Publish docs @ 89f360d

Author: github-actions[bot]

faq/index.html

@@ -781,14 +781,21 @@ <h2 class="anchored" data-anchor-id="why-is-this-variable-being-treated-as-rando
 <span id="cb3-6"><a href="#cb3-6" aria-hidden="true" tabindex="-1"></a>m2 <span class="op">=</span> <span class="fu">f2</span>() <span class="op">|</span> (<span class="pp">@varname</span>(x[<span class="fl">1</span>]) <span class="op">=&gt;</span> <span class="fl">1.0</span>)</span>
 <span id="cb3-7"><a href="#cb3-7" aria-hidden="true" tabindex="-1"></a><span class="fu">sample</span>(m2, <span class="fu">NUTS</span>(), <span class="fl">100</span>) <span class="co"># This doesn't work!</span></span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
 <p>The key insight is that <code>filldist</code> creates a single distribution (not N independent distributions), which is why you cannot condition on individual elements. The distinction is not just about what appears on the LHS of <code>~</code>, but whether you’re dealing with separate distributions (<code>.~</code> with univariate) or a single distribution over multiple values (<code>~</code> with multivariate or <code>filldist</code>).</p>
-<p>To understand more about how Turing determines whether a variable is treated as random or observed, see: - <a href="../core-functionality">Core Functionality</a> - basic explanation of the <code>~</code> notation and conditioning</p>
+<p>To understand more about how Turing determines whether a variable is treated as random or observed, see:</p>
+<ul>
+<li><a href="../core-functionality">Core Functionality</a> - basic explanation of the <code>~</code> notation and conditioning</li>
+</ul>
 </section>
 <section id="can-i-use-parallelism-threads-in-my-model" class="level2">
 <h2 class="anchored" data-anchor-id="can-i-use-parallelism-threads-in-my-model">Can I use parallelism / threads in my model?</h2>
 <p>Yes, but with important caveats! There are two types of parallelism to consider:</p>
 <section id="parallel-sampling-multiple-chains" class="level3">
 <h3 class="anchored" data-anchor-id="parallel-sampling-multiple-chains">1. Parallel Sampling (Multiple Chains)</h3>
-<p>Turing.jl fully supports sampling multiple chains in parallel: - <strong>Multithreaded sampling</strong>: Use <code>MCMCThreads()</code> to run one chain per thread - <strong>Distributed sampling</strong>: Use <code>MCMCDistributed()</code> for distributed computing</p>
+<p>Turing.jl fully supports sampling multiple chains in parallel:</p>
+<ul>
+<li><strong>Multithreaded sampling</strong>: Use <code>MCMCThreads()</code> to run one chain per thread</li>
+<li><strong>Distributed sampling</strong>: Use <code>MCMCDistributed()</code> for distributed computing</li>
+</ul>
 <p>See the <a href="../core-functionality/#sampling-multiple-chains">Core Functionality guide</a> for examples.</p>
 </section>
 <section id="threading-within-models" class="level3">
@@ -801,17 +808,30 @@ <h3 class="anchored" data-anchor-id="threading-within-models">2. Threading Withi
 <span id="cb4-5"><a href="#cb4-5" aria-hidden="true" tabindex="-1"></a>        y[i] <span class="op">~</span> <span class="fu">Normal</span>(x[i])  <span class="co"># `observe` statements are okay</span></span>
 <span id="cb4-6"><a href="#cb4-6" aria-hidden="true" tabindex="-1"></a>    <span class="cf">end</span></span>
 <span id="cb4-7"><a href="#cb4-7" aria-hidden="true" tabindex="-1"></a><span class="kw">end</span></span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
-<p><strong>Important limitations:</strong> - <strong>Observe statements</strong>: Generally safe to use in threaded loops - <strong>Assume statements</strong> (sampling statements): Often crash unpredictably or produce incorrect results - <strong>AD backend compatibility</strong>: Many AD backends don’t support threading. Check the <a href="https://turinglang.org/ADTests/">multithreaded column in ADTests</a> for compatibility</p>
+<p><strong>Important limitations:</strong></p>
+<ul>
+<li><strong>Observe statements</strong>: Generally safe to use in threaded loops</li>
+<li><strong>Assume statements</strong> (sampling statements): Often crash unpredictably or produce incorrect results</li>
+<li><strong>AD backend compatibility</strong>: Many AD backends don’t support threading. Check the <a href="https://turinglang.org/ADTests/">multithreaded column in ADTests</a> for compatibility</li>
+</ul>
 <p>For safe parallelism within models, consider vectorized operations instead of explicit threading.</p>
 </section>
 </section>
 <section id="how-do-i-check-the-type-stability-of-my-turing-model" class="level2">
 <h2 class="anchored" data-anchor-id="how-do-i-check-the-type-stability-of-my-turing-model">How do I check the type stability of my Turing model?</h2>
-<p>Type stability is crucial for performance. Check out: - <a href="../usage/performance-tips">Performance Tips</a> - includes specific advice on type stability - Use <code>DynamicPPL.DebugUtils.model_warntype</code> to check type stability of your model</p>
+<p>Type stability is crucial for performance. Check out:</p>
+<ul>
+<li><a href="../usage/performance-tips">Performance Tips</a> - includes specific advice on type stability</li>
+<li>Use <code>DynamicPPL.DebugUtils.model_warntype</code> to check type stability of your model</li>
+</ul>
 </section>
 <section id="how-do-i-debug-my-turing-model" class="level2">
 <h2 class="anchored" data-anchor-id="how-do-i-debug-my-turing-model">How do I debug my Turing model?</h2>
-<p>For debugging both statistical and syntactical issues: - <a href="../usage/troubleshooting">Troubleshooting Guide</a> - common errors and their solutions - For more advanced debugging, DynamicPPL provides <a href="https://turinglang.org/DynamicPPL.jl/stable/api/#Debugging-Utilities">the <code>DynamicPPL.DebugUtils</code> module</a> for inspecting model internals</p>
+<p>For debugging both statistical and syntactical issues:</p>
+<ul>
+<li><a href="../usage/troubleshooting">Troubleshooting Guide</a> - common errors and their solutions</li>
+<li>For more advanced debugging, DynamicPPL provides <a href="https://turinglang.org/DynamicPPL.jl/stable/api/#Debugging-Utilities">the <code>DynamicPPL.DebugUtils</code> module</a> for inspecting model internals</li>
+</ul>
 </section>
 <section id="what-are-the-main-differences-between-turing-bugs-and-stan-syntax" class="level2">
 <h2 class="anchored" data-anchor-id="what-are-the-main-differences-between-turing-bugs-and-stan-syntax">What are the main differences between Turing, BUGS, and Stan syntax?</h2>
@@ -845,11 +865,22 @@ <h2 class="anchored" data-anchor-id="what-are-the-main-differences-between-turin
 </section>
 <section id="which-automatic-differentiation-backend-should-i-use" class="level2">
 <h2 class="anchored" data-anchor-id="which-automatic-differentiation-backend-should-i-use">Which automatic differentiation backend should I use?</h2>
-<p>The choice of AD backend can significantly impact performance. See: - <a href="../usage/automatic-differentiation">Automatic Differentiation Guide</a> - comprehensive comparison of ForwardDiff, Mooncake, ReverseDiff, and other backends - <a href="../usage/performance-tips#choose-your-ad-backend">Performance Tips</a> - quick guide on choosing backends - <a href="https://turinglang.org/ADTests/">AD Backend Benchmarks</a> - performance comparisons across various models</p>
+<p>The choice of AD backend can significantly impact performance. See:</p>
+<ul>
+<li><a href="../usage/automatic-differentiation">Automatic Differentiation Guide</a> - comprehensive comparison of ForwardDiff, Mooncake, ReverseDiff, and other backends</li>
+<li><a href="../usage/performance-tips#choose-your-ad-backend">Performance Tips</a> - quick guide on choosing backends</li>
+<li><a href="https://turinglang.org/ADTests/">AD Backend Benchmarks</a> - performance comparisons across various models</li>
+</ul>
 </section>
 <section id="i-changed-one-line-of-my-model-and-now-its-so-much-slower-why" class="level2">
 <h2 class="anchored" data-anchor-id="i-changed-one-line-of-my-model-and-now-its-so-much-slower-why">I changed one line of my model and now it’s so much slower; why?</h2>
-<p>Small changes can have big performance impacts. Common culprits include: - Type instability introduced by the change - Switching from vectorized to scalar operations (or vice versa) - Inadvertently causing AD backend incompatibilities - Breaking assumptions that allowed compiler optimizations</p>
+<p>Small changes can have big performance impacts. Common culprits include:</p>
+<ul>
+<li>Type instability introduced by the change</li>
+<li>Switching from vectorized to scalar operations (or vice versa)</li>
+<li>Inadvertently causing AD backend incompatibilities</li>
+<li>Breaking assumptions that allowed compiler optimizations</li>
+</ul>
 <p>See our <a href="../usage/performance-tips">Performance Tips</a> and <a href="../usage/troubleshooting">Troubleshooting Guide</a> for debugging performance regressions.</p>