update documentation

docs/_build/html/_sources/getting_started.rst.txt

@@ -1,3 +1,50 @@
 Tutorial
 ========
 
+*AnalogAINAS* is a framework that aims at building analog-aware efficient deep learning models. AnalogNAS is built on top of the [AIHWKIT](https://github.com/IBM/aihwkit). IBM Analog Hardware Acceleration Kit (AIHWKIT) is an open source Python toolkit for exploring and using the capabilities of in-memory computing devices in the context of artificial intelligence.
+
+In a high-level AnalogAINAS consists of 4 main building blocks which (can) interact with each other:
+
+* Configuration spaces: a search space of architectures targeting a specific dataset.
+* Evaluator: a ML predictor model to predict: 
+    * 1-day Accuracy: the evaluator models the drift effect that is encountered in Analog devices. The accuracy after 1 day of drift is then predicted and used as an objective to maximize. 
+    * The Accuracy Variation for One Month (AVM): The difference between the accuracy after 1 month and the accuracy after 1 sec. 
+    * The 1-day accuracy standard deviation: The stochasticity of the noise induces different variation of the model's accuracy depending on its architecture. 
+* Optimizer: a optimization strategy such as evolutionary algorithm or bayesian optimization. 
+* Worker: A global object that runs the architecture search loop and the final network training pipeline
+
+Create a Configuration Space
+----------------------------
+
+AnalogNAS presents a general search space composed of ResNet-like architectures. 
+
+The macro-architecture defined in the file ```search_spaces/resnet_macro_architecture.py``` is customizable to any image classification dataset, given an input shape and output classes. 
+
+.. warning::
+    The hyperparameters in the configuration space should have a unique name ID each. 
+
+Evaluator 
+---------
+
+To speed up the search, we built a machine learning predictor to evaluate the accuracy and robustness of any given architecture from the configuration space. 
+
+Search Optimizer and Worker
+---------------------------
+
+In this example, we will use evolutionary search to look for the best architecture in CS using our evaluator. 
+
+::
+
+    from analogainas.search_algorithms.ea_optimized import EAOptimizer
+    from analogainas.search_algorithms.worker import Worker
+
+    optimizer = EAOptimizer(evaluator, population_size=20, nb_iter=10)  
+
+    NB_RUN = 2
+    worker = Worker(CS, optimizer=optimizer, runs=NB_RUN)
+
+    worker.search()
+
+    worker.result_summary()
+
+

docs/_build/html/getting_started.html

@@ -19,6 +19,7 @@
     <script src="_static/js/theme.js"></script>
     <link rel="index" title="Index" href="genindex.html" />
     <link rel="search" title="Search" href="search.html" />
+    <link rel="next" title="API Reference" href="api_references.html" />
     <link rel="prev" title="Installation" href="install.html" /> 
 </head>
 
@@ -44,7 +45,17 @@
               <p class="caption" role="heading"><span class="caption-text">Get started</span></p>
 <ul class="current">
 <li class="toctree-l1"><a class="reference internal" href="install.html">Installation</a></li>
-<li class="toctree-l1 current"><a class="current reference internal" href="#">Tutorial</a></li>
+<li class="toctree-l1 current"><a class="current reference internal" href="#">Tutorial</a><ul>
+<li class="toctree-l2"><a class="reference internal" href="#create-a-configuration-space">Create a Configuration Space</a></li>
+<li class="toctree-l2"><a class="reference internal" href="#evaluator">Evaluator</a></li>
+<li class="toctree-l2"><a class="reference internal" href="#search-optimizer-and-worker">Search Optimizer and Worker</a></li>
+</ul>
+</li>
+</ul>
+<p class="caption" role="heading"><span class="caption-text">References</span></p>
+<ul>
+<li class="toctree-l1"><a class="reference internal" href="api_references.html">API Reference</a></li>
+<li class="toctree-l1"><a class="reference internal" href="references.html">Paper References</a></li>
 </ul>
 
         </div>
@@ -73,13 +84,60 @@
 
   <section id="tutorial">
 <h1>Tutorial<a class="headerlink" href="#tutorial" title="Permalink to this heading"></a></h1>
+<p><em>AnalogAINAS</em> is a framework that aims at building analog-aware efficient deep learning models. AnalogNAS is built on top of the [AIHWKIT](<a class="reference external" href="https://github.com/IBM/aihwkit">https://github.com/IBM/aihwkit</a>). IBM Analog Hardware Acceleration Kit (AIHWKIT) is an open source Python toolkit for exploring and using the capabilities of in-memory computing devices in the context of artificial intelligence.</p>
+<p>In a high-level AnalogAINAS consists of 4 main building blocks which (can) interact with each other:</p>
+<ul class="simple">
+<li><p>Configuration spaces: a search space of architectures targeting a specific dataset.</p></li>
+<li><dl class="simple">
+<dt>Evaluator: a ML predictor model to predict:</dt><dd><ul>
+<li><p>1-day Accuracy: the evaluator models the drift effect that is encountered in Analog devices. The accuracy after 1 day of drift is then predicted and used as an objective to maximize.</p></li>
+<li><p>The Accuracy Variation for One Month (AVM): The difference between the accuracy after 1 month and the accuracy after 1 sec.</p></li>
+<li><p>The 1-day accuracy standard deviation: The stochasticity of the noise induces different variation of the model’s accuracy depending on its architecture.</p></li>
+</ul>
+</dd>
+</dl>
+</li>
+<li><p>Optimizer: a optimization strategy such as evolutionary algorithm or bayesian optimization.</p></li>
+<li><p>Worker: A global object that runs the architecture search loop and the final network training pipeline</p></li>
+</ul>
+<section id="create-a-configuration-space">
+<h2>Create a Configuration Space<a class="headerlink" href="#create-a-configuration-space" title="Permalink to this heading"></a></h2>
+<p>AnalogNAS presents a general search space composed of ResNet-like architectures.</p>
+<p>The macro-architecture defined in the file <code class="docutils literal notranslate"><span class="pre">`search_spaces/resnet_macro_architecture.py`</span></code> is customizable to any image classification dataset, given an input shape and output classes.</p>
+<div class="admonition warning">
+<p class="admonition-title">Warning</p>
+<p>The hyperparameters in the configuration space should have a unique name ID each.</p>
+</div>
+</section>
+<section id="evaluator">
+<h2>Evaluator<a class="headerlink" href="#evaluator" title="Permalink to this heading"></a></h2>
+<p>To speed up the search, we built a machine learning predictor to evaluate the accuracy and robustness of any given architecture from the configuration space.</p>
+</section>
+<section id="search-optimizer-and-worker">
+<h2>Search Optimizer and Worker<a class="headerlink" href="#search-optimizer-and-worker" title="Permalink to this heading"></a></h2>
+<p>In this example, we will use evolutionary search to look for the best architecture in CS using our evaluator.</p>
+<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">analogainas.search_algorithms.ea_optimized</span> <span class="kn">import</span> <span class="n">EAOptimizer</span>
+<span class="kn">from</span> <span class="nn">analogainas.search_algorithms.worker</span> <span class="kn">import</span> <span class="n">Worker</span>
+
+<span class="n">optimizer</span> <span class="o">=</span> <span class="n">EAOptimizer</span><span class="p">(</span><span class="n">evaluator</span><span class="p">,</span> <span class="n">population_size</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">nb_iter</span><span class="o">=</span><span class="mi">10</span><span class="p">)</span>
+
+<span class="n">NB_RUN</span> <span class="o">=</span> <span class="mi">2</span>
+<span class="n">worker</span> <span class="o">=</span> <span class="n">Worker</span><span class="p">(</span><span class="n">CS</span><span class="p">,</span> <span class="n">optimizer</span><span class="o">=</span><span class="n">optimizer</span><span class="p">,</span> <span class="n">runs</span><span class="o">=</span><span class="n">NB_RUN</span><span class="p">)</span>
+
+<span class="n">worker</span><span class="o">.</span><span class="n">search</span><span class="p">()</span>
+
+<span class="n">worker</span><span class="o">.</span><span class="n">result_summary</span><span class="p">()</span>
+</pre></div>
+</div>
+</section>
 </section>
 
 
            </div>
           </div>
           <footer><div class="rst-footer-buttons" role="navigation" aria-label="Footer">
         <a href="install.html" class="btn btn-neutral float-left" title="Installation" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left" aria-hidden="true"></span> Previous</a>
+        <a href="api_references.html" class="btn btn-neutral float-right" title="API Reference" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right" aria-hidden="true"></span></a>
     </div>
 
   <hr/>