build based on 16db639

dev/.documenter-siteinfo.json

@@ -1 +1 @@
-{"documenter":{"julia_version":"1.6.7","generation_timestamp":"2023-10-05T05:11:03","documenter_version":"1.1.0"}}
+{"documenter":{"julia_version":"1.6.7","generation_timestamp":"2023-10-10T03:29:15","documenter_version":"1.1.0"}}

dev/index.html

@@ -13,7 +13,7 @@
 preds = m(x_train)</code></pre><h3 id="Tables-and-DataFrames-input"><a class="docs-heading-anchor" href="#Tables-and-DataFrames-input">Tables and DataFrames input</a><a id="Tables-and-DataFrames-input-1"></a><a class="docs-heading-anchor-permalink" href="#Tables-and-DataFrames-input" title="Permalink"></a></h3><p>When using a <code>Tables</code> compatible input such as <code>DataFrames</code>, features with element type <code>Real</code> (incl. <code>Bool</code>) and <code>Categorical</code> are automatically recognized as input features. Alternatively, <code>fnames</code> kwarg can be used. </p><p><code>Categorical</code> features are treated accordingly by the algorithm. Ordered variables will be treated as numerical features, using <code>≤</code> split rule, while unordered variables are using <code>==</code>. Support is currently limited to a maximum of 255 levels. <code>Bool</code> variables are treated as unordered, 2-levels cat variables.</p><pre><code class="language-julia hljs">dtrain = DataFrame(x_train, :auto)
 dtrain.y .= y_train
 m = fit_evotree(config, dtrain; target_name=&quot;y&quot;);
-m = fit_evotree(config, dtrain; target_name=&quot;y&quot;, fnames=[&quot;x1&quot;, &quot;x3&quot;]);</code></pre><h3 id="GPU-Acceleration"><a class="docs-heading-anchor" href="#GPU-Acceleration">GPU Acceleration</a><a id="GPU-Acceleration-1"></a><a class="docs-heading-anchor-permalink" href="#GPU-Acceleration" title="Permalink"></a></h3><p>If running on a CUDA enabled machine, training and inference on GPU can be triggered through the <code>device</code> kwarg: </p><pre><code class="language-julia hljs">m = fit_evotree(config, dtrain; target_name=&quot;y&quot;, device=&quot;gpu&quot;);
+m = fit_evotree(config, dtrain; target_name=&quot;y&quot;, fnames=[&quot;x1&quot;, &quot;x3&quot;]);</code></pre><h3 id="GPU-Acceleration"><a class="docs-heading-anchor" href="#GPU-Acceleration">GPU Acceleration</a><a id="GPU-Acceleration-1"></a><a class="docs-heading-anchor-permalink" href="#GPU-Acceleration" title="Permalink"></a></h3><p>EvoTrees supports training and inference on Nvidia GPU&#39;s with <a href="https://github.com/JuliaGPU/CUDA.jl">CUDA.jl</a>. Note that on Julia ≥ 1.9 CUDA support is only enabled when CUDA.jl is installed and loaded, by another package or explicitly with e.g.</p><pre><code class="language-julia hljs">using CUDA</code></pre><p>If running on a CUDA enabled machine, training and inference on GPU can be triggered through the <code>device</code> kwarg: </p><pre><code class="language-julia hljs">m = fit_evotree(config, dtrain; target_name=&quot;y&quot;, device=&quot;gpu&quot;);
 p = m(dtrain; device=&quot;gpu&quot;)</code></pre><h2 id="Reproducibility"><a class="docs-heading-anchor" href="#Reproducibility">Reproducibility</a><a id="Reproducibility-1"></a><a class="docs-heading-anchor-permalink" href="#Reproducibility" title="Permalink"></a></h2><p>EvoTrees models trained on cpu can be fully reproducible.</p><p>Models of the gradient boosting family typically involve some stochasticity.  In EvoTrees, this primarily concern the the 2 subsampling parameters <code>rowsample</code> and <code>colsample</code>. The other stochastic operation happens at model initialisation when the features are binarized to allow for fast histogram construction: a random subsample of <code>1_000 * nbins</code> is used to compute the breaking points. </p><p>These random parts of the algorithm can be deterministically reproduced on cpu by specifying an <code>rng</code> to the model constructor. <code>rng</code> can be an integer (ex: <code>123</code>) or a random generator (ex: <code>Random.Xoshiro(123)</code>).  If no <code>rng</code> is specified, <code>123</code> is used by default. When an integer <code>rng</code> is used, a <code>Random.MersenneTwister</code> generator will be created by the EvoTrees&#39;s constructor. Otherwise, the provided random generator will be used.  </p><p>Consequently, the following <code>m1</code> and <code>m2</code> models will be identical:</p><pre><code class="language-julia hljs">config = EvoTreeRegressor(rowsample=0.5, rng=123)
 m1 = fit_evotree(config, df; target_name=&quot;y&quot;);
 config = EvoTreeRegressor(rowsample=0.5, rng=123)
@@ -40,4 +40,4 @@
  &quot;a&quot;
  &quot;b&quot;
  &quot;missing value&quot;</code></pre><h3 id="Target"><a class="docs-heading-anchor" href="#Target">Target</a><a id="Target-1"></a><a class="docs-heading-anchor-permalink" href="#Target" title="Permalink"></a></h3><p>Target variable must have its element type <code>&lt;:Real</code>. Only exception is for <code>EvoTreeClassifier</code> for which <code>CategoricalValue</code>, <code>Integer</code>, <code>String</code> and <code>Char</code> are supported.</p><h2 id="Save/Load"><a class="docs-heading-anchor" href="#Save/Load">Save/Load</a><a id="Save/Load-1"></a><a class="docs-heading-anchor-permalink" href="#Save/Load" title="Permalink"></a></h2><pre><code class="language-julia hljs">EvoTrees.save(m, &quot;data/model.bson&quot;)
-m = EvoTrees.load(&quot;data/model.bson&quot;);</code></pre></article><nav class="docs-footer"><a class="docs-footer-nextpage" href="models/">Models »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option><option value="auto">Automatic (OS)</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.1.0 on <span class="colophon-date" title="Thursday 5 October 2023 05:11">Thursday 5 October 2023</span>. Using Julia version 1.6.7.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>
+m = EvoTrees.load(&quot;data/model.bson&quot;);</code></pre></article><nav class="docs-footer"><a class="docs-footer-nextpage" href="models/">Models »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option><option value="auto">Automatic (OS)</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.1.0 on <span class="colophon-date" title="Tuesday 10 October 2023 03:29">Tuesday 10 October 2023</span>. Using Julia version 1.6.7.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>