>
|
|
<link rel="shortcut icon" href="/icon.png"/>
|
|
<link rel="stylesheet" href="../../pylit.css">
|
|
<link rel="canonical" href="https://nn.labml.ai/transformers/knn/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">knn</a>
|
|
</p>
|
|
<p>
|
|
|
|
<a href="https://github.com/lab-ml/labml_nn/tree/master/labml_nn/transformers/knn/__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>k-Nearest Neighbor Language Models</h1>
|
|
<p>This is a <a href="https://pytorch.org">PyTorch</a> implementation of the paper
|
|
<a href="https://arxiv.org/abs/1911.00172">Generalization through Memorization: Nearest Neighbor Language Models</a>.
|
|
It uses k-nearest neighbors to improve perplexity of autoregressive transformer models.</p>
|
|
<p>An autoregressive language model estimates $p(w_t | \color{yellowgreen}{c_t})$,
|
|
where $w_t$ is the token at step $t$
|
|
and $c_t$ is the context, $\color{yellowgreen}{c_t} = (w_1, w_2, …, w_{t-1})$.</p>
|
|
<p>This paper, improves $p(w_t | \color{yellowgreen}{c_t})$ using a k-nearest neighbor search
|
|
on key-value pairs $\big(f(c_i), w_i\big)$, with search key $f(\color{yellowgreen}{c_t})$.
|
|
Here $f(\color{yellowgreen}{c_t})$ is an embedding of the context $\color{yellowgreen}{c_t}$.
|
|
The paper (and this implementation) uses the <strong>input to the feed-forward layer of the
|
|
final layer of the transformer</strong> as $f(\color{yellowgreen}{c_t})$.</p>
|
|
<p>We use <a href="https://github.com/facebookresearch/faiss">FAISS</a> to index $f(c_i)$.</p>
|
|
<h3>Implementation</h3>
|
|
<p>So to run $k$NN-LM we need to:</p>
|
|
<ul>
|
|
<li><a href="train_model.html">Train a transformer model</a></li>
|
|
<li><a href="build_index.html">Build an index</a> of $\big(f(c_i), w_i\big)$</li>
|
|
<li><a href="eval_knn.html">Evaluate kNN-ML</a> using $k$NN seach on $\big(f(c_i), w_i\big)$
|
|
with $f(\color{yellowgreen}{c_t})$</li>
|
|
</ul>
|
|
<p>This experiment uses a small dataset so that we can run this without using up a few hundred giga-bytes
|
|
of disk space for the index.</p>
|
|
<p>The official implementation of $k$NN-LM can be found <a href="https://github.com/urvashik/knnlm">here</a>.</p>
|
|
</div>
|
|
<div class='code'>
|
|
<div class="highlight"><pre></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> |