https://elated-raman-42e0c2.netlify.app/tags/bayesopt/Recent content in Bayesopt on Sawyer Zheng's BlogHugozh-cnWed, 24 Sep 2025 14:10:50 +0800https://elated-raman-42e0c2.netlify.app/post/notes/python/smac/Tue, 31 Dec 2024 00:00:00 +0000https://elated-raman-42e0c2.netlify.app/post/notes/python/smac/<div id="outline-container-headline-1" class="outline-2"> <h2 id="headline-1"> 入口类 </h2> <div id="outline-text-headline-1" class="outline-text-2"> <ol> <li>Facade</li> <li>文档最完善的入口类： <code class="verbatim">smac.facade.smac_ac_facade.SMAC4AC</code></li> </ol> </div> </div> <div id="outline-container-headline-2" class="outline-2"> <h2 id="headline-2"> 程序入口 </h2> <div id="outline-text-headline-2" class="outline-text-2"> <ol> <li><code>Facade.optimize()</code> –&gt; <code>SMBO.run()</code> (Facade.solver().run())</li> <li> <p><code>SMBO.run()</code> 迭代循环</p> <ol> <li> <p><code>Intensifier.get_next_run()</code> (SMBO.run()–&gt; SMBO.intensifier.get_next_run()) –&gt; <code>AbstractRacer._next_challenger()</code></p> <ul> <li>AbstractRacer 是 Intensifier 的基类</li> <li>选择下一个 challenger (候选配置config)</li> <li> <p>方法：</p>https://elated-raman-42e0c2.netlify.app/post/notes/python/smac2/Tue, 31 Dec 2024 00:00:00 +0000https://elated-raman-42e0c2.netlify.app/post/notes/python/smac2/<div id="outline-container-headline-1" class="outline-2"> <h2 id="headline-1"> multi-objectives </h2> <div id="outline-text-headline-1" class="outline-text-2"> <div id="outline-container-headline-2" class="outline-3"> <h3 id="headline-2"> runhistory 转换成 单个cost 的工具 </h3> <div id="outline-text-headline-2" class="outline-text-3"> <p><code>RunHistoryLogScaledEncoder</code> 基类： <code>AbstractRunHistoryEncoder</code></p> <p> 通过重载 <code>_build_matrix()</code> 方法实现</p> </div> </div> <div id="outline-container-headline-3" class="outline-3"> <h3 id="headline-3"> 数据收集用来做模型训练的方法 </h3> <div id="outline-text-headline-3" class="outline-text-3"> <p><code>ConfigSelector</code></p> <ul> <li>通过 <code>self._collect_data()</code> 实现收集</li> <li>通过 <code>self._get_x_best()</code> 实现选取最优点 (X,y) 单目标 cost 数据</li> </ul> </div> </div> <div id="outline-container-headline-4" class="outline-3"> <h3 id="headline-4"> 多目标代理模型类 </h3> <div id="outline-text-headline-4" class="outline-text-3"> <p><code>MultiObjectiveModel</code></p>https://elated-raman-42e0c2.netlify.app/post/notes/python/smac2.sync-conflict-20250905-133604-irqgsmp/Tue, 31 Dec 2024 00:00:00 +0000https://elated-raman-42e0c2.netlify.app/post/notes/python/smac2.sync-conflict-20250905-133604-irqgsmp/<div id="outline-container-headline-1" class="outline-2"> <h2 id="headline-1"> multi-objectives </h2> <div id="outline-text-headline-1" class="outline-text-2"> <div id="outline-container-headline-2" class="outline-3"> <h3 id="headline-2"> runhistory 转换成 单个cost 的工具 </h3> <div id="outline-text-headline-2" class="outline-text-3"> <p><code>RunHistoryLogScaledEncoder</code> 基类： <code>AbstractRunHistoryEncoder</code></p> <p> 通过重载 <code>_build_matrix()</code> 方法实现</p> </div> </div> <div id="outline-container-headline-3" class="outline-3"> <h3 id="headline-3"> 数据收集用来做模型训练的方法 </h3> <div id="outline-text-headline-3" class="outline-text-3"> <p><code>ConfigSelector</code></p> <ul> <li>通过 <code>self._collect_data()</code> 实现收集</li> <li>通过 <code>self._get_x_best()</code> 实现选取最优点 (X,y) 单目标 cost 数据</li> </ul> </div> </div> <div id="outline-container-headline-4" class="outline-3"> <h3 id="headline-4"> 多目标代理模型类 </h3> <div id="outline-text-headline-4" class="outline-text-3"> <p><code>MultiObjectiveModel</code></p>