mirror of
https://github.com/labmlai/annotated_deep_learning_paper_implementations.git
synced 2025-08-14 01:13:00 +08:00
📚 ppo intro
This commit is contained in:
Varuna Jayasiri
@ -75,14 +75,22 @@
|
||||
<h1>Proximal Policy Optimization (PPO)</h1>
|
||||
<p>This is a <a href="https://pytorch.org">PyTorch</a> implementation of
|
||||
<a href="https://arxiv.org/abs/1707.06347">Proximal Policy Optimization - PPO</a>.</p>
|
||||
<p>PPO is a policy gradient method for reinforcement learning.
|
||||
Simple policy gradient methods one do a single gradient update per sample (or a set of samples).
|
||||
Doing multiple gradient steps for a singe sample causes problems
|
||||
because the policy deviates too much producing a bad policy.
|
||||
PPO lets us do multiple gradient updates per sample by trying to keep the
|
||||
policy close to the policy that was used to sample data.
|
||||
It does so by clipping gradient flow if the updated policy
|
||||
is not close to the policy used to sample the data.</p>
|
||||
<p>You can find an experiment that uses it <a href="experiment.html">here</a>.
|
||||
The experiment uses <a href="gae.html">Generalized Advantage Estimation</a>.</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">17</span><span></span><span class="kn">import</span> <span class="nn">torch</span>
|
||||
<span class="lineno">18</span>
|
||||
<span class="lineno">19</span><span class="kn">from</span> <span class="nn">labml_helpers.module</span> <span class="kn">import</span> <span class="n">Module</span>
|
||||
<span class="lineno">20</span><span class="kn">from</span> <span class="nn">labml_nn.rl.ppo.gae</span> <span class="kn">import</span> <span class="n">GAE</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">26</span><span></span><span class="kn">import</span> <span class="nn">torch</span>
|
||||
<span class="lineno">27</span>
|
||||
<span class="lineno">28</span><span class="kn">from</span> <span class="nn">labml_helpers.module</span> <span class="kn">import</span> <span class="n">Module</span>
|
||||
<span class="lineno">29</span><span class="kn">from</span> <span class="nn">labml_nn.rl.ppo.gae</span> <span class="kn">import</span> <span class="n">GAE</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-1'>
|
||||
@ -91,6 +99,7 @@ The experiment uses <a href="gae.html">Generalized Advantage Estimation</a>.</p>
|
||||
<a href='#section-1'>#</a>
|
||||
</div>
|
||||
<h2>PPO Loss</h2>
|
||||
<p>Here’s how the PPO update rule is derived.</p>
|
||||
<p>We want to maximize policy reward
|
||||
<script type="math/tex; mode=display">\max_\theta J(\pi_\theta) =
|
||||
\mathop{\mathbb{E}}_{\tau \sim \pi_\theta}\Biggl[\sum_{t=0}^\infty \gamma^t r_t \Biggr]</script>
|
||||
@ -186,7 +195,7 @@ J(\pi_\theta) - J(\pi_{\theta_{OLD}})
|
||||
</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">23</span><span class="k">class</span> <span class="nc">ClippedPPOLoss</span><span class="p">(</span><span class="n">Module</span><span class="p">):</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">32</span><span class="k">class</span> <span class="nc">ClippedPPOLoss</span><span class="p">(</span><span class="n">Module</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-2'>
|
||||
@ -197,8 +206,8 @@ J(\pi_\theta) - J(\pi_{\theta_{OLD}})
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">122</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="lineno">123</span> <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">133</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="lineno">134</span> <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-3'>
|
||||
@ -209,8 +218,8 @@ J(\pi_\theta) - J(\pi_{\theta_{OLD}})
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">125</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">log_pi</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">sampled_log_pi</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="lineno">126</span> <span class="n">advantage</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">clip</span><span class="p">:</span> <span class="nb">float</span><span class="p">)</span> <span class="o">-></span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">:</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">136</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">log_pi</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">sampled_log_pi</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="lineno">137</span> <span class="n">advantage</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">clip</span><span class="p">:</span> <span class="nb">float</span><span class="p">)</span> <span class="o">-></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-4'>
|
||||
@ -222,7 +231,7 @@ J(\pi_\theta) - J(\pi_{\theta_{OLD}})
|
||||
<em>this is different from rewards</em> $r_t$.</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">129</span> <span class="n">ratio</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="n">log_pi</span> <span class="o">-</span> <span class="n">sampled_log_pi</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">140</span> <span class="n">ratio</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="n">log_pi</span> <span class="o">-</span> <span class="n">sampled_log_pi</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-5'>
|
||||
@ -230,7 +239,8 @@ J(\pi_\theta) - J(\pi_{\theta_{OLD}})
|
||||
<div class='section-link'>
|
||||
<a href='#section-5'>#</a>
|
||||
</div>
|
||||
<p>
|
||||
<h3>Cliping the policy ratio</h3>
|
||||
<p>
|
||||
<script type="math/tex; mode=display">\begin{align}
|
||||
\mathcal{L}^{CLIP}(\theta) =
|
||||
\mathbb{E}_{a_t, s_t \sim \pi_{\theta{OLD}}} \biggl[
|
||||
@ -257,14 +267,14 @@ Large deviation can cause performance collapse;
|
||||
but it reduces variance a lot.</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">156</span> <span class="n">clipped_ratio</span> <span class="o">=</span> <span class="n">ratio</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span><span class="nb">min</span><span class="o">=</span><span class="mf">1.0</span> <span class="o">-</span> <span class="n">clip</span><span class="p">,</span>
|
||||
<span class="lineno">157</span> <span class="nb">max</span><span class="o">=</span><span class="mf">1.0</span> <span class="o">+</span> <span class="n">clip</span><span class="p">)</span>
|
||||
<span class="lineno">158</span> <span class="n">policy_reward</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">ratio</span> <span class="o">*</span> <span class="n">advantage</span><span class="p">,</span>
|
||||
<span class="lineno">159</span> <span class="n">clipped_ratio</span> <span class="o">*</span> <span class="n">advantage</span><span class="p">)</span>
|
||||
<span class="lineno">160</span>
|
||||
<span class="lineno">161</span> <span class="bp">self</span><span class="o">.</span><span class="n">clip_fraction</span> <span class="o">=</span> <span class="p">(</span><span class="nb">abs</span><span class="p">((</span><span class="n">ratio</span> <span class="o">-</span> <span class="mf">1.0</span><span class="p">))</span> <span class="o">></span> <span class="n">clip</span><span class="p">)</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span>
|
||||
<span class="lineno">162</span>
|
||||
<span class="lineno">163</span> <span class="k">return</span> <span class="o">-</span><span class="n">policy_reward</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">169</span> <span class="n">clipped_ratio</span> <span class="o">=</span> <span class="n">ratio</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span><span class="nb">min</span><span class="o">=</span><span class="mf">1.0</span> <span class="o">-</span> <span class="n">clip</span><span class="p">,</span>
|
||||
<span class="lineno">170</span> <span class="nb">max</span><span class="o">=</span><span class="mf">1.0</span> <span class="o">+</span> <span class="n">clip</span><span class="p">)</span>
|
||||
<span class="lineno">171</span> <span class="n">policy_reward</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">ratio</span> <span class="o">*</span> <span class="n">advantage</span><span class="p">,</span>
|
||||
<span class="lineno">172</span> <span class="n">clipped_ratio</span> <span class="o">*</span> <span class="n">advantage</span><span class="p">)</span>
|
||||
<span class="lineno">173</span>
|
||||
<span class="lineno">174</span> <span class="bp">self</span><span class="o">.</span><span class="n">clip_fraction</span> <span class="o">=</span> <span class="p">(</span><span class="nb">abs</span><span class="p">((</span><span class="n">ratio</span> <span class="o">-</span> <span class="mf">1.0</span><span class="p">))</span> <span class="o">></span> <span class="n">clip</span><span class="p">)</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span>
|
||||
<span class="lineno">175</span>
|
||||
<span class="lineno">176</span> <span class="k">return</span> <span class="o">-</span><span class="n">policy_reward</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-6'>
|
||||
@ -273,6 +283,7 @@ Large deviation can cause performance collapse;
|
||||
<a href='#section-6'>#</a>
|
||||
</div>
|
||||
<h2>Clipped Value Function Loss</h2>
|
||||
<p>Similarly we clip the value function update also.</p>
|
||||
<p>
|
||||
<script type="math/tex; mode=display">\begin{align}
|
||||
V^{\pi_\theta}_{CLIP}(s_t)
|
||||
@ -289,7 +300,7 @@ V^{\pi_\theta}_{CLIP}(s_t)
|
||||
significantly from $V_{\theta_{OLD}}$.</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">166</span><span class="k">class</span> <span class="nc">ClippedValueFunctionLoss</span><span class="p">(</span><span class="n">Module</span><span class="p">):</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">179</span><span class="k">class</span> <span class="nc">ClippedValueFunctionLoss</span><span class="p">(</span><span class="n">Module</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-7'>
|
||||
@ -300,10 +311,10 @@ V^{\pi_\theta}_{CLIP}(s_t)
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">185</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">value</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">sampled_value</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">sampled_return</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">clip</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
|
||||
<span class="lineno">186</span> <span class="n">clipped_value</span> <span class="o">=</span> <span class="n">sampled_value</span> <span class="o">+</span> <span class="p">(</span><span class="n">value</span> <span class="o">-</span> <span class="n">sampled_value</span><span class="p">)</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span><span class="nb">min</span><span class="o">=-</span><span class="n">clip</span><span class="p">,</span> <span class="nb">max</span><span class="o">=</span><span class="n">clip</span><span class="p">)</span>
|
||||
<span class="lineno">187</span> <span class="n">vf_loss</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">max</span><span class="p">((</span><span class="n">value</span> <span class="o">-</span> <span class="n">sampled_return</span><span class="p">)</span> <span class="o">**</span> <span class="mi">2</span><span class="p">,</span> <span class="p">(</span><span class="n">clipped_value</span> <span class="o">-</span> <span class="n">sampled_return</span><span class="p">)</span> <span class="o">**</span> <span class="mi">2</span><span class="p">)</span>
|
||||
<span class="lineno">188</span> <span class="k">return</span> <span class="mf">0.5</span> <span class="o">*</span> <span class="n">vf_loss</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">200</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">value</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">sampled_value</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">sampled_return</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">clip</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
|
||||
<span class="lineno">201</span> <span class="n">clipped_value</span> <span class="o">=</span> <span class="n">sampled_value</span> <span class="o">+</span> <span class="p">(</span><span class="n">value</span> <span class="o">-</span> <span class="n">sampled_value</span><span class="p">)</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span><span class="nb">min</span><span class="o">=-</span><span class="n">clip</span><span class="p">,</span> <span class="nb">max</span><span class="o">=</span><span class="n">clip</span><span class="p">)</span>
|
||||
<span class="lineno">202</span> <span class="n">vf_loss</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">max</span><span class="p">((</span><span class="n">value</span> <span class="o">-</span> <span class="n">sampled_return</span><span class="p">)</span> <span class="o">**</span> <span class="mi">2</span><span class="p">,</span> <span class="p">(</span><span class="n">clipped_value</span> <span class="o">-</span> <span class="n">sampled_return</span><span class="p">)</span> <span class="o">**</span> <span class="mi">2</span><span class="p">)</span>
|
||||
<span class="lineno">203</span> <span class="k">return</span> <span class="mf">0.5</span> <span class="o">*</span> <span class="n">vf_loss</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
126
docs/rl/ppo/readme.html
Normal file
126
docs/rl/ppo/readme.html
Normal file
@ -0,0 +1,126 @@
|
||||
<!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=""/>
|
||||
|
||||
<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="Proximal Policy Optimization (PPO)"/>
|
||||
<meta name="twitter:description" content=""/>
|
||||
<meta name="twitter:site" content="@labmlai"/>
|
||||
<meta name="twitter:creator" content="@labmlai"/>
|
||||
|
||||
<meta property="og:url" content="https://nn.labml.ai/rl/ppo/readme.html"/>
|
||||
<meta property="og:title" content="Proximal Policy Optimization (PPO)"/>
|
||||
<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="Proximal Policy Optimization (PPO)"/>
|
||||
<meta property="og:description" content=""/>
|
||||
|
||||
<title>Proximal Policy Optimization (PPO)</title>
|
||||
<link rel="shortcut icon" href="/icon.png"/>
|
||||
<link rel="stylesheet" href="../../pylit.css">
|
||||
<link rel="canonical" href="https://nn.labml.ai/rl/ppo/readme.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">rl</a>
|
||||
<a class="parent" href="index.html">ppo</a>
|
||||
</p>
|
||||
<p>
|
||||
|
||||
<a href="https://github.com/lab-ml/labml_nn/tree/master/labml_nn/rl/ppo/readme.md">
|
||||
<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'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-0'>#</a>
|
||||
</div>
|
||||
<h1><a href="https://nn.labml.ai/rl/ppo/index.html">Proximal Policy Optimization (PPO)</a></h1>
|
||||
<p>This is a <a href="https://pytorch.org">PyTorch</a> implementation of
|
||||
<a href="https://arxiv.org/abs/1707.06347">Proximal Policy Optimization - PPO</a>.</p>
|
||||
<p>PPO is a policy gradient method for reinforcement learning.
|
||||
Simple policy gradient methods one do a single gradient update per sample (or a set of samples).
|
||||
Doing multiple gradient steps for a singe sample causes problems
|
||||
because the policy deviates too much producing a bad policy.
|
||||
PPO lets us do multiple gradient updates per sample by trying to keep the
|
||||
policy close to the policy that was used to sample data.
|
||||
It does so by clipping gradient flow if the updated policy
|
||||
is not close to the policy used to sample the data.</p>
|
||||
<p>You can find an experiment that uses it <a href="https://nn.labml.ai/rl/ppo/experiment.html">here</a>.
|
||||
The experiment uses <a href="https://nn.labml.ai/rl/ppo/gae.html">Generalized Advantage Estimation</a>.</p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
|
||||
</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>
|
||||
@ -120,7 +120,7 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/normalization/batch_norm/readme.html</loc>
|
||||
<lastmod>2021-02-02T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
@ -155,7 +155,7 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/index.html</loc>
|
||||
<lastmod>2021-02-07T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
@ -302,7 +302,7 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/index.html</loc>
|
||||
<lastmod>2021-02-07T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
@ -379,14 +379,14 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/switch/index.html</loc>
|
||||
<lastmod>2021-02-10T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/switch/readme.html</loc>
|
||||
<lastmod>2021-02-01T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
@ -414,21 +414,35 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/mha.html</loc>
|
||||
<lastmod>2021-02-17T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/compressive/index.html</loc>
|
||||
<lastmod>2021-02-17T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/compressive/experiment.html</loc>
|
||||
<lastmod>2021-02-17T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/compressive/index.html</loc>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/compressive/readme.html</loc>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/compressive/experiment.html</loc>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
@ -442,21 +456,21 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/xl/experiment.html</loc>
|
||||
<lastmod>2021-02-07T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/xl/index.html</loc>
|
||||
<lastmod>2021-02-10T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-18T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/transformers/xl/readme.html</loc>
|
||||
<lastmod>2021-02-07T16:30:00+00:00</lastmod>
|
||||
<lastmod>2021-02-19T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
147
docs/transformers/compressive/readme.html
Normal file
147
docs/transformers/compressive/readme.html
Normal file
@ -0,0 +1,147 @@
|
||||
<!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=""/>
|
||||
|
||||
<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="Compressive Transformer"/>
|
||||
<meta name="twitter:description" content=""/>
|
||||
<meta name="twitter:site" content="@labmlai"/>
|
||||
<meta name="twitter:creator" content="@labmlai"/>
|
||||
|
||||
<meta property="og:url" content="https://nn.labml.ai/transformers/compressive/readme.html"/>
|
||||
<meta property="og:title" content="Compressive Transformer"/>
|
||||
<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="Compressive Transformer"/>
|
||||
<meta property="og:description" content=""/>
|
||||
|
||||
<title>Compressive Transformer</title>
|
||||
<link rel="shortcut icon" href="/icon.png"/>
|
||||
<link rel="stylesheet" href="../../pylit.css">
|
||||
<link rel="canonical" href="https://nn.labml.ai/transformers/compressive/readme.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">compressive</a>
|
||||
</p>
|
||||
<p>
|
||||
|
||||
<a href="https://github.com/lab-ml/labml_nn/tree/master/labml_nn/transformers/compressive/readme.md">
|
||||
<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'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-0'>#</a>
|
||||
</div>
|
||||
<h1><a href="https://nn.labml.ai/transformers/compressive/index.html">Compressive Transformer</a></h1>
|
||||
<p>This is an implementation of
|
||||
<a href="https://arxiv.org/abs/1911.05507">Compressive Transformers for Long-Range Sequence Modelling</a>
|
||||
in <a href="https://pytorch.org">PyTorch</a>.</p>
|
||||
<p>This is an extension of <a href="https://nn.labml.ai/transformers/xl/index.html">Transformer XL</a> where past memories
|
||||
are compressed to give a longer attention range.
|
||||
That is, the furthest $n_{cm} c$ memories are compressed into
|
||||
$n_{cm}$ memories, where $c$ is the compression rate.</p>
|
||||
<h2>Compression operation</h2>
|
||||
<p>The compression operation is defined as
|
||||
$f_c: \mathbb{R}^{nc \times d} \rightarrow \mathbb{R}^{n \times d}$.
|
||||
The paper introduces multiple choices for $f_c$ and we have only implemented
|
||||
1D convolution which seems to give the best results.
|
||||
Each layer has a separate compression operation $f_c^{(i)}$ where
|
||||
$i$ is the layer number.</p>
|
||||
<h2>Training compression operation</h2>
|
||||
<p>Since training compression with BPTT requires maintaining
|
||||
a very large computational graph (many time steps), the paper proposes
|
||||
an <em>auto-encoding loss</em> and an <em>attention reconstruction loss</em>.
|
||||
The auto-encoding loss decodes the original memories from the compressed memories
|
||||
and calculates the loss.
|
||||
Attention reconstruction loss computes the multi-headed attention results
|
||||
on the compressed memory and on uncompressed memory and gets a mean squared error
|
||||
between them.
|
||||
We have implemented the latter here since it gives better results.</p>
|
||||
<p>This implementation uses pre-layer normalization
|
||||
while the paper uses post-layer normalization.
|
||||
Pre-layer norm does the layer norm before FFN[../feedforward.html) and
|
||||
self-attention, and the pass-through in the residual connection is not normalized.
|
||||
This is supposed to be more stable in standard transformer setups.</p>
|
||||
<p>Here are <a href="https://nn.labml.ai/transformers/compressive/experiment.html">the training code</a> and a notebook for training a compressive transformer
|
||||
model on the Tiny Shakespeare dataset.</p>
|
||||
<p><a href="https://colab.research.google.com/github/lab-ml/nn/blob/master/labml_nn/transformers/compressive/experiment.ipynb"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a>
|
||||
<a href="https://web.lab-ml.com/run?uuid=0d9b5338726c11ebb7c80242ac1c0002"><img alt="View Run" src="https://img.shields.io/badge/labml-experiment-brightgreen" /></a></p>
|
||||
</div>
|
||||
<div class='code'>
|
||||
|
||||
</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>
|
||||
@ -10,6 +10,15 @@ summary: >
|
||||
This is a [PyTorch](https://pytorch.org) implementation of
|
||||
[Proximal Policy Optimization - PPO](https://arxiv.org/abs/1707.06347).
|
||||
|
||||
PPO is a policy gradient method for reinforcement learning.
|
||||
Simple policy gradient methods one do a single gradient update per sample (or a set of samples).
|
||||
Doing multiple gradient steps for a singe sample causes problems
|
||||
because the policy deviates too much producing a bad policy.
|
||||
PPO lets us do multiple gradient updates per sample by trying to keep the
|
||||
policy close to the policy that was used to sample data.
|
||||
It does so by clipping gradient flow if the updated policy
|
||||
is not close to the policy used to sample the data.
|
||||
|
||||
You can find an experiment that uses it [here](experiment.html).
|
||||
The experiment uses [Generalized Advantage Estimation](gae.html).
|
||||
"""
|
||||
@ -24,6 +33,8 @@ class ClippedPPOLoss(Module):
|
||||
"""
|
||||
## PPO Loss
|
||||
|
||||
Here's how the PPO update rule is derived.
|
||||
|
||||
We want to maximize policy reward
|
||||
$$\max_\theta J(\pi_\theta) =
|
||||
\mathop{\mathbb{E}}_{\tau \sim \pi_\theta}\Biggl[\sum_{t=0}^\infty \gamma^t r_t \Biggr]$$
|
||||
@ -128,6 +139,8 @@ class ClippedPPOLoss(Module):
|
||||
# *this is different from rewards* $r_t$.
|
||||
ratio = torch.exp(log_pi - sampled_log_pi)
|
||||
|
||||
# ### Cliping the policy ratio
|
||||
#
|
||||
# \begin{align}
|
||||
# \mathcal{L}^{CLIP}(\theta) =
|
||||
# \mathbb{E}_{a_t, s_t \sim \pi_{\theta{OLD}}} \biggl[
|
||||
@ -167,6 +180,8 @@ class ClippedValueFunctionLoss(Module):
|
||||
"""
|
||||
## Clipped Value Function Loss
|
||||
|
||||
Similarly we clip the value function update also.
|
||||
|
||||
\begin{align}
|
||||
V^{\pi_\theta}_{CLIP}(s_t)
|
||||
&= clip\Bigl(V^{\pi_\theta}(s_t) - \hat{V_t}, -\epsilon, +\epsilon\Bigr)
|
||||
|
||||
16
labml_nn/rl/ppo/readme.md
Normal file
16
labml_nn/rl/ppo/readme.md
Normal file
@ -0,0 +1,16 @@
|
||||
# [Proximal Policy Optimization (PPO)](https://nn.labml.ai/rl/ppo/index.html)
|
||||
|
||||
This is a [PyTorch](https://pytorch.org) implementation of
|
||||
[Proximal Policy Optimization - PPO](https://arxiv.org/abs/1707.06347).
|
||||
|
||||
PPO is a policy gradient method for reinforcement learning.
|
||||
Simple policy gradient methods one do a single gradient update per sample (or a set of samples).
|
||||
Doing multiple gradient steps for a singe sample causes problems
|
||||
because the policy deviates too much producing a bad policy.
|
||||
PPO lets us do multiple gradient updates per sample by trying to keep the
|
||||
policy close to the policy that was used to sample data.
|
||||
It does so by clipping gradient flow if the updated policy
|
||||
is not close to the policy used to sample the data.
|
||||
|
||||
You can find an experiment that uses it [here](https://nn.labml.ai/rl/ppo/experiment.html).
|
||||
The experiment uses [Generalized Advantage Estimation](https://nn.labml.ai/rl/ppo/gae.html).
|
||||
Reference in New Issue
Block a user